Strabismus and Amblyopia

WENNER-GREN CENTER
INTERNATIONAL SYMPOSIUM SERIES
VOLUME 49
STRABISMUS AND AMBLYOPIA

Experimental Basis for Advances
in Clinical Management
Proceedings of an International Symposium held at

The Wenner-Gren Center, Stockholm, June 24th - 26th, 1987
Edited by
Gunnar Lennerstrand
Department of Ophthalmology
Karolinska Institute
Stockholm, Sweden
Gunter K. von Noorden

Baylor College of Medicine
Houston, Texas, USA
and
Emilio C. Campos
Department of Ophthalmology,
University of Modena
Modena, Italy
M
MACMILLAN
PRESS
Scientific & Medical
© Th e Wenner-Gren Center 1988
Softcover reprint of the hardcover 1st edition 1988 978-0-333-47173-9
All rights reserved. No reproduction, copy or

transmission of this publication may be made
without written permission.
No paragraph of this publication may be

reproduced, copied or transmitted save with
written permission or in accordance with the
provisions of the Copyright Act 1956 (as
amended), or under the terms of any licence
permitting limited copying issued by the
Copyright Licensing Agency, 7, Ridgmount
Street, London WC1E 7AE.
Any person who does any unauthorised act in

relation to this publication may be liable to
criminal prosecution and civil claims for
damages.
First published 1988
Published in Great Britain by

THE MACMILLAN PRESS LTD
Houndmills, Basingstoke, Hampshire RG21 2XS
and London
Companies and representatives throughout the world
British Library Cataloguing in Publication Data

Strabismus and amblyopia : experimental
basis for advances in clinical management:
proceedings of an international symposium
held at the Wenner-Gren Center, Stockholm,
June 24-26, 1987.-(Wenner-Gren Center
international symposium series, ISSN
0083-7989; 49).
1. Strabismus
I. Title II. von Noorden, Gunter K.
III. Campos, E.C. IV. Series
617.7'62 RE771
ISBN 978-1-349-10405-5 ISBN 978-1-349-10403-1 (eBook)
DOI 10.1007/978-1-349-10403-1
CONTENTS
Preface lX
Invited Contributors and Participants Xl
Opening Address: Bjorn Tengroth 1
Session I
Oculomotor Control and Strabismus
1. Motor Dysfunction in Strabismus G. Lennerstrand 5

2. Aetiology, Classification and Clinical Characteristics of
Esotropia in Infancy B. Harcourt 23
3. Morphology of the Extraocular Muscles in Relation to the

Clinical Manifestation of Strabismus R.F. Spencer and
K.W.McNeer 37
4. Motor and Sensory Functions of Normal and Strabismic

Extraocular Muscle G. Lennerstrand 47
5. Kinetics of the Eye H.E.A. Bicas 65
6. Phyletic Organization of Brainstem Neuronal Circuits and

the Etiology of Strabismus R. Baker 77
7. Neural Basis for Changes of the Optokinetic Reflex in

Animals and Men with Strabismus and Amblyopia
K.-P. Hoffmann 89
8. Ocular Motor Phenomena in Infantile Strabismus

G. Kommerell 99
9. Phasic-Tonic Organization of Accommodation and Vergence

C.M. Schor 111
10. Clinical Aspects ofVergent Mechanisms B. Bagolini 121
11. Normative Oculomotor Development in Human Infants

R.N. Aslin 133
v
VI CONTENTS
12. DISCUSSION: Oculomotor Control and Strabismus

Moderator: G. Lennerstrand 143
Session II
Normal and Abnormal Visual Development
13. Amblyopia in Humans and Clinical Relevance of Animal
Models G.K. vonNoorden 169
14. Normal Anatomical Development of the Primate Primary

Visual Pathway L. Garey 173
15. Effects of Abnormal Visual Experience on the Morphology of

Lateral Geniculate Neurons in the Infant Primate
M.P. Headon,].]. Sloper and T.P.S. Powell 185
16. The Influence of the Period of Deprivation on Experimental

Refractive Errors E.L. Smith III, R.S. Harwerth,
M.L.]. CrawfordandG.K. vonNoorden 197
17. Electrophysiology of Cortical Neurons under different

Conditions ofVisual Deprivation M.L.].Crawford 207
18. The Sensitive Periods of the Monkey's Visual Cortex

C. Blakemore 219
19. Psychophysical Studies of Visual System Plasticity During

Critical Periods of Development R. S. Harwerth and
E.L. Smith III 235
20. Normal Visual Development and its Deviations

R. Held 247
21. Neuronal Mechanisms of Deprivation Amblyopia

W. Singer 259
22. Abnormal Binocular Interaction:Evidence in Humans

G.K. vonNoorden 275
23. DISCUSSION: Normal and Abnormal Visual Development

Moderator: G.K. von Noorden 285
Session III
Psychophysics Related to Strabismus and Amblyopia
24. Visual Perception in Strabismus E. C. Campos 311
25. Role of Oculomotor Proprioception in the Visual System of

the Cat L. Maffei and A. Fiorentini 319
CONTENTS VII
26. Extraocular Muscle Proprioception and Visual Function:

Psychophysical Aspects M.]. Steinbach, M.A. Musarella
and B.L. Gallie 327
27. Extraocular Muscle Proprioception and Visual Functions:

Clinical Aspects E.C. Campos 337
28. On the Physiological Basis for Stereopsis

M.L.]. Crawford 345
29. Stereopsis and Strabismus E.M. Helveston 359
30. Psychophysical Consequences of Image Degradation and

Binocular Misregistration on the Developing Visual Nervous
System D.M. Levi and S.A. Klein 369
31. Amblyopia: Clinical Aspects W. Haase 381
32. DISCUSSION: Psychophysics Related to Strabismus and

Amblyopia Moderator: E. C. Campos 393
33. Concluding Remarks G. Westheimer 413
34. FINAL DISCUSSION Moderator: G. Westheimer 417
35. Global Stereopsis B. fulesz 427
Index 431
PREFACE
During the last two decades a large amount of experimental research has
been devoted to visual and oculomotor physiology. Normal visual
development in animals and humans is quite well established and the
electrophysiological and behavioural effects of different forms of visual
deprivation in animals are also extensively studied. Investigations on the
developmental aspects of ocular motility are just beginning.
Parallel to this upsurge of experimental work, clinical research on

strabismus and its effects on the visual system has been very active.
However, there has been rather limited interaction and exchange of ideas
between clinical and basic scientists. The relevance of the enormous body
of basic knowledge for the diagnosis and treatment of strabismus and
amblyopia has therefore not been properly evaluated. The present meeting
has been an attempt to increase communication between the two groups,
and to reduce the time-lag between laboratory findings and clinical
applications.
The meeting had a workshop character with much time for discussion. It
was divided into three sessions. The first dealt with oculomotor control
aspects in strabismus, the second with normal and abnormal visual
development, and the third part with the psychophysics of strabismus and
amblyopia. Subsections of each part were reviewed by prominent workers
in each field. The presentations as well as the discussions of the findings
and the clinical ramifications are included in this publication.
We want to thank the participants for all the work they have invested in
the preparations and during the conference, and for the help that they have
given in editing the discussion parts. We are deeply indebted to Professor
David Ottoson for providing support on the scientific part and for the use
of the excellent conference facilities of the Wenner-Gren Center in
Stockholm. We also acknowledge the generous financial support of the
ix
X PREFACE
Wenner-Gren Center Foundation, the Swedish Medical Research Council
and the Swedish Government.
Stockholm, June 1987
Gunnar Lennerstrand
Gunter K. von Noorden
Emilio C. Campos
INVITED CONTRIBUTORS AND PARTICIPANTS
Richard Aslin Emilio Campos

Department of Psychology Department of Ophthalmology
University of Rochester University of Modena
ROCHESTER Via del Pozzo 71
N.Y. 14627 USA I-41100 MODENA
Italy
Bruno Bagolini
Eye Clinic M.L.J. Crawford
Catholic University Sensory Sciences Center and
Largo Gemelli 8 Department of Ophthalmology
I-00168 ROMA The University of Texas at
Italy Houston
HOUSTON
Robert Baker Texas 77030 USA
Department of Physiology
New York University Medi- Howard Eggers
cal Center Edward S. Harkness Eye
NEW YORK Institute
N.Y. 10016 USA 635 West 165th Street
NEW YORK
Harley Bicas N.Y. 10032 USA
Faculty of Medicine Mildred El Azazi
University of Sao Paolo Department of Ophthalmology
14.100 RIBEIRAO PRETO Huddinge Hospital
Est. Sao Paolo S-141 86 HUDDINGE
Brazil Sweden
Colin Blakemore Juan Gallo

University Laboratory of Department of Ophthalmology
Physiology Karolinska Hospital
Parks Road S-104 01 STOCKHOLM
OXFORD OXl 3PT Sweden
U.K.
Xl
xii INVITED CONTRIBUTORS and PARTICIPANTS
Laurence Garey Eugene Helveston

Institute of Anatomy Department of Ophthalmology
University of Lausanne Indiana University Medical
Rue du Bugnon 9 Center
CH-1005 LAUSANNE 702 Rotary Circle, 3rd fl
Switzerland INDIANAPOLIS
Indiana 46223 USA
Mirdza Germanis
Department of Ophthalmology Klaus-Peter Hoffmann
Huddinge Hospital Department of Neurobiology
S-141 86 HUDDINGE University of Ulm
S>V'eden P 0 Box 4066
D-7900 ULM
Hagen Goller FR Germany
Akademiska Hospital Claes von Hofsten
S-751 85 UPPSALA Department of Psychology
Sweden University of Umea
S-901 87 UMEA
Wolfgang Haase Sweden
University Hospital Annemarie Hohmann
Martinistrasse 52 Dietrichstrasse 58
D-2000 HAMBURG D-4790 PADERBORN
FR Germany FR Germany
Brian Harcourt Gerd Holmstrom

Department of Ophthalmology Department of Ophthalmology
General Infirmary Danderyd Hospital
Great George Street S-182 88 DANDERYD
LEEDS LSl 3EX Sweden
U.K.
Peter Jakobsson
Ronald Harwerth Department of Ophthalmology
College of Optometry University Hospital
University of Houston S-581 85 LINK~PING
4800 Calhoun Road Sweden
HOUSTON
Texas 77004 USA Bela Julesz
Bell Laboratories
Mauricio Headen 600 Murray Hills
Department of Human Anatomy MURRAY HILLS
South Parks Road New Jersey 07974
OXFORD OXl 3QX USA
U.K.
Guntram Kommerell
Richard Held Department of Ophthalmology
Massachusetts Institute of Killianstrasse 5
Technology D-7800 FREIBURG
79 Amherst Street El0-145 FR Germany
CAMBRIDGE
MA 02139 USA
INVITED CONTRIBUTORS and PARTICIPANTS XIII
Gunnar Lennerstrand Elisabeth Schultz

Department of Ophthalmology Department of Ophthalmology
Huddinge Hospital University Hospital
S-141 86 HUDDINGE Martinistrasse 52
Sweden D-2000 HAMBURG
FR Germany
Dennis Levi
College of Optometry Wolf Singer
University of Houston Department of Neurophysiology
4901 Calhoun Road Max-Planck-Institut
HOUSTON Deutschordenstrasse 46
Texas 77004 USA D-6000 FRANKFURT a.M. 71
FR Germany
Sivert Lindstrom
Department of Physiology Johan Sjostrand
University of Goteborg Department of Ophthalmology
S-400 33 G~TEBORG Sahlgrenska Hospital
Sweden S-413 45 G~TEBORG
Sweden
Bjorn Lundh
Department of Ophthalmology John Sloper
University Hospital Department of Human Anatomy
S-581 85 LINK~PING University of Oxford
Sweden South Park Road
OXFORD OXl 3QX
Lamberto Maffei U.K.
Via S. Zeno 51 E . L . Smith II I
I-56100 PISA College of Optometry
Italy University of Houston
4800 Calhoun Road
Birgitta Neikter HOUSTON
Department of Ophthalmology Texas 77004 USA
Danderyd Hospital
S-182 88 DANDERYD Robert Spencer
Sweden Department of Anatomy
Medical College of Virginia
Agneta Rydberg RICHMOND
Department of Ophthalmology Virginia 23298 USA
Karolinska Hospital
S-104 01 STOCKHOLM Martin Steinbach
Sweden Atkinson College
York University
Clifton Schor NORTH YORK
School of Optometry Ontario
University of California Canada M3J 1P3
BERKELEY
California 94620 USA Goran Stigmar
University Hospital
S-221 85 LUND
S-v1eden
XIV INVITED CONTRIBUTORS and PAK!'ICIPANTS
Bjorn Tengroth Olof Wennhall

Department of Ophthalmology Departme~t of Ophthalmology
Karolinska Hospital Central Hospital
S-104 01 STOCKHOLM S-721 89 VASTERAS
Sweden Sweden
Per Udden Gerald Westheimer

Hofstrasse 1 Department of Physiology-
CH-6064 KERNS Anatomy
Switzerland University of California
BERKELEY
Gunter K. von Noorden California 94720 USA
Texas Childrens Hospital Jan Ygge
P 0 Box 20269 Department of Anatomy
HOUSTON Karolinska Institute
Texas 77225 USA S-104 01 STOCKHOLM
Sweden
Peter Wanger
Sabbatsberg Hospital
Box 6401
S-113 82 STOCKHOLM
Sweden
OPENING ADDRESS
BJÖRN TENGROTH
It is a great pleasure and privilege for me to greet you all heartly

welcome to Stockholm and to this Wenner-Gren Center International
Symposium. Many of you were present at the meeting here in 1974 -
a meeting which marked, to a certain extent, a new era in modern
strabology.
Almost 250 years ago Chevalier John Taylor presented himself in this
city as the doctor who could put the eyes straight. From his coll-
eagues at the time as well as from the public in general, remarks
were made as to whether he was a humbug or not. Since then strabology
in this country as well as in the rest of the world has passed a
stage which more resembles theology than a biological science. With
our increased knowledge and understanding of the basic neurophysiolo-
gy of vision and the oculomotor system, an increasing knowledge of
the pathophysiology of strabismus and amblyopia has developed.
At the Stockholm meeting in 1974 it was mentioned that the meeting

could be looked upon as a marriage between neurophysiology and
strabology. However, at the end of the meeting a statement was made
that this was merely an engagement to get married. I believe in long
engagements before getting married, and I think that the time lapse
is far longer than what is usually common between a man and a woman.
After having looked through the program I am convinced that we are

not ready for a wedding party but we know that we are on the right
track. There is always a risk when clinicians and basic scientists
get together. The clinical knowledge is very often based on empirical
signs and sometimes only on impressions. The interpretation has very
often led into a dead end. Also the clinician's admiration for the
basic scientists, mainly because of the clinican's lack of knowledge,
might lull him into false conclusions on where the frontline goes.
The basic scientist forms his hypothesis and starts_prpjects in fields

where the questions are far from relevant for solving clinical prob-
1
2 B.TENGROTH
lems. An interesting result wrongly interpreted and applied in the

clinic might result in nothing. In the world of strabology and oculo-
motor function this is not an uncommon event. I doubt that there will
be a strabologist in the future that will be able to accomplish some-
thing new in his field without a background in basic physiology.
Electrophys1ology has characterized the last century in neurophysio-

logy.However, the modern trends show us that not only psychophysics
but also modern biochemistry and n1olecular biology have entered the
picture.
Our understanding of the normal interaction between sensory and motor

in the development of the visual system is far from clear. Recent
results in the development of the eye itself suggest an interaction
between neuropeptides and growth factors and the influence of the
neural activities on these systems. The biochemists and the physio-
logists have to have a closer relationship than before. The patho-
physiology is much more complex than we clinicians understood.
With this perspective one wonders how the clinicians will ever be
able to grasp the theoretical basis and use this knowledge to create
modern methods for diagnosis and treatment. However, the scientifi-
cally trained clinical mind will always be able to use the clinical
signs in order to form the relevant questions. All scientists in-
volved are dependent upon these questions in order to enhance re-
search projects which might lead to the right answers.
In my mind, a meeting of this kind therefore fulfills a purpose and

will always result in a step forward. It is my hope that the meeting
will be characterized by the only important and since long married
r.ouple - excellence and creativity. Separated the two might lead into
aullness or disaster.
Session I
OCULOMOTOR CONTROL
AND STRABISMUS
1
MOTOR DYSFUNCTION IN STRABISMUS
GUNNAR LENNERSTRAND
The intension of this symposium is to combine

experimental and clinical findings, andthereby
establish a platform for future research in the
area of strabismus. It is a difficult task,particu-
lary on the motor side. Basic research in the field
of oculomotor physiology has advanced rapidly
during the last decades and the new knowledge has
been incorporated to a large degree into research
and clinical practice of neurology and neuro-
ophthalmology (Lennerstrand et al 1982, Leigh & Zee
1983). However, the influence on strabismus re-
search has been rather limited. We hope that this
workshop will stimulate a more intensive exchange
of ideas between basic and clinical research also
in strabismus, and that the interaction will be
mutually beneficial. With this in mind I have tried
to make my review of motor dysfunction in strabis-
mus rather broad. I have included in the review a
very brief and somewhat superficial description of
strabismus from the point of view of epidemiology,
inheritance patterns, characteristics of different
types, etiology and treatment. This is done in
order to acquaint the laboratory research workers
with the basic concepts and terminology of stra-
bismology. For a more extensive description of all
aspects of strabismus, the excellent textbook by
von Noorden (1985) is highly recommended.
The second part of the review concentrates on
specific motor factors that are known or thought to
be of relevance in strabismus. The material in this
part will be dealt with in much greater detail by
the other speakers of this session, and it is only
in order to present an over all view of the field
and set the scene for the later discussion, that
the different topics are introduced already here.
5
6 G.LENNERSTRAND
GENERAL CONSIDERATIONS
Strabismus or squint is a condition that is
closely linked to the ability of binocular vision.
Binocular functions always suffer in strabismus,
but it is mainly in its effects on monocular vi-
sion, i.e. as a creator of amblyopia that strabis-
mus causes ophthalmological concern. These aspects
of strabismus will be treated extensively later at
this meeting. Strabismus is defined as a patho-
logical deviation of one eye with respect to the
other. In this eye position a visual object is
imaged on non-corresponding retinal areas, since
the visual axes of the two eyes do not intersect at
the fixation point (in horizontal and vertical
strabismus) or the vertical planes through the
visual axes are tilted with respect to each other
(in pure cyclotorsional strabismus).
Depending on the eye position, a strabismus
can be subdivided into the following pure forms:
esotropia or convergent strabismus, exotropia or
divergent strabismus, hyper- or hypotropia, i.e.
vertical strabismus, and incyclotropia and excyclo-
tropia or torsional strabismus. These forms can be
mixed in one and the same patient. Based on the
manner of fixation one can distinguish between
alternating strabismus, in which each eye can
fixate and the other be strabismic, or monolateral
strabismus where one eye fixates most of the time
and the other is deviated. Amblyopia is most pre-
valent in the latter form. Depending on the stra-
bismus situation in terms of time, a distinction is
made between permanent of manifest strabismus
(hetero-tropia) and intermittent squint. In the
latter condition, strabismus is seen at some times
and during other periods the eyes are straight.
Depending on the magnitude of squint angle in
various directions of gaze, a separation is made
between concomitant and incomitant strabismus. The
concomitant type infers that the angle of squint is
constant or nearly so, independent of the direction
of gaze, while the incomitant form shows an angle
of squint that varies with the gaze direction. The
cause of the latter form is usually ocular muscle
palsy and it is therefore often referred to as
paralytic strabismus.
In addition to the above mentioned manifest
types there are also forms of latent strabismus or
heterophoria. They occur mostly after the interup-
tion of binocular vision by occlusion. Depending
upon the direction of the deviation following oc-
clusion the terms exophoria, hyperphoria etc are
used. Heterophoria is very common and is regarded
MOTOR DYSFUNCTION IN STRABISMUS 7
to exist in as high a proportion as 80 % of the

population.
Strabismus is generally a childhood disease
with a strong hereditary trait. Manifest strabismus
is seen in 3-4 % of a Caucasian population and is
somewhat more frequent in girls than in boys. Non-
caucasian races show percentages of strabismus that
is considerably lower. tn USA only 0.6 % of blacks
have strabismus. In Gabon with a negroid population
the incidence of strabismus was 0.52 %. There are
also racial variations with regard to type of stra-
bismus. The ratio between esotropia and exotropia
in Europe is 5:1. In Japan the divergent form is
more prevalent and in Indonesia the convergent type
is rarely observed. For a long time the occurance
of natural strabismus was thought to be restricted
to the human race, but recently strabismus has been
discovered in monkey colonies at primate centers.
The incidence is 4 %with a high preponderance of
esotropia (Boothe et al, 1985).
If either of the parents has strabismus the
risk to develop squint is 4 times higher than in an
unselected population. 60 % of children with squint
have a close relative with the same disease. How-
ever, it is unknown how the strabismus trait is
transmitted. The inheritance pattern is multifacto-
rial and it would seem likely that it is not stra-
bismus in itself that is inherited but some other
factor that predisposes for squint. This includes
structural anomalies, refraction anomalies, accom-
modation-convergence anomalies, defects in visual
pathways etc. For example strabismus is almost
obligate in albinism (of the tyrosinase-negative
form) and a high incidence of strabismus is seen in
achromatopsia, probably due to abnormal retina-
striate connections.
SPECIFIC TYPES OF STRABISMUS

Concomitant strabismus
Infantile esotropia
This type of esotropia will be presented in
great detail by Harcourt (this symposium).
Esotropia of late onset
This form is more often familial. It starts
after 6 months of age and mostly at 1-2 years.
Accommodative factors play an important role, in
combination with hyperopia and anisometropia.
8 G.LENNERSTRAND
Disturbances of the relation between accommodation
and convergence (the AC/A-ratio), results in a
rather instabile angle of squint, usually larger at
near fixation than at distance fixation (conver-
gence excess). Esotropia of late onset is seldom
combined with so called A- or V-incomitances,
latent nystagmus or alternating occlusion hyper-
tropia which are common in infantile esotropia.
Since the esotropic eye position occurred after
binocular functions had been developing for a
period of time, some binocularity can usually be
restored after successful alignment of the eyes.
Exodeviations
The onset is very variable. Sometimes it is
congenital, but mostly it developes over a period
and progresses from a latent stage over an inter-
mittent phase and finally to a manifest divergent
strabismus that can occur at any age during child-
hood. The angle of squint is usually quite large.
Overactions of the oblique muscles and A- V-
patterns are often seen. When the angle of strabis--
mus is the same for near and distance, the condi-
tion is called basic exotropia. Convergence in-
sufficiency implies larger angle of strabismus at
near than at distance, and divergence excess a
larger deviation at distance than at near.
Microstrabismus
This entity implies that one eye has abandomed
central fixation, is positioned at an angle of
strabismus of less than a couple of degrees and has
a central scotoma and slight amblyopia. However,
this definition is somewhat artificial and the
transition to strabismus of larger angle is smooth.
Secondary microstrabismus is often the terminal
result of treatment for esotropia of early onset.
The binocular vision is better developed than in
other forms of convergent strabismus.
Treatment or concomitant strabismus
Correction of refractive errors can sometimes
reduce the angle of strabismus, as in esotropia of
the accommodative type, when correction of a hyper-
opia under optimal conditions can result in ortho-
phoria or latent strabismus. If this is not at-
tained with optical correction, orthoptic treatment
has to be instituted in order to preserve monocular
and binocular vision. Pharmacological treatment is
sometimes used. In esotropia of the convergence
excess type, drugs that induce ciliary muscle con-

traction can be instilled topically as a part of
the treatment. Recently a method of injecting weak
solutions of botulinum toxin directly into the eye
muscle has been introduced (Scott, 1981) in order
to paralyze the muscles temporarily (for 2-6
months) and create a new state of visuomotor con-
trol. In cases of concomitant strabismus this
seldom leads to orthophoria but usually a strabis-
mus of smaller angle than before the injection can
be attained.
Surgical treatment remains the most important
means of remediation in strabismus. The goal is
mostly to align the eyes in a position for the
sensory machinery to function as perfectly as
possible. This may not always imply perfect paral-
lelism of the eyes, but may involve over- and
undercorrection of the strabismus in order to reach
a satisfying sensory result, i.e. to avoid
diplopia, enchance fusion capabilities etc. In
planning and calculating for surgery, all influen-
ces on eye position of accomodation and convergence
have to be eliminated. An elaborate scheme for
establishing the mechanical position of rest has
been developed by Collins and Jampolsky (1982). It
is particularly useful in reoperations when orbital
mechanics are often abnormal. The eyes are aligned
according to measurments done both when the pa-
tients is awake and under anesthesia, but the
muscles are put on adjustable sutures so that the
final positioning can be performed in the awake
state after the operation. In this way additional
innervational effects, which are hard to predict,
can be compensated for.
Incomitant forms of strabismus

This group of strabismus contains conditions
of neuromuscular dysfunction, usually of well
established origin. The angle of strabismus varies
with the direction of gaze. Excluded are the in-
comitancies seen in combination with the concomi-
tant strabismus types, e.g. the A- and V- patterns,
the vertical deviations on lateral gaze, the alter-
nate occlusion hypertropia in early onset esotropia
etc.
Paralytic strabismus
Myogenic types include myastenia gravis, different
forms of primary myogenic dystrophies and myo-
pathies. Some types of eye muscle malformation and
10 G.LENNERSTRAND
maldevelopment such as the superior oblique tendon
sheat syndrom of Brown, the retraction syndrome of
Stilling-TUrk-Duane, or effects of muscle entrap-
ment in orbital fractures should also be mentioned
here.
Neurogenic types include strabismus caused by
lesions of the third, fourth and sixth nerves and
their nuclei, whether congenital or acquired. The
angle of strabismus and particularly the incomi-
tancy may change with time due to 'spread of
comitance' as a part of the change in balance of
eye muscle forces that occurs in reinnervation of
the paralytic muscle(s), and contracture and
overaction in the non-paralyzed muscles. Other
types of incomitant strabismus are caused by
dysfunction in the supranuclear control of ocular
motility. They will be described more closely in
the section on specific motor factors in strabis-
mus.
Treatment of an incomitant strabismus
The first concern is always to establish the
diagnosis of the neuromuscular or CNS disease of
which the strabismus is a sign and institute proper
treatment. The ensuing strabismic problem has often
to be looked after as well, particularly in paretic
varieties of strabismus. Prisms or operations are
usually needed to relieve diplopia in acquired
paralytic squint of adulthood. In children there is
the added problem of disrupted binocular vision and
risk for subsequent amblyopia that needs attention.
For the surgical treatment different new techniques
have evolved. They in- elude the use of adjustable
sutures, "Faden" or posterior myopexia operations
and different kinds of transposition procedures to
compensate for paralysis of the involved muscles.
Botulinum injection to correct strabismus has its
main indications in the field of paralytic strabis-
mus. Deviations due to defects in supranuclear
oculomotor control are more difficult to treat and
for the most part they can only be dealt with in a
symptomatic way with regard to strabismus and
diplopia.
GENERAL PATHOPHYSIOLOGY OF STRABISMUS

As will be evident from the following, the
cause or causes of manifest strabismus are known to
a very limited extent. However, some general mecha-
nisms to produce strabismus are fairly well estab-
lished. Other theories are more speculative but may
still hold a grain or truth. A description or them

will serve as a background for the review on pos-
sible specific causes or eye motility dysfunction
in section IV.
Abnormalities or the fusion mechanism
Interference with the fusion mechanism in
children and adults is an established factor in the
causation or strabismus, but how this inadequacy
comes about in childhood strabismus is a matter or
controversy, particularly whether it is congenital
as proposed by Worth or acquired as suggested by
Chavasse (von Noorden, 1985). With poor fusion the
eyes may start to squint under the influence or
precipitating factors such as hypertropia, aniso-
metropia, motor anomalies, mental disturbances,
injury during birth, trauma or illness in child-
hood, occlusion or one eye etc. To these a number
or motor factors will be added in a following
section or this paper. It should be noted that loss
or fusion in childhood most commonly leads to an
esodeviation, but to a divergent strabismus in
adulthood. The reason for this difference is un-
known.
Brain damage
In patients with brain abnormalities, for
instance those with Down's syndrome, there is a
high incidence or strabismus (40-60 % in different
materials). In children with cerebral palsy, hydro-
cephalus and other more general brain disease,
strabismus is common in addition to other abnor-
malities or ocular motility. Children suffering
from general disease, e.g. congenital heart le-
sions, have 4-6 times higher frequency or strabis-
mus than the general population. Dericiencis or
postural control has been reported in esotropic
children, but exotropics were normal in this re-
spect (Sandstedt et al, 1986). Thus, it is possible
that minor motor problems exist in children with
some types or strabismus, as a sign or general CNS
dysfunction. This may be postulated in cyclic
heterotropia with rhythmic variations on alternate
days between heterotropia and normal binocular
vision (von Noorden, 1985).
Neuromuscular anomalies
These will be described more extensively in
the following section and in other papers or this
session. They include pathophysiological factors
12 G.LENNERSTRAND
connected with orbital mechanics, eye muscle func-
tion, brain stem and cerebellar function, accom-
modation-convergence coupling, the eye movement
systems and the development of oculomotor func-
tions.
Reflexological theories
It has been suggested that strabismus is a
disturbance of the optomotor reflexes (Keiner,
1951). The position of the eyes during fetal life
would depend upon subcortical reflexes initiated by
stimulation of eye muscle proprioceptors (that in-
duce a monocular duction reflex) and the vestibular
organs (that induce a binocular reflex for ver-
sions). Light stimulation after birth initiates
the development of the optomotor reflexes, which
supersede the older, subcortical reflexes of pro-
prioception and vestibular action. Esotropia would
thus depend on an abnormal development of the opto-
motor reflexes and consists of a predominance of
the monocular adduction reflexes over those for
conjugate movements and abduction. These ideas have
been retaken by Mitsui and Tamura (1986) in their
work on the effects of muscle stretch on eye posi-
tion in esotropia and exotropia (see also a sub-
sequent paper by Lennerstrand, this symposium).
However, the disturbances in optomotor responses
could be secondary to changes in the sensory
systems, as suggested by studies on the development
of the optokinetic nystagmus (Schor, 1983;
Hoffmann, Kommerell, this symposium) and need not
be the primary cause of strabismus.
SPECIFIC MOTOR FACTORS RELATED TO STRABISMUS

In the following we will examine different
mechanisms, mainly on the motor side but also
sensory motor reflexes, which may cause an im-
balance in eye position and eye movements and
which, in isolation or in combinations, can lead to
disruption of fusion and subsequently cause
manifest strabismus. Many of these factors will be
described in much greater detail by other speakers
in this session.
Mechanical factors in the orbit

Previous investigators have put strong
emphasis on anomalies of the check ligaments that
connect the muscles and surroundings tissues. They

believed that these anatomical variations were of
prime importance in the etiology of strabismus.
More recently it has proposed that strabismus is
caused mainly by anatomical variations in the
insersions of EOM on the globe, and particularly
those of the oblique muscles (see von Noorden,
1985).
Elaborate examinations of the orbit have shown
that there exists a complicated system of connec-
tive tissue septa, between the muscles and the
surrounding orbital structures including the bone
walls (Koornneef & Zonnervald, 1985). Severance of
these septa by accidental or surgical trauma may
cause motility problems of restrictive type.
Several types of functional testing of orbital
mechanics have been developed which are of great
help in evaluating restrictive problems in the
orbit (Metz, 1983).
Mechanical factors are most likely the cause
of strabismus in cranio-facial malformations (where
also aplasia of eye muscles are common), the
Brown's syndrome and other types of defects of eye
muscles and orbital structures. Such mechanisms can
be evaluated by means of CT-scans and biomechanical
models (Clement, 1986), and they are important in
the planning of strabismus operations.
Extraocular muscles (EOM)
The different aspects on EOM motor and sensory
functions that may be of importance for strabismus
will be described more extensively in subsequent
papers by Spencer and by Lennerstrand (this sym-
posium).
Brain stem control of eye movements
A description of the premotor areas where the
signals to the motoneurons are produced, and the
programs that govern the different types of eye
movements will be presented by Baker (this sym-
posium).
With respect to strabismus some specific
lesions of brain stem structures can be identified.
They in- elude the Duane's and M6bius' syndromes,
with abducens palsy due to a hypoplasia or aplasia
of the abducens nucleus. In Duane's syndrome there
is often a coinnervation of the lateral rectus
muscle from the oculomotor complex, leading to
different forms of co-contraction patterns in the
oculomotor innervated muscles and the lateral
rectus muscles (Huber, 1974).
14 G.LENNERSTRAND
There are reports of abnormal vestibula-ocular

function in patiens with strabismus of early onset
(von Noorden, 1985. It has been shown that children
with esotropia, but not those with exotropia, have
slight abnormalies of balance and gait, without any
other obvious neurological deficits (Sandstedt et
al, 1986). These disturbances might represent signs
of dysfunction in the cerebellopontine control of
gait and postural control, and they may be linked
with the VOR dysfunction and other brain stem
abnormalities, e.g. those disclosed in studies of
brain stem auditory evoked potentials in children
with early onset esotropia.
Some of the supranuclear ocular motility dis-
turbances observed in lesions of different areas of
the brain stem resemble motility problems in stra-
bismus (Leigh & Zee, 1983). These conditions in-
clude see-saw nystagmus with an alternating ele-
vation and intorsion of one eye and depression and
extorsion of the other. The lesion is usually in
the upper brain stem, mostly in mecencephalon, in-
volving the reticular formation and the nucleus of
Cajal. Convergence insufficiency, sometimes com-
bined in an alternate fashion with nystagmic and
spasmic convergence movements, is a sign of dys-
function of the vertical gaze center. Skew devia-
tion, i.e. a vertical misalignment with right
hypertropia in right gaze and left hypertropia in
left gaze, occurs in lesions of the vestibular
system, particularly the otolith pathways, in
different parts of the brain stem. It is sometimes
hard to differentiate from vertical strabismus of
the kind caused by over- and underaction of the
oblique muscles, but there is no certain indication
that the two conditions have a common cause.
Internuclear ophthalmoplegia is caused by a
lesion in the medial longitudinal fasciculus (MLF),
the pathways for the axons of the abducens inter-
nuclear neurons to the oculomotor complex. In gaze
to the side of the lesion, an adduction deficit
causes exotropia and diplopia. A more pronounced
exodeviation, seen also in the primary position, is
found in lesions involving both the horizontal
gaze center and the MLF in the pons.
A disturbance of thalamic function is known to
cause forced movements downward and inward of the
eyes and the subject is "peering at his nose".
Possibly the lesion involves the mecencephalic
portion of the MLF. A one-sided lesion in this area
could, at least theoretically, induce vertical
strabismus on attempted vertical gaze and may
possibly be the cause of the so called double-
elevator palsy, with underaction of both the
superior rectus and the inferior oblique of the

same side. It is also possible that a dysfunction
of the vertical gaze center could cause overactions
of the oblique muscles known as strabismus sursu-
ductorius and deorsoductorius.
Children with congenital ocular motor apraxia,
i.e. an unability to perform saccadic movements in
the horizontal plane but normal vertical movements,
also often show strabismus, mainly exotropia and
convergence insufficiency. The cause of the ocular
motor apraxia is unknown, but in some cases agen-
esis of the corpus callosum and dysplasia of the
cerebellar cortex has been demonstrated (Fielder et
al, 1986).
Recent developments in brain scanning (computer-
tomography, magnetic resonance imaging and posi-
tron emission tomography) should be of great value
for studies of the oculomotor effects of specific
lesions in the CNS, and help bridge the gap between
animal research and clinical research in this area.
However, as shown in the following section, the
developmental aspects on the lesions have also to
be taken into account.
Ocular motility defects in the Arnold-Chiari mal-

formation
As mentioned previously strabismus is very
common in children with meningomyelocele, a
congenital defect of the brain and the spinal cord
combined with hydrocephalus and the Arnold-Chiari
malformation. The latter implies downward displace-
ment of the brain stem and cerebellum and herni-
ation of these structures through the foramen
magnum into the cervical spinal canal. We
(Lennerstrand, Gallo & Samuelsson, unpublished
observations) have studied ocular motility in 28
patients with meningomyelocele and Arnold-Chiari
malformations documented by means of magnetic
resonance imaging.
Strabismus was found in 21 (76%) of the
patients, with heterotropia in 11 and heterophoria
in 10. Esotropia was the most common type of mani-
fest strabismus and was seen in 9 patients. Exo-
phoria was found in 6 patients. An A-syndrome,
usually in combination with esotropia and over-
action of the superior oblique muscles, was
observed in 5 patients. Ocular motility disturb-
ances were common, most often in the form of a
defect of optokinetic nystagmus (vertically and
horizontally) and spontaneous nystagmus particu-
larly of the gaze paretic type.
::;::
TABLE
Patient data, ocular motility observations and CNS deformations in four
patients with meningomyelocele. All had normal visual acuity and visual
fields.
Ocular Motility Deformation of
Pat. Sex Age Strab. Binoc. Gaze + Sacc Purs OKN HC Mec M.O. - Shunt
-
r-- - - - -Cer -
type func + nyst - - - - - QP_.
================' ================ p=~==================== =====================' ==========I
ELN F 20 ! ET-A none LNy N(?) p p ++ +++ + ++ 16y
I GP I
C'l
LC F 11 XF-V 240 GPNy p N p + ++ ++ +++ 3w
r
i I:Tl
p p
z
TJ M 10 ortho 60 GPNy N + + + + 4y zI:Tl
l (/)
MG F 25 or tho 60 N N N N +++ +++ ++ ++ none "'...,

I "'~t:1
Explanations:
F = female, M= male, ET-A= esotropia with A-syndrome, XP-V = exophoria
with V-syndrome, Binoc func = stereoacuity in sec.,
LNy = latent nystagmus, GP =gaze paresis, GPNy = gaze paretic nystagmus
HC = hydrocephalus, Mec = mecencephalon, Cer =cerebellum,
M.O. =medulla oblongata, N =normal, P =pathological,+= slight,++= moderate,
+++ = marked.
The amount of CNS malformation, i.e. hydro-

cephalus and brain stem and cerebellar displace-
ment, was correlated with strabismus and ocular
motility problems in the group. All patients with
heterotropia and most of those with heterophoria
had ocular motility defects as well. Also in ortho-
phoric patients motility problems were often seen.
However, in the individual patient motility defects
were not always in parallel with the extent of CNS
malformation as shown by the examples presented in
the Table. Among the four patients presented the
CNS pathology was most marked in patient M.G. with
orthophoria and normal motility. It should be noted
that ocular motility disturbances in this material
were correlated to the CNS changes observed at the
present stage and that we know very little about
the extent of the malformations when the patients
were young. It is therefore possible that the
patients with the more pronounced ocular motility
defects had more extensive early CNS damage than
the patients with normal ocular motility. In the
strabismic patients the adaptive mechanisms might
have been insufficient for a normal oculomotor
development, while in the patients with normal
motility the oculomotor system had time to adapt to
a rather slowly progressing deformation of the
brain stem and cerebellum.
Role of accommodation and refraction in comitant
strabismus
Danders discovered the close relationship
between accommodation and convergence. Uncorrected
hyperopia may cause an improportionally large con-
vergence impulse and esophoria-esotropia. Myopia
may lead to exophoria-exotropia by inhibiting the
convergence since no accommodation is elicited.
Fusion is usually sufficient to keep phoria under
control, but with inadequate fusional amplitudes or
if the fusion mechanism is impaired, manifest devi-
ation might occur. Correction of the refractive
error may cause concomitant esotropia to disappear
or revert into exophoria in up to 1/3 of the cases
(von Noorden, 1985). Obviously, refractive errors
play an important part in the etiology of squint.
However, Danders theory cannot be taken to
account quantatively for the relationship between
hyperopia and esotropia, and non-accommodative
forms of esotropia certainly exist. Some indi-
viduals show an exaggerated convergence respons to
accommodation and esodeviation will result also
with a small hyperopia, or even without it. Never-
theless Danders theory remains the best substan-
18 G.LENNERSTRAND
tiated theory of the causation of strabismus, as

demonstrated by a large number of patients. Ver-
gence mechanisms in strabismus will be described by
Schor and Bagolini (this symposium).
Kinematics of eye movements; Dander's and Listing's
laws
The kinematics of eye rotations involve
movements in the horizontal and vertical planes as
well as torsional movements around the fixational
axis. However, the order of freedom in eye rotation
is limited to two according to Dander's law, which
states that for each direction of gaze there is
only one orientation of the globe in the orbit, and
Listing's law which states that there is a specific
torsion of the eye at any gaze direction (von
Noorden, 1985). Thus, the orientation of the eye
can be predicted by assuming that the eye has
followed the shortest path from the primary
position to any other fixation position. Listing's
law is upheld by the central nervous system, and
only those combinations of eye muscle innervation
that would point the eye in the desired direction
are permitted and all others excluded. However,
Listing's law is violated under conditions of
convergence and head tilt, it can be overcome with
voluntary effort and it appears to break down
during anesthesia and sleep (Nakayama, 1983). The
possibilities that the stereotyped innervational
pattern to execute movements according to Listing's
law may be disrupted in strabismus should also be
contemplated.
Oculomotor systems and neural plasticity
The ocular motility system has a large
potential to adapt to different kinds of disturb-
ances (Berthoz & Melvill Jones, 1985; Robinson,
1982). Experimental findings relevant to the area
of strabismus include:
i) changes in gain of the vestibula-ocular reflex
induced by optical devices such as telescope
lenses, dove prisms etc., ii) adaption of the
saccadic gain control to tenotomy of the horizontal
eye muscles, iii) change in vergence tonus from
wearing base-in prisms or in the AC/A ratio from
wearing periscopic glasses.
All the adaptive changes evidently occur quite
rapidly in the adult human with an intact olivo-
cerebellar system. The capability for plasticity of
the vergence and saccadic systems might be of
importance in compensation for misalignment of the

eyes in concomitant and paralytic strabismus, but
it is not known to what extent this can occur in
children with an immature nervous system or in
strabismus. Developmental aspects on the oculomotor
systems relevant in this context will be presented
by Baker and by Aslin (this symposium).
Longitudinal studies of the natural course of
strabismus in humans (Helveston, 1986) and monkeys
(Boothe et al, 1985) are just beginning. These
should be supplemented with studies on eye movement
development in patients with disturbances at dif-
ferent levels of the oculomotor plant, i.e. EOM
palsies of neurogenic or myogenic origin, lesions
in the brain stem, cerebellum, mecencephalon
including the tectum, and cerebral disturbances of
frontal eye fields, occipital-parietal areas etc.
Such investigations could be expected to supply
insights into adaptional process in the ocular
motility systems during childhood. Patients with
Duane's syndrome, meningomyelocele, oculomotor
apraxia etc., might be suitable patient groups to
study in this respect.
SUMMARY
Motor characteristics of concomitant and in-
comitant strabismus and the different treatment
methods currently in use for alignment of the eyes
have been briefly reviewed. The hereditary aspects
and racial differences in the occurance of strabis-
mus were pointed out. The present understanding of
the general pathophysiology of strabismus was
briefly reported. A number of specific ocular motor
factors related to strabismus were reviewed and
suggested for further research.
ACKNOWLEDGEMENTS
The research reported from the author's
laboratory has been supported by grants from the
Swedish Medical Research Council (No 4751) and
Karolinska institutets fonder.
REFERENCES
Berthoz, A, and Melvill Jones, G.M. (Eds.), (1985).
Adaptive Mechanisms in Gaze Control. Reveiws in
Oculomotor Research, Vol. I. Elsevier, Amsterdam.
20 G.LENNERSTRAND
Boothe, R.G., Dobson. V. and Teller, D.Y. (1985).
Postnatal development of vision in human and non-
human primate. Am. Rev. Neurosci., ~. 495-545.
Clement, R.A. (1986). A comparison of different
models of extraocular muscle cooperation. Ophthal.
Physiol. Opt., ~. 165-170.
Collins, c.c. and Jampolsky, A. (1982). Objective
calculation of strabismus surgery. In Functional
Basis of Ocular Motility Disorders. (eds.
G. Lennerstrand, D.S. Zee and E.L. Keller). Pp 185-
194, Pergamon Press, Oxford.
Fielder, A.R., Gresty, M.A., Dodd, K.L., Mellor,
D.H. and Levene, M.I. (1986). Congenital ocular
motor apraxia. Trans. Ophthalmol. Soc. UK., 105,
589-598.
Helveston, E.M. (1986). Esotropia in the first year
of life. In Pediatric Ophthalmology and Strabismus.
Transactions of the New Orleans Academy of Ophthal-
mology. Raven Press, New York.
Huber, A. (1974). Electrophysiology of the retrac-
tion syndromes. Brit. J. Ophthalmol., 58, 293-300.
Keiner, G.B.J (1951). New Viewpoints on the Origin
of Squint. Martinus Nijhoff, The Hague.
Koorneef, L. and Zonneveld, F.W. (1985). Orbital
anatomy, the direct scanning of the orbit in three
planes and their bearings on the treatment of
motility disturbances of the eye after orbital
"blow-out" fractures. Acta. Morphol. Neerl. Scand.,
23, 229-246.
Leigh, R.J. and Zee, D.S. (1983). The Neurolo~y of
Eye Movements. F.A. Davies Co., Philadelphia.
Lennerstrand, G., Zee, D.S. and Keller, E.L. (Eds)
(1982). Functional Basis of Ocular Motility
Disorders, Pergamon Press, Oxford.
Metz, B.S. (1983). Restrictive factors in strabis-
mus. Surv. Ophthalmol., 28, 71-83.
Mitsui, Y, and Tamura, 0. (1986). Strabismus and
the sensory motor reflex. Excerpta Medica Clin.
Pract. Series no. 3, Amsterdam.
Nakayama, K. (1983). Kinematics of normal and

strabismic eyes. In Vergence Eye Movements. (Eds.
C. Schor and K. Ciuffreda). Pp, 543-564.
Butterworths, New York.
Von Noorden, G.K. (1985). Burian & von Noorden's:
Binocular Vision and Ocular Motility. C.V. Mosby,
St. Louis.
Robinson, D.A. (1982). Plasticity in the oculomotor
system. Fedr. Proceed., 41, 2153-2155.
Sandstedt, P., Odenrick, P. and Lennerstrand, G.
(1986). Gait and postural control in children with
divergent strabismus. Binocular Vision 1. 141-146.
Schor, C.M. (1983). Subcortical binocular suppres-
sion affects the development of latent and opto-
kinetic nystagmus. Am. J. Optom. & Physiol.
Optics., 60, 481-502.
Scott, A.B. (1981). Botulinum toxin injection of
eye muscles to correct strabismus. Trans. Am.
Ophthalmol. Soc., 79, 734-770.
2
AETIOLOGY, CLASSIFICATION AND CLINICAL
CHARACTERISTICS OF ESOTROPIA IN INFANCY
BRIAN HARCOURT
Normal oculomotor function depends basically upon mechanisms

which control ocular versions and vergences. Failure of normal
versions is manifest clinically in gaze palsies; failure of approp-
riate vergences may give rise to non-paralytic strabismus. Fail-
ures of either or both may cause nystagmus. Even if the supra-
nuclear systems and the ocular muscles develop normally, incomitant
strabismus will ensue if there are abnormalities in ocular motor
nerve functions or mechanical defects limiting ocular rotations. In
considering the relationship between the development of oculomotor
control and the incidence of strabismus, particular attention must
therefore be concentrated on the vergence mechanisms and on the
developmental factors which can cause paralytic strabismus of neuro-
genic or mechanical origin.
The commonest form of strabismus noted in early life is eso-

tropia. This presentation considers ways in which meticulous
clinical assessment of early onset esotropia aids its sub-
classification and helps to indicate its aetiology. It also
attempts to assess the anomalies in the development of normal oculo-
motor coordination which may be the cause of strabismus. Although
the primary concern is with motor development, there is some in-
evitable extension into normal and abnormal sensory aspects of
binocular fixation. Understanding the abnormal may give additional
insights into normal developmental processes, and it also assists in
planning the most rational management of early onset strabismus.
The subject is bedevilled by problems of nomenclature, but the

terms used here follow closely those of von Noorden (1984). The
term 'infantile' esotropia infers a manifest convergent strabismus
first noted before the age of 6 months, and is preferred to
'congenital' esotropia which should be limited to instances in which
the strabismus has definitely been present from the time of the
patient's birth. Infantile esotropia is a very mixed group of dis-
orders, and expanding somewhat on von Noorden's classification, the
following sub-groups can be recognised:-
23
24 B. HARCOURT
~WITH NYSTAGMUS
ESSENTIAL INFANTILE ESOTROPIA
~WITHOUT NYSTAGMUS
NYSTAGMUS BLOCKAGE SYNDROME
EARLY ONSET ACCOMMODATIVE ESOTROPIA
CONGENITAL OR INFANTILE SIXTH CRANIAL NERVE PALSIES
DUANE'S SYNDROME
MOBIUS SYNDROME
STRABISMUS FIXUS
SYMPTOMATIC SENSORY STRABISMUS
Description here is limited to the first 6 groups, in which

developmental motor factors are thought to play an important role.
ESSENTIAL INFANTILE ESOTROPIA
The characteristic features of this disorder, apart from its

early onset, are a large and relatively stable angle of strabismus,
commonly with cross fixation and an absence of amblyopia. Affected
patients are usually emmetropic and rarely have more than a low
degree of hypermetropia or hypermetropic astigmatism.
This group of patients may, somewhat tentatively, be further

subdivided into 2 sub-groups; those who do, and those who do not
exhibit nystagmus.
a) with nystagmus:
Ciancia (1962) described the association in infants of eso-

tropia, jerky horizontal nystagmus and a face turn. Lang (1968)
delineated a group of patients with infantile strabismus, latent
nystagmus (LN) and dissociated vertical divergence (DVD), often with
an abnormal head posture. Other associated features in this group
are anomalous uniocular horizontal optokinetic responses (OKN), a
high incidence of A and V patterns and poor binocular single vision
even after intensive early treatment.
The nystagmus is latent, occurring only when the fixation of

one eye is embarrassed, or manifest latent (MLN), being present
without uniocular occlusion. The fast phase is towards the side of
the fixing eye and therefore reverses in direction depending upon
which eye is fixing, and not according to the direction of gaze.
Although the amplitude of nystagmus may be very small, so as to be
missed on initial external examination, direct ophthalmoscopy or
electro-nystagmography (ENG) indicate that nystagmus is manifest in
nearly all affected patients, for only one eye is taking part in the
'viewing' process. The intensity of nystagmus often increases as
the fixing eye moves into an abducted position, and this can account
for the compensatory head posture of face-turn towards the side of
the fixing eye.
ESOTROPIA IN INFANCY 25
The limitation of abduction seen associated with cross-

fixation is often a pseudo-paralysis; good abduction can be demon-
strated following prolonged contralateral total occlusion, and less
adequately by spinning the child around a vertical axis while the
head is in a normal position. This latter manoeuvre is complicated
by the optokinetic stimulation which may be induced and by the
suggestion that some of these children at least may have defects in
their vestibula-ocular responses (VOR) (Hoyt, 1982). In those
patients who do have persistent limitation of abduction despite
these manoeuvres, a forced duction test is rarely positive under
general anaesthesia, so that the limitation of movement has central
oculomotor neurological rather than peripheral mechanical causes.
Dell'Osso et al (1983) considered that LN arises when there is

confusion between egocentric direction referable to the 'cyclopean
eye' and oculocentric direction referable to the fixing eye when
changing from binocular to uniocular viewing. Lang (1982) drew
attention to the nasal retinal fixation preference exhibited by
lower mammals in whom the temporal visual field dominates, probably
through the extra-genicula-striate visual system. This phenomenon
may also be present as a primitive immature response in human neo-
nates; Lang suggested that LN results when this primitive system
persists. When both eyes are viewing, the tendency to shift fixat-
ion onto the nasal retina of each eye is cancelled out, but when
only one eye is taking part in the viewing process, fixation contin-
uously drifts onto the nasal retina giving the slow phase of the
nystagmus, and recovers to foveal fixation during the fast phase.
Kommerell and Mehdorn (1982) suggested that defective development of
the optokinetic control system could be the basic cause of both
congenital nystagmus (CN) and LN/MLN. Further understanding of the
mechanisms which cause LN and MLN may come from the study of those
patients who have either had one eye enucleated in early infancy, or
have had extremely poor vision in one eye from the same period.
There are several reports (Haase, 1971; Reinecke, 1984; Harcourt and
Spencer, 1985; Helveston et al, 1985) indicating that at least some
such patients exhibit a pattern of MLN, compensatory face turn
towards the side of the fixing eye and asymmetrical OKN responses
identical to that exhibited by patients with essential infantile
esotropia with nystagmus, even though the seeing eye appears
entirely healthy and there is no indication of any neurological
disorder. The inference from this would seem to be that primitive
oculomotor mechanisms may persist when there is not the possibility
of the visual processes of the 2 eyes becoming coordinated during
the critical first few months after birth.
Lang (1968) reported an incidence of abnormal head posture in

70%, and Harcourt and Mein (1982) one of 65% in comparable series
of patients with essential infantile esotropia with nystagmus. A
face turn in most cases compensates for the nystagmus, which
increases in intensity on abduction of the fixing eye. The reasons
for supposing that the face turn is not commonly due to limitation
of abduction of the fixing eye are that uniocular occlusion or
26 B. HARCOURT
spinning may demonstrate that abduction is potentially full, that

forced duction tests show no mechanical limitation of movements,
and that the intensity of the nystagmus and the face turn increase
in unison on increasing effort to see, for instance when reading
small print or the Snellen chart; this effort makes no difference to
the range of ocular movements. The face turn also commonly persists
after strabismus surgery unless the nystagmus intensity has itself
been reduced by that procedure.
A head tilt is more difficult to understand; it may be another

distinctive aspect of the neurological defect in these patients.
There is little to support the view of Crone (1954) that it compen-
sates for the cyclodeviation present in patients with DVD. The
direction of tilt often changes depending on which eye fixates, but
not always towards the side of the fixing eye.
Dissociated vertical divergence (DVD) is a characteristic

feature of essential infantile esotropia with nystagmus but it does
not develop before the age of some 18 months. A slow and curious
elevation and extorsion of one or other eye occurs intermittently.
The elevation occurs both spontaneously and when the amount of light
entering the affected eye is reduced, for instance during a cover
test, or sometimes when the eye is adducted and the bridge of the
nose may interfere with vision. There is often an associated A
pattern or V pattern of ocular movements, and DVD must be differ-
entiated from inferior oblique overaction. DVD is essentially the
same in degree regardless of the horizontal gaze position of the
eye, whereas inferior oblique overaction produces a hyperdeviation
which increases when gaze is directed towards the opposite side
(contralateral version). Inferior oblique overaction is never
associated with an A pattern of ocular movements.
The cause of DVD is not really understood at all although its

characteristics and associations have been thoroughly described
(Helveston, 1980). The area of greatest disagreement is whether
DVD develops and persists in any patients who exhibit a good quality
of binocular single vision (BSV). Mein and Johnson (1981) reported
very poor and unstable binocular responses in a series of 100
patients with DVD and this was confirmed by Mein and Harcourt
(1986b). If patients with established DVD all have very poor BSV,
then this primitive abnormal ocular movement could be caused by a
lack of the stabilising effect of fusion on the eye position, but
also by some separate associated brain-stem anomaly. If some
patients with good quality BSV really do exhibit DVD, then its
cause is likely to be an anomalous motor input.
Patients with DVD also show an associated asymmetry of their

horizontal optokinetic responses (OKN) when each eye is tested
separately. When the stimulating stripes are moved in a temporal
to nasal direction, the responses are normal, but when the stripes
are moved in the opposite direction there are only random eye
movements or no movements at all. The vertical responses are normal,
as are those when both eyes are uncovered. These anomalous res-
ponses are not found in patients who have essential infantile
esotropia without DVD and they are not related to the presence or
absence of binocular vision (t1ein, 1983; Flynn et al, 1984).
Although the clinical tests used and their interpretation have

varied very widely, there is general agreement that however soon
after initial examination, and however adequately, surgical correct-
ion of essential infantile esotropia with nystagmus is carried out,
the prognosis for the development and retention into adult life of
even a good quality of anomalous binocualr vision is poor. This
strongly suggests that there may be a 'congenital defect of the
fusion faculty' in these patients as suggested by Worth (1903),
accounting for their strabismus and its attendant anomalies of
ocular motility (nystagmus and DVD). The fact that other patients
denied the use of a fusion faculty by having one eye blind or
enucleated from infancy similarly develop MLN and DVD adds weight
to this idea.
b) without nystagmus:
The most characteristic pattern of ocular motility in these

patients is a V esotropia with elevation of each eye in adduction.
This has been explained by Gobin (1968) as being due to abnormal
anatomical alignment of the inferior oblique muscles. He avers that
when these muscles lie in a more antero-posterior axis (sagittal-
isation) they are less active extorters of the eyes, leading to
incyclotropia counteracted by inferior oblique contraction and
inhibition of the intortors (superior obliques and superior recti)
leading to horizontal and vertical muscle imbalance and a V pattern.
The anatomical anomaly is thus considered to cause abnormal inner-
vational responses which are the direct instigators of the V pattern
esotropia. Similarly, an A pattern esotropia can arise from
sagittalisation of the superior oblique tendon insertions.
NYSTAGMUS BLOCKAGE SYNDROME
This disorder is characterised by a very variable esotropia,

often unilateral and with a common incidence of amblyopia. At times
there appears to be simultaneous convergence of both eyes. Jerky
and essentially horizontal nystagmus increases proportionately in
intensity the further that the fixing eye moves from an adducted
position. The fast phase of the nystagmus is in the direction of
gaze of the fixing eye. Convergence inhibits (blocks) the nystagmus
partially or completely. Abduction may be limited even after pro-
longed contralateral occlusion, but the forced duction test is
negative. A compensatory face turn towards the side of the fixing
eye allowing fixation in the preferred adducted position is an
almost universal finding. The angle of strabismus is unchanged by
correction of any associated hypermetropic refractive error, or by
additional convex lenses, and the AC/A ratio is normal, The intra-
28 B. HARCOURT
duction of a base-out prism in front of the fixing eye fails to
induce abduction of the squinting eye. The strabismus is often
absent under general anaesthesia.
There are clearly some similarities between nystagmus blockage

syndrome and essential infantile esotropia with nystagmus. The fact
that the reported incidence of nystagmus blockage within undiffer-
entiated series of early onset esotropia is so variable (76% by
Muhlendyck, 1976; 12% by von Noorden and Avilla, 1984) suggests
that there is either overlap or confusion. The main differentiating
features are as set out in Table I (adapted from Mein and Harcourt,
(1986a).
Hoyt (1982) reported that vestibule-ocular reflexes (VOR) were

normal in infants with nystagmus blockage syndrome but were
moderately or severely defective in those with congenital esotropia
(essential infantile esotropia); he suggested that this is an
additional aid in differential diagnosis.
TABLE I
Essential infantile Nystagmus blockage

esotropia with nystagmus syndrome
Angle of deviation large and stable very variable
DVD common rare
Amblyopia rare common
Results of predictable unpredictable

strabismus surgery
EARLY ONSET ACCOMMODATIVE ESOTROPIA
Although commonly associated with a somewhat later age of

onset, non-paralytic esotropia with an accommodative element
certainly does arise in some children below the age of 6 months.
Baker & Parks (1980) described a group of 21 such patients with a
mean age of 4~ months at the time of onset of strabismus. They
showed characteristics associated with convergence excess in that
the angle of strabismus was greater for near than for distance, or
was present only on near fixation, and was reduced or abolished by
convex lenses or miotics. There was a high incidence of hyper-
metropia and the AC/A ratio was also high in some 50% of cases.
However, binocular vision was subnormal; half the patients reviewed
were at risk of amblyopia and the esotropia recurred after treatment
in almost half the cases. The outcome was therefore much inferior
to that which would be expected in a similar group of patients with
accommodative convergent strabismus of later onset. Whether this is
due to the immaturity of the binocular fixation reflexes at this
earlier age, or whether the conditions described are in fact essent-
ially different cannot at present be stated with any certainty.
CONGENITAL SIXTH CRANIAL NERVE PALSIES
There are undoubtedly a few patients who are otherwise healthy

and who exhibit permanent congenital abducent nerve palsies with
persistent incomitance. If the palsy is unilateral, the limitation
of abduction of the affected eye with a face turn towards that side
is characteristic and must not be confused with unilateral Duane's
syndrome. Bilateral symmetrical palsy is more difficult to diagnose
with certainty, but there is bilateral limitation of abduction
without nystagmus. Lateral rectus palsy is more common in young
infants with associated neurological defects, especially hydroceph-
alus. It may then rarely be congenital and developmental, but it is
more commonly seen in infancy arising as the result of suddenly
raised intracranial pressure from blockage of a Spitz-Halter valve,
when it is usually bilateral, though often asymmetrical.
Of more doubtful incidence and significance are neonatal sixth

nerve palsies which recover but give rise to a persistent, permanent
and virtually concomitant esotropia. Although this has been a
popular explanation for much infantile esotropia in the past, there
is little evidence for its occurrence in prospective series of
healthy children examined in the neonatal period. Nixon et al
(1985) examined 1219 normal neonates within the first 3 days of
life. They noted only 3 patients with signs of sixth nerve palsy,
and the 2 who could be followed up reverted to normal ocular
motility without strabismus within 3 months. Hoyt (1977) suggested
that nystagmus blockage syndrome was common in children with hydro-
cephalus, as the result of sixth nerve palsies, but this has not
been confirmed by other authors.
DUANE'S SYNDROME
Although there are undoubtedly mechanical aspects to Duane's

syndrome especially in older patients, as demonstrated by the
limitation of passive movement on forced duction testing, there is
at least indirect evidence of a neurological origin. Not only do
electromyographic (EMG) recordings show evidence of bizarre patterns
of innervational activity in the ocular muscles, but anomalies of
the sixth cranial nerves have also been described. Hotchkiss et al
(1976) reported on autopsy of specimens from 2 affected patients in
30 B. HARCOURT
which the sixth nerve was absent and the abducent nucleus absent or
rudimentary. The. third cranial nerve innervated the only part of
the lateral rectus muscle which was normally formed; the rest of the
muscle was poorly innervated and fibrotic. Huber (1984) reported on
the intraoperative findings in a patient with Duane's syndrome who
was undergoing surgery for removal of an acoustic neuroma. The
sixth cranial nerve on the side affected by Duane's syndrome
comprised only 2 thin atrophic strands which emerged separately
from the dura. Malformation or acquired abnormality of the
abducent nerve supply to the lateral rectus with consequent mis-
direction of third cranial nerve fibres to that muscle is now
considered as a very likely cause of Duane's syndrome, the mechan-
ical element being a secondary phenomenon. Further indirect
evidence of a central neurological aetiology comes [rom the work of
Jay and Hoyt (1980) who reported abnormal brain-stem auditory-
evoked potentials in some patients with Duane's syndrome. A very
early embryonic origin during the second month of embryonic life is
suggested by the association of Duane's syndrome with other abnor-
malities known to be the result of abnormal embryonic processes
dating from that time, particularly thalidomide embryopathy,
Goldenhar's syndrome, Klippel-Feil syndrome and spinal meningiocele.
A summary of the possible causes of the convergent strabismus

in these 6 different types of infantile esotropia is set out in
Table II.
TABLE II
POSSIBLE CAUSES OF ESOTROPIA IN INFANCY
Inherent defects in binocular fusion potential (essential infantile

esotropia with nystagmus)
Anomalous insertion of oblique muscles, termed sagittalisation

(essential infantile esotropia without nystagmus)
Nystagmus (nystagmus blockage syndrome)
Anomalous accommodation/convergent relationships (early onset

accommodative esotropia)
Sixth cranial nerve palsy (persistent or recovered)
Anomalous innervation of rectus muscles (Duane's syndrome)
Mechanical restrictions of movement (Duane's syndrome)
These descriptions allow consideration of those areas which are

likely to be fruitful in the further investigation of the normal and
abnormal development of the mechanisms which control ocular
motility. The principal questions which need to be answered seem

to be:-
a) What are the exact subgroups of infantile esotropia?
Despite a number of recent attempts to set out the essential

features which differentiate the subgroups of the disorder, there is
still some confusion both in published work and in strabismologists'
personal practice. For instance, is there overlap between essential
infantile esotropia with nystagmus, and nystagmus blockage syndrome?
In particular, does DVD really occur in patients with nystagmus
blockage, and do any patients with DVD really develop and retain a
good quality of sensory fusion? Again, exactly what proportions of
patients with different types of infantile esotropia maintain what
quality of BSV after adequate surgery at what age? Is a division
into essential infantile esotropia with and without nystagmus really
justified nosologically? Does DVD and asymmetrical OKN really occur
in patients with no nystagmus, or is this thought to be the case
only because fine nystagmus passes unnoticed.
b) Are there demonstrable pathological anomalies of the oculo-

motor control systems in patients with infantile esotropia?
If there are identifiable neuropathological defects in these

patients, are there clinico-pathological correlates with the nature
of the consequent strabismus? In particular, is there recognisable
pathology in the regions of the brain controlling vestibula-ocular
and smooth pursuit (tracking) movements and optokinetic responses
and in the visual cortical association areas. We need to know
whether essential infantile esotropia with nystagmus can be the
final common effect of a number of different developmental or
acquired defects in brain function, or whether it is a specific
disorder. Is there any anatomical evidence that there might be a
developmental lack of fusion faculty in these patients? It certain-
ly is the case that associated neurological anomalies are common in
patients with essential infantile esotropia with nystagmus and those
with nystagmus blockage syndrome (von Noorden and Avilla, 1984;
Lang, 1968; Harcourt and Mein, 1982).
In Duane's syndrome, is there further confirmation of sixth

nerve malformation and third cranial nerve misinnervation of the
lateral rectus muscle?
c) What are the roles of abnormal OKN and VOR systems?
Are there indeed innate defects in the vestibulo-ocular and

smooth tracking eye movement systems in certain patients with
infantile esotropia, and if so exactly what clinical features are
associated with what defect?
d) When and in what way does strabismus arise in the first few
months of life?
32 B. HARCOURT
Nixon et al (1985) have shown that, at least up to the age of
3 months, persistent esotropia is extremely rare, certainly by
comparison with the generally accepted 1% incidence of infantile
esotropia within the population (Helveston, 1986). De Decker (1987)
is reaching a similar conclusion. The inference is that many more
babies begin to squint between the ages of 3 and 6 months, around
the stage at which normal ocular motor responses become firmly est-
ablished. If so, what are the earliest features of the developing
strabismus? In particular, which comes first, the strabismus or the
nystagmus; and do skew deviations occur as a preparatory feature in
a significant proportion of patients as suggested by Hoyt et al
(1980).
e) What is the long-term natural history of affected patients?
Several authors, but especially Hiles et al (1980) in very

long-term follow-up studies have noted late decompensation of
initially apparently well-controlled essential infantile esotropia.
Mein and Harcourt (1986b) have commented on the late deterioration
or loss of anomalous binocular single vision, associated with
recurrent esotropia or consecutive exotropia in a considerable
proportion of cases, despite apparently adequate early surgery.
Further large long-term prospective studies are required.
f) Are there anatomical defects in the structure and orientation

of the extraocular muscles?
Confirmation of Gobin's work on sagittalisation of the oblique

muscles is needed, together with an assessment as to whether, even
if this is found in certain patients with infantile esotropia, it is
really the sole significant determining factor. The mechanical
features in the early stages of Duane's and Mobius syndromes also
require further elucidation.
Firmer answers to these questions should .allow a more rational

approach to treatment and the possibility of formulating a more
accurate long-term prognosis for infantile esotropia at an early
stage. In particular, if there is an irremediable congenital defect
in fusion faculty (Worth, 1903), then surgery is cosmetic only and
can safely be postponed to an age at which the exact characteristics
of the strabismus can be assessed with greater accuracy and the
risks of the unrecognised development of amblyopia after surgical
abolition of cross-fixation can be decreased.
If the normal grounding of the binocular fixation reflexes is

an acquired process which can be disrupted in early childhood by
adverse factors as stated by Chavasse (1939), then attempts need to
be made to see whether even earlier surgery,(between the ages of 3
and 6 months, immediately the strabismus is first distinguished)
gives a better long-term prognosis by avoiding the pre-operative
development of anomalous binocular responses.
If accommodative esotropia of a type indistinguishable from
that occurring in older children really can affect young infants,
and has a potentially favourable prognosis, then more consistent
methods of assessment and treatment need to be laid down. In part-
icular, the diagnostic use of phospholine iodide and the immediate
and total correction of any hypermetropia, using extended wear soft
contact lenses if spectacles are not well tolerated, need to be
emphasised.
The possibilities are exciting, but the clinical and laboratory

research methods required are extremely exacting, particularly the
size of series required for acceptable levels of statistical signif-
icance, and the difficulties of making firm diagnostic judgements
from the clinical examination of ocular motility in very young
children. Long-term follow-up by the same examiner is also very
difficult, especially in contemporary health care systems. There
are great difficulties in realising the rare opportunities which do
arise for detailed neuropathological studies of this largely healthy
group of patients. It will require time and determination and the
continuing close cooperation of scientists and clinicians to achieve
worthwhile progress, and those possibilities should be enhanced by
this Conference.
REFERENCES
Baker, J.D. & Parks, M.M. (1980). Early onset accommodative eso-
tropia. American Journal of Ophthalmology, 2Q, 11-18.
Chavasse, F.B. (1939). Worth's Squint or The Binocular reflexes and

the Treatment of Strabismus. Seventh Edition. p 113. Bailliere,
Tindall & Cox, London.
Ciancia, A. (1962). La esotropia en la lactante, diagnostico y

tratamiento. Archives of Chilean Ophthalmology, ~. 117.
Crone, R.A. (1954). Alternating hyperphoria. British Journal of

Ophthalmology, 38, 591-604.
de Decker, W. (1987). Personal communication.
Dell'Osso, L.F., Traccis, S. & Abel, L.A. (1983). Strabismus, a

necessary condition for latent and manifest latent nystagmus.
Neuro-ophthalmology, }, 247-57.
Flynn, J.T., Pritchard, C. & Lasley, D. (1984). Binocular VlSlon

and OKN asymmetry in strabismus patients. In Strabismus II. 35-44.
(ed Reinecke R.D.). Grune & Stratton, New York.
Gobin M.H. (1968). Sagittalisation of the oblique muscles as a

possible cause for the A, V and X phenomena. British Journal of
34 B. HARCOURT
Haase, W. (1971). Zur operativen Therapie de Kopffehlantung

infolge Nystagmus bei Monophthalmus. Klinische Monatsblatter der
Augenheilkunde, ~. 35-41.
Harcourt, B. & Mein J. (1982). Early onset esotropia. In

Documenta Ophthalmologica Proceedings Series 32. 79-82. (eds Balen
A. Th.M. van & Houtman W.A.). Junk, The Hague.
Harcourt, B. & Spencer, F. (1985). Manifest latent nystagmus

affecting patients with uniocular congenital blindness. In
Transactions of the Fourteenth Meeting of the European Strabismol-
ogical Association. 259-64. (ed Gregerson E.). APS, Copenhagen.
Helveston, E.M. (1980). Dissociated vertical deviation: a clinical

and laboratory study. Transactions of the American Ophthalmological
Society, 78, 734-79.
Helveston, E.M. (1986). Origins of Congenital Esotropia. American

Orthoptic Journal, ~. 40-48.
Helveston, E.M., Pinchoff, B., Ellis, F.D. & Miller, K. (1985).

Unilateral esotropia after enucleation in infancy. American Journal
of Ophthalmology, ~. 96-99.
Hiles, D.A., Watson, B.A., & Biglan, A.W. (1980). Characteristics

of infantile esotropia following early bimedial rectus recession.
Archives of Ophthalmology, ~. 697-703.
Hotchkiss, M.G., Miller, N.R., Clark, A.W. & Green W.R. (1980).
Bilateral Duane's retraction syndrome. A clinico-pathologi~ case
report. Archives of Ophthalmology, ~' 870-4.
Hoyt, C.S. (1977). The nystagmus compensation syndrome.

(Correspondence). American Journal of Ophthalmology, 83, 423.
Hoyt, C.S. (1982). Abnormality of the vestibula-ocular response in

congenital esotropia. American Journal of Ophthalmology, 2l•
704-8.
Hoyt, C.S., Mousel, D.K. & Weber, A.A. (1980). Transient Supra-
nuclear disturbances of gaze in healthy neonates. American Journal
of Ophthalmology, 89, 708-13.
Huber, A. (1984). Duane's retraction syndrome. Considerations on

pathophysiology and aetiology. In Transactions of the Fifth Inter-
national Orthoptic Congress. 119-25. (eds Ravault, A.P. &
Lenk, M.). LIPS, Lyon.
Jay, W.M. & Hoyt, C.S. (1980). Abnormal brainstem auditory-evoked

potentials in Stilling-Turk-Duane retraction syndrome. American
Journal of Ophthalmology, 89, 814-18.
Kommerell, G. & Mehdorn, E. (1982). Is an optokinetic defect the

cause of congenital and latent nystagmus? In Functional Basis of
Ocular Motility Disorders. 159-67. (ed Lennerstrand G.) Pergamon
Press, London.
Lang, J. (1968). Squint dating from birth or with early onset. In

Transactions of the First International Congress of Orthoptists.
231-7. Kimpton, London.
Lang, J. (1982). A new hypothesis on latent nystagmus and on the

congenital squint syndrome. In Documenta Ophthalmologica Proceed-
ings Series 32. 83-6. (eds Balen A.Th.M. van & Houtman W.A.).
Junk, The Hague.
Mein, J. (1983). The OKN response in early onset strabismus.

Australian Orthoptic Journal, 20, 13-17.
Mein, J. & Harcourt, B. (1986a). Diagnosis and Management of

Ocular Motility Disorders, p 228. Blackwell Scientific Publications,
Oxford.
Mein, J. & Harcourt, B. (1986b). Diagnosis and Management of

Ocular Motility Disorders, p 234. Blackwell Scientific Publications,
Oxford.
Mein, J. & Johnson, F. (1981). Dissociated vertical divergence and

its association with nystagmus. In Orthoptics, Research and
Practice, 14-16. (eds Mein, J. & Moore, S.). Kimpton, London
Muhlendyck, H. (1976). Diagnosis of convergent strabismus with

nystagmus and its treatment with Cuppers' faden operation. In
Orthoptics, Past, Present and Future, 143-54. (eds Moore, S.,
Mein, J. & Sockbridge, L.). Symposia Specialists, Miami.
Nixon, R.B., Helveston, E.M., Miller, K., Archer, S.M. & Ellis, F.D.
(1985). Incidence of Strabismus in Neonates. American Journal of
Noorden, G.K. von (1984). Infantile esotropia, a continuing riddle.

American Orthoptic Journal, 34' 52-62.
Noorden, G.K. von & Avilla, C. (1984). Nystagmus blockage Syndrome:

revisited. In Strabismus II. 75-82. (ed Reinecke, R.D.). Grune
and Stratton, New York.
Reinecke, R.D. (1984). Nystagmus blockage syndrome in the uni-

laterally blind patient. Documenta Ophthalmologica, 58, 125-130.
Worth, C. (1903). Squint; Its Causes, Pathology and Treatment.

p 55. Bale and Danielsson, London.
3
MORPHOLOGY OF THE EXTRAOCULAR MUSCLES
IN RELATION TO THE CLINICAL
MANIFESTATION OF STRABISMUS
ROBERT F. SPENCER and KEITH W. McNEER
INTRODUCTION
Strabismus is a congenital or acquired disorder of ocular

motility that is characterized by misalignment of the eyes either
in the primary position and/or during conjugate movements in
specific directions. The etiology of acquired non-commitant or
commitant forms of strabismus can be traced in some instances to a
central lesion affecting either the motor nerve or areas of the
brainstem involving the extraocular motor nuclei or the premotor
structures and pathways related to oculomotor control (see papers
by Lennerstrand and Harcourt in this volume). Congenital forms of
strabismus, on the other hand, might be attributable to
developmental abnormalities that affect the extraocular muscles,
their motor innervation, or the central connections of the
oculomotor system (see paper by Baker in this volume).
Previous studies of the histopathology of human strabismic

extraocular muscles have revealed a censtellation of structural
abnormalities, most of which are rather non-specific (Martinez et
al. • 1976; Berard-Badier et al. • 1978; Spencer and McNeer, 1980;
Martinez et al •• 1980). The extent to which these changes are the
cause or effect to, or are unrelated to, the clinical manifestation
of strabismus is presently unclear. A major limitation of the
examination or human extraocular muscle is the surgical resection
procedure, which limits sampling to the distal portion of the
muscle and thus, in most cases, does not include an entire cross-
sectional profile of the muscle, particularly the orbital layer and
the region of motor innervation. Incieed, the experimental studies
of the postnatal development and surgical and pharmacological
denervation of the extraocular muscle~ that will be summarized
indicate that one fiber type in the orbital layer may be most
susceptible to innervational changes, the sequelae of which might
be manifested as certain forms of strabismus.
37
38 R.F. SPENCER and K.W. McNEER
Figure 1. Phase contrast light micrographs of muscle fiber types

in the orbital (A), intermediate (B), and global (C) layers of the
lateral rectus muscle in Rhesus monkey. 1,3-5 - singly-innervated
muscle fibers; 2,6 - multiply-innervated muscle fibers. Bar 25 ~·
NORMAL MORPHOLOGY OF ADULT EXTRAOCULAR MUSCLES
The four rectus and two oblique extraocular muscles in

mammals, including humans, have an intrinsic organization
characterized by an outer orbital layer and inner global layer
separated by an intermediate transition zone. The differential
distribution of at least six basic morphological types of muscle
fibers within these layers forms the basis for the concept that the
fibers in the orbital layer are slower and recruited first during
eye movement, whereas those in the global layer are faster and
recruited later (Scott and Collins, 1973; Barmack, 1978).
The orbital layer contains one type of singly-innervated and

one type of multiply-innervated muscle fiber, while the
intermediate and global layers contain three types of singly-
innervated fibers and another type of multiply-innervated fiber
(Asmussen et al., 1971; Mayr, 1971; Spencer and Porter, 1981;
Pachter, 1982, 1983; Pachter and Colbjornsen, 1983). The most
apparent difference between the singly-innervated fiber types is in
the size, number, and disposition of the mitochondria (Fig. 1) and
the development of the internal membrane system. The mitochondrial
differences correlate well with the different oxidative and/or
glycolytic enzyme activities of the various fiber types (Durston,
1974; Ringel et al., 1978), while the extent of development ofT-
tubules and sarcoplasmic reticulum probably relates to differences
in the speed of contraction. The extent to which the simultaneous
expression of at least six different myosin heavy chain (MHC) genes
at the mRNA level and the synthesis of at least four different MHC
proteins in extraocular muscle (Wieczorek et al., 1985) are related
to the six morphological types of muscle fibers is presently
MORPHOLOGY OF EXTRAOCULAR MUSCLES 39
unclear.
Variations in the capillary vascular network also are apparent

in the different regions of the muscle (Ringel et al., 1978). The
orbital layer exhibits the most extensive microvasculature that is
associated specifically with the singly-innervated fiber (Fig. 1).
This fiber type is characterized by prominent central and
subsarcolemmal aggregations of mitochondria in the end-plate
region, and a moderate internal membrane system that separates the
myofibrils. The muscle fiber is focally innervated by synaptic
endings that literally encircle the fiber, but that exhibit little
subjunctional folding. It is perhaps this feature of its
innervation that led to the earlier misinterpretation of this
"coarse" fiber as multiply-innervated (Durston, 1974; Ringel et
al., 1978), which subsequently was resolved by the absence of slow
myosin immunoreactivity (Pierobon-Bormioli et al., 1979, 1980).
The morphological characteristics of this fiber type thus suggest
that it is highly oxidative and, in all likelihood, the most
fatigue resistant of the singly-innervated muscle fibers.
PRENATAL AND POSTNATAL DEVELOPMENT OF EXTRAOCULAR MUSCLES
Subsequent to their embryological origin from premandibular

and maxillo-mandibular mesodermal condensations (Gilbert, 1957),
the anlage of human extraocular muscles proceed through six
prenatal developmental stages in the formation of muscle fibers
(Sevel, 1981). While the appearance of muscle fiber striations and
motor innervation, including polyneuronal innervation, occurs early
in the prenatal period (Martinez et al., 1977; Gamble et al.,
1978), the differentiation and maturation of the six muscle fiber
types occur later during gestation and extend well into the
postnatal period. At birth, the fate of the extraocular muscle
fibers as fast-twitch singly-innervated or slow multiply-
innervated, on the basis of myosin ATPase activity, appears to be
determined (Hanson et al., 1980). The expression of MHC mRNAs,
including fast-glycolytic and muscle-specific, that characterize
adult extraocular muscle is incomplete, however, and the
predominant MHC protein is a neonatal form (Wieczorek et al.,
1985).
The maturation of the muscle fibers into their distinct adult

fiber type characteristics occurs largely during the postnatal
period (Nag and Cheng, 1982), as does the diameter spectrum of
their axonal innervation (Kerns, 1980). The postnatal maturation of
the muscle fibers is manifested by increases in cross-sectional
size and the appearance of metabolic enzymes that are related to
the oxidative/glycolytic capacities of the fibers (Schonfelder et
al., 1977; Hanson et al., 1980). In particular, the orbital
singly-innervated fiber appears to be the last fiber type to attain
its adult features, as demonstrated by oxidative enzyme activity
(Hanson et al., 1980) correlated with mitochondrial content and
40 R .F. SPENCER and K.W. McNEER
Figure 2. Phase contrast light micrographs of muscle fibers in the

orbital layer of lateral rectus muscles in cats 1 (A), 29 (B), 56
(C), and 63 (D) days postnatal. Bar 10 ~·
microvascular supply (Fig. 2). Perhaps not coincidently, these

changes occur concommitant with the development of interocular
alignment (Sherman, 1972).
The postnatal development of the morphological, and

physiological (Lennerstrand and Hanson, 1978a,b), properties of the
extraocular muscles may be largely influenced by the activity of
the motoneurones that innervate them. The unique and complex
pattern of gene regulation in extraocular muscle suggests that a
high degree of plasticity at the level of MHC gene transcription
may be responsive, at least during development, to neural and/or
environmental (e.g., hormonal) factors that influence its
expression (Wieczorek et al., 1985). Like the sensory visual
system, the premotor systems (e.g., vestibule-ocular, optokinetic)
that regulate motoneurone behaviour also develop postnatally
(Flandrin et al., 1979). It is thus not surprising that the
consequences of alterations in visual experience during the early
postnatal period, either experimentally or genetically, include
changes in the morphological and physiological properties of the
extraocular muscles (Lennerstrand, 1979, 1980; Lennerstrand and
Hanson, 1979), in addition to the well documented changes in the
sensory visual system that result in amblyopia. Such conditions

affect primarily the oxidative enzymes and capillary network
(Lennerstrand, 1980), both of which are correlated with the
decreased fatigue resistance of the extraocular muscles
(Lennerstrand, 1979; Lennerstrand and Hanson, 1979) and are
especially associated with the singly-innervated fiber type in the
orbital layer.
SURGICAL DENERVATION OF EXTRAOCULAR MUSCLES
The effects of intracranial and intraorbital transection of

the IIIrd nerve have been studied in several species (Cheng-Minoda
et al., 1968; Drachman et al., 1969; Durston, 1974; Asmussen and
Kiessling, 1975; Ringel et al., 1978; Baker et al., 1982). In the
acute stages of denervation, changes in the extraocular muscles
include degeneration of the neuromuscular junctions, disruption of
myofibrillar organization, reorganization of myonuclei into chains,
and infiltration by mononuclear inflammatory cells. Atrophy and/or
hypertrophy of the muscle fibers is not consistent.
A major conclusion to be drawn from these studies is that

extraocular muscles respond to denervation in a unique fashion. In
contrast to skeletal muscle, major degeneration and loss of the
muscle fibers and fiber type grouping that typically accompany
regeneration do not occur in extraocular muscle. Individual fiber
architecture is only minimally disrupted and changes in the size
and distribution of muscle fiber types are minor. The one fiber
type that exhibits the most prominent alterations, however, is the
orbital singly-innervated fiber, which is characterized by
persistent hypertrophy and hyperplasia. Indeed, the contractile
properties of long-term denervated extraocular muscles may be
attributable to this fiber type (Asmussen and Gaunitz, 1981).
PHARMACOLOGICAL DENERVATION OF EXTRAOCULAR MUSCLES
Botulinum toxin causes blockade of neuromuscular transmission

in skeletal muscle by interference with the calcium-dependent
neurogenic quantal and spontaneous non-quantal release of
acetylcholine (Thesleff and Molgo, 1983; Thesleff, 1984). This
presynaptic action of the toxin induces denervation-like
alterations in the motor innervation of skeletal muscle fibers with
consequential changes in the physiological, histochemical, and
ultrastructural features of the muscle fibers. The toxin induces a
third type of neurotransmitter release that is insensitive to
transmembrane ionic fluxes or nerve terminal membrane
depolarization and that may have a trophic function on the muscle
fibers, thus limiting the severity of the postsynaptic changes to a
lesser extent than those produced by nerve transection.
Although the overall effect of botulinum toxin paralysis of

Figure 3. Phase contrast light micrographs of orbital singly-

innervated muscle f i bers in medial rec tus muscles of Rhesus monkeys
7 (A), 14 (B), 28 (C), and 56 (D) days after intramuscular
injection of botulinum toxin. Bar 25 ~ ·
neuromuscular transmission in extraocular muscles is temporary

disuse of all muscle fiber types, the orbital singly-innervated
muscle fiber and its associated microvasculature demonstrate the
most prof ound changes (Spencer and McNeer, 1986). In adult medial
rectus musc les, the most apparent change in the morphology of this
muscle fiber in the short term (1-3 weeks) is the dispersion of the
central aggregates of mitochondria toward the periphery with the
formation of massive subsarcolemmal accumulations that distort the
surface profiles of the fibers (Fig. 3) . The hypertrophy of these
muscle f ibers is accompanied by withdrawal of the capillary
vascular net work with which this fiber type is associated. In the
l ong term (6-8 weeks), with recovery of function, the morphology of
this muscle fiber and the density of the capillary network appear
normal. Although t he cross-sectional areas of the orbital singly-
innervated muscle fibers and the vasculature in the orbital layer
are less than normal, the ratio of the myofiber:vascular area is
the same as normal.
The f indi ngs are interpreted to indicate that the

morphological changes in the orbital singly-innervated muscle
fibers probably are secondary to the withdrawal of the capillary

microvascular supply upon which this fiber type is dependent for
oxidative metabolism. Other fiber types are less affected since
they rely more upon glycolytic pathways. Most significantly, they
demonstrate that an intimate relationship exists whereby the
capillary network adapts proportionately to the demand for
oxidative metabolism as a consequence of the change in the force
dynamics of the opposing muscles that, in the long term with return
of function, renders the toxin-paralyzed muscle fibers smaller, and
consequently weaker, than normal.
CONCLUSION
The orbital singly-innervated muscle fiber may have a

prominent role in ocular motility. On the basis of morphological
and histochemical features alone, this fiber type is highly
oxidative and fatigue resistant, both factors dependent upon its
extensive capillary network. It is thus ideally suited to maintain
fixation and probably is initially recruited in various types of
eye movements, but especially vergence. It is furthermore the last
of the six basic morphological fiber types to develop its adult
features and is most susceptible to alterations in its innervation,
both neural and vascular. The relative changes in length-tension
characteristics that are the hallmark of strabismus, therefore, may
be attributable predominantly to variations from normal in the
orbital singly-innervated muscle fiber and its associated
microvasculature under the influence of neural and/or environmental
factors.
ACKNOWLEDGEMENT
Supported by U.S. Public Health Service Research Grant EY02191

from the National Eye Institute.
REFERENCES
Asmussen, G., and Gaunitz, U. (1981). Changes in mechanical

properties of the inferior oblique muscle of the rabbit after
denervation. Pflugers Arch.,392, 198-205.
Asmussen, G. and Kiessling, A. (1975). Hypertrophy and atrophy of

mammalian extraocular muscle fibres following denervation.
Experientia,31, 1186-1187.
Asmussen, G., Kiessling, A., and Wohlrab, F. (1971).

Histochemical characteristics of muscle fibre types in the
mammalian extraocular muscles. Acta Anat.,79, 526-545.
Baker, R.S., Millett, A.J., Young, A.B. and Markesbery, W.R.

(1982). Effects of chronic denervation on the histology of canine

extraocular muscle. Invest. Ophthalmol. Vis. Sci.,22, 701-705.
Barmack, N.H. (1978). Laminar organization of the extraocular

muscles of the rabbit. Exp. Neurol.,59, 304-321.
Berard-Badier, M., Pellissier, J.F., Toga, M., Mouillac, N., and

Berard, P.V. (1978). Ultrastructural studies of extraocular
muscles in ocular motility disorders. II. Morphological analysis of
38 biopsies. Albrecht v. Graefes Arch. klin. exp. Ophthal.,208,
193-205.
Cheng-Minoda, K., Ozawa, T. and Breinin, G.M. (1968).

Ultrastructural changes in rabbit extraocular muscles after
oculomotor nerve section. Invest. Ophthalmol.,I, 599-616.
Drachman, D.A., Wetzel, N., Wasserman, M. and Naito, H. (1969).

Experimental denervation of ocular muscles. Arch. Neurol.,21, 170-
183.
Durston, J.H.J. (1974). Histochemistry of primate extraocular

muscles and the changes of denervation. Brit. J. Ophthalmol.,58,
193-216.
Flandrin, J.M., Courjon, J.H. and Jeannerod, M. (1979).

Development of vestibule-ocular response in the kitten. Neurosci.
Lett.,12, 295-299.
Gamble, H.J., Fenton, J., and Allsopp, G. (1978). Electron

microscope observations on human fetal striated muscle. J. Anat.
(Lond.),126, 567-589.
Gilbert, P.W. (1957). The or1g1n and development of the human

extrinsic ocular muscles. Contrib. Embryol.,36, 61-78.
Hanson, J.,. Lennerstrand, G. and Nichols, K.C. (1980). The

postnatal development of the inferior oblique muscle of the cat.
III. Fiber sizes and histochemical properties. Acta Physiol.
Scand.,108, 61-71.
Kerns, J. M. (1980). Postnatal differentiation of the rat

trochlear nerve. J. Comp. Neurol.,189, 291-306.
Lennerstrand, G. (1979). Contractile properties of extraocular

muscle in Siamese cat. Acta Ophthalmol. (Copenh.),57, 1030-1038.
Lennerstrand, G. (1980). Histochemical studies on the inferior

oblique muscle of Siamese cats and domestic cats with unilateral
lid suture. Exp. Eye Res.,30, 619-639.
Lennerstrand, G. and Hanson, J. (1978a). The postnatal

development of the inferior oblique muscle of the cat. I.
Isometric twitch and tetanic properties. Acta Physiol. Scand.,103,

132-143.
Lennerstrand, G. and Hanson, J. (1978b). The postnatal

development of the inferior oblique muscle of the cat. II.
Effects of repetitive stimulation on isometric tension responses.
Acta Physiol. Scand.,103, 144-153.
Lennerstrand, G., and Hanson, J. (1979). Contractile properties

of extraocular muscle in cats reared with monocular lid closure and
artificial squint. Acta Ophthalmol.,57, 591-599.
Martinez, A.J., Biglan, A.W., and Hiles, D.A. (1980). Structural

features of extraocular muscles of children with strabismus. Arch.
Ophthalmol.,98, 533-539.
Martinez, A.J., Hay, S. and McNeer, K.W. (1976). Extraocular

muscles. Light microscopy and ultrastructural features. Acta
Neuropathol. (Berl.),34, 237-253.
Martinez, A.J., McNeer, K.W., Hay, S.H. and Watson, A. (1977).

Extraocular muscles: morphogenetic study in humans. Light
microscopy and ultrastructural features. Acta Neuropathol.
(Berl.),38, 87-93.
Mayr, R. (1971). Structure and distribution of fibre types in the

external eye muscles of the rat. Tissue Cell,l, 433-462.
Nag, A.C. and Cheng, M. (1982). Differentiation of fibre types in

an extraocular muscle of the rat. J. Embryol. Exp. Morphol.,Zl,
171-191.
Pachter, B.R. (1982). Fiber composition of the superior rectus

extraocular muscle of the Rhesus monkey. J. Morphol.,174, 237-250.
Pachter, B.R. (1983). Rat extraocular muscle. I. Three

dimensional cytoarchitecture, component fibre populations and
innervation. J. Anat. (Lond.),137, 143-159.
Pachter, B.R. (1984). Rat extraocular muscle. 3. Histochemical

variability along the length of multiply-innervated fibers of the
orbital surface layer. Histochemistry,80, 535-538.
Pachter, B.R. and Colbjornsen, C. (1983). Rat extraocular muscle.

2. Histochemical fibre types. J. Anat. (Lond.),137, 161-170.
Pierobon-Bormioli, S.P., Sartore, S., Vitadello, M., and

Schiaffino, S. (1980). "Slow" myosins in vertebrate skeletal
muscle. An immunofluorescence study. J. Cell Biol.,85, 672-681.
Pierobon-Bormioli, S.P., Torresan, P., Sartore, S., Moschini, G.B.,

and Schiaffino, S. (1979). Immunohistochemical identification of
slow-tonic fibers in human extrinsic eye muscles. Invest. Ophthal.

Vis. Sci.,18, 303-306.
Ringel, S.P., Engel, W.K., Bender, A.N., Peters, N.D. and Yee, R.D.
(1978). Histochemistry and acetylcholine receptor distribution in
normal and denervated monkey extraocular muscles. Neurology
(Minn.),28, 55-63.
Schonfelder, J., Asmussen, G., and Schaaf, P. (1977). Zur

postnatalen ontogenese der ausseren Augenmuskeln des Kaninchens.
Verb. Anat. Ges.,Zl, 1187-1191.
Scott, A.B. and Collins, C.C. (1973). Division of labor in human

extraocular muscle. Arch. Ophthalmol.,90, 319-322.
Sevel, D. (1981). A reappraisal of the origin of human

extraocular muscles. Ophthalmol.,88, 1330-1338.
Sherman, S.M. (1972). Development of interocular alignment in

cats. Brain Res.,37, 187-203.
Spencer, R.F., and McNeer, K.W. (1980). Structural alterations in

overacting inferior oblique muscles. Arch. Ophthalmol.,98, 128-
133.
Spencer, R.F., and McNeer, K.W. (1986). Morphological basis of

junctional or muscle adaptation in pharmacological denervation of
monkey extraocular muscles. In Adaptive Processes in Visual and
Oculomotor Systems. (eds. E.L. Keller and D.S. Zee). Pergamon
Press, Oxford, pp. 13-20.
Spencer, R.F. and Porter, J.D. (1981). Innervation and structure

of extraocular muscles in the monkey in comparison to those of the
cat. J. Comp. Neurol.,198, 649-665.
Thesleff, S. (1984) Transmitter release in botulinum-poisoned

muscles. J. Physiol. (Paris),79, 192-195.
Thesleff, S., and Molgo, J. (1983) A new type of transmitter

release at the neuromuscular junction. Neuroscience,~, 1-9.
Wieczorek, D.F., Periasamy, M., Butler-Browne, G.S., Whalen, R.G.,

and Nadal-Ginard, B. (1985). Co-expression of multiple myosin
heavy chain genes, in addition to a tissue-specific one, in
extraocular musculature. J. Cell Biol.,lQl, 618-629.
4
MOTOR AND SENSORY FUNCTIONS OF NORMAL
AND STRABISMIC EXTRAOCULAR MUSCLE
GUNNAR LENNERSTRAND
Extraocular muscle (EOM) show functional

properties that are different in many respects from
the ordinary skeletal, striated muscle. However, we
know very little about how EOM properties and dys-
function might be involved in concomitant strabis-
mus or how EOM is affected in this disease. Most of
the studies on EOM have been performed in animals,
but some information is also available from humans.
The first part of this paper deals with the EOM
muscle fibers and motor units: the electrophysio-
logy and the mechanical properties, the recruitment
pattern in different types of eye movements and the
effects of denervation. The changes in the func-
tional properties of EOM during development is
another important aspect of eye muscle function,
which is very relevant in the discussion of stra-
bismus and its causes. The second part of the paper
on muscle receptors and their role in ocular motil-
ity and strabismus is intended to supplement the
more extensive coverage of EOM proprioceptive
effects on visual development and spatial percep-
tion that will be presented later on during this
conference (see papers by Maffei, Steinbach and
Campos).
MOTOR FUNCTION OF EOM

Neurophysiology of motor units
The morphological composition of EOM differs
appreciably from that of ordinary skeletal muscle.
At least five different fiber types have been
identified in eye muscles of different mammals
including man (Chiarandini and Davidowitz, 1979;
Morgan & Proske, 1984). The morphological
47
48 G.LENNERSTRAND
charateristics, innervation, histochemistry and
distribution within the eye muscle of the different
fiber types have just been described by Spencer
(this symposium). Also in lower vertebrates several
distinct fiber types have been reported (Morgan and
Proske, 1984; Lennerstrand and Baker, 1987).
In the physiological sense EOM is unique in
that it combines high contractile speed and rather
extreme fatigue resistance during continous acti-
vation (Fuchs and Binder, 1983). Both properties
are usually much more developed than in ordinary
skeletal muscle of the same species. Fatigue re-
sistance of EOM is higher in species with good bi-
nocular vision such as the cat, than in species
with poor binocular vision (rat, guinea pig, fish)
(Lennerstrand, 1982; Lennerstrand and Baker, 1987).
In the cat the fatigue resistance is higher in
animals with normal binocular vision than in cats
with congenitally abnormal binocularity (Siamese
cats) or acquired defects (Lennerstrand, 1982).
An EOM motor unit consist of the motoneuron in
the brain stem and the muscle fibers it innervates
in the muscle itself. All fibers in a unit are pre-
sumably of the same type (Burke, 1981; Gueritaud et
al, 1985). In the EOM a motoneuron innervates only
6-10 muscle fibers on an average while the inner-
vation ratio may be up to 1:1000 for large hind
limb muscles. Studies of EOM motor units have been
performed mainly in the cat, by recording the
mechanical and electrical responses in the muscle
to stimulation of single motoneurons or their axons
(Lennerstrand, 1975, Nelson et al, 1986). It has
been possible to differentiate 2 or 3 types of
motor units. The units with singly innervated
fibres (SI units) have rapid contractions, among
the fastest in the body. Their fatigue resistance
ranges from very low to intermediate values. The
fiber type that Spencer (this symposium) has sug-
gested to be involved in strabismus, i.e. the or-
bital, singly innervated fiber, is probably the
slowest and most fatigue resistant among the SI
units. The units with multiply innervated, non-
conduction fibers (MINC units) contract very slowly
and are extremely fatigue resistant. Properties of
units with multiple innervation but conducted
action potentials (MIC units) are intermediate to
SI and MINC units with regard to speed of contrac-
tion and fatigue resistance. Thus, it seems likely
that the rapid contractions of EOM are due to SI
unit properties and the high endurance to MIC and
MINC unit activation.
EOM MOTOR AND SENSORY FUNCTION 49
In the cat the distribution of motor units,

identified as fast twitch units (probably both SI
and MIC units) or slow non-twitch units (probably
MINC units), is the same in the medial rectus (MR)
and lateral rectus muscles (LR), with over 90% of
the units being fast and less than 10 % slow
(Meredith and Goldberg, 1986).
However, on an average the fast units of the MR
were found to contract faster than LR units, but
the maximal tension output of each unit was some-
what lower in MR than in LR units. The difference
between MR and LR units in tension output is
probably compensated by a higher rate of discharge,
on an average, of MR motoneurons than of LR neurons
(Delgado-Garcia et al, 1986a). This in turn may be
due to premotor influences, since it is known that
the internuclear neurons of the abducens nucleus
are driving the MR unit at higher discharge rates
than that seen in the LR motoneurons (Delgado-
Garcia et al 1986b). It is not known if similar
differences occur in other antagonistic pairs of
EOM in the cat, or if they exist in other species.
Physiological studies of this kind should also be
done in monkeys, with natural strabismus, discove-
red by Kiorpes et al. (1985). In humans EMG of EOM,
preferably with techniques for single fiber record-
ing, could be used in studies of electrophysio-
logical properties of muscle fibers of normal and
strabismic muscles. It has been possible to
determine muscle fiber electrical properties in
extracellular recordings of rabbit and cat EOM
(Chiarandini and Davidowitz, 1979).
Development of eye muscle function

The morphological aspects have already been
described by Spencer (this symposium).
In physiological studies of EOM development in
the cat, it has been shown that speed of contrac-
tion and fatigue resistance was low at birth but
that they increase with age (Lennerstrand, 1982).
Speed of contraction continued to rise until adult
age, while fatigue resistance reached a plateau
at about 6 weeks of age. The latter stage corre-
sponds in time to the phase of rapid development of
binocular and monocular visual functions into the
cat. It was suggested that the time course of
development of fatigue properties might reflect the
need for fibers and motor units with high endurance
to be used in fixation and binocular vision. It was
also found that Siamese cats with a congenital
fusion defect and domestic cats with defects of
binocular vision induced by squint or monocular
so G.LENNERSTRAND
deprivation from an early age, showed reduced EOM
speed of contraction and fatique resistance in
comparison with normal, binocular cats. The
structural correlates to the binocular defects
seemed to be a reduction in the over all fibers
size and a reduction of the number of capillaries
in the muscles. However, there were no changes in
the histochemical properties of the fibers,
in the distribution pattern or in the relative pro-
portions of the different (histochemical) fiber
types.
These studies demonstrate that changes in the
visual input can modify oculomotor behaviour even
at the most peripheral part of the plant, i.e. at
the level of eye muscle function. The structural
changes resemble most closely those seen in disuse
of muscle and not what is observed after partial
denervation. It should be of interest to examine if
similar changes in EOM structure and function occur
also in humans with defects of binouclar function,
and if they can be reversed with proper treatment
of strabismus and improvement of binocular vision.
Recruitment of EOM motor units

Two mechanisms exist whereby the force of a
muscle may be regulated: the recruitment of
individual members of the muscle population of
motor units, and the modulation of discharge
frequency of motor units that already have been
recruited. In EOM the muscle force appears to be
regulated mainly by the latter mechanism, i.e. fre-
quency modulation. The frequency range of firing is
extensive (up to 600 imp/s in monkey EOM mota-
neurons). However, recruitment of slow units is
also an important factor, particularly for con-
trolling the very precise eye motor acts which
probably employ fatige resistant units with finely
graded tension output.
Studies on the spinal cord have shown that the
neurons within a motoneuron pool are recruited
according to size, with the smallest neurons being
activated first. In most reflex actions and other
stereotyped inputs, the motor units are recruited
in the following order: slow units - fast fatigue
resistant units fast fatiguing units (Burke,
1981). Nonstereo-typed, 'faculative' movements
might well utilize other neural organizations to
bypass the constraints inherent in a hardwired
network. However, the load on the eye muscles is
predictable and non-changeable, except under some
pathological conditions, and most eye movement
patterns would therefore be suitable for fixed
innervational programs.
Recent work by Nelson and collaborators (1986)
on trochlear motoneurons has shown that the size of
the EOM motoneuron is fairly well correlated with
the speed of contraction and force production of
its muscle unit, in the same way as in the limb
muscles. This suggests that the recruitment order
in EOM follows the size principle outlined for limb
muscles. It is likely that slow EOM motor units are
activated before fast units, but this recruitment
order may not be as strict as in limb muscles,
since there is evidence that some of the slow
units are innervated by large axons and motoneurons
(Lennerstrand, 1975), and they would be recruited
late during muscle activation. Possibly they supply
the steady contraction with low fatigue necessary
for maintenance of eye deviation in lateral gaze.
Support for the view that slow and fast units can
be recruited separately are supplied in the
experiments on human EOM done by Scott and Collins
(1973). It is therefore tempting to propose that
the slow units with characteristics of amphibian
and avian slow-tonic fibers, are used predominantly
for fixations in different positions of gaze and
for slow eye movements like vergences, while the
fast units are used mainly for faster eye movements
like persuit and saccades, although we know that
there is no selective activation of any type of
unit for a specific type of eye movement in the
monkey or the cat (Delgado-Garcia et al 1986a).
Spencer (this symposium) has suggested a role also
for the orbital, singly innervated fibers in
control of slow eye movements and fixation.
With respect to frequency modulation of motor
unit discharge, variations between EOM motor units
have been demonstrated in the relation between
firing rate and eye position (so called k-value)-
(Delgado-Garcia et al, 1986a). Generally the units
and motoneurons with low threshold of excitation
(mostly the slow units) have lower k-value than the
high-threshold units that usually contract faster.
Most low threshold units had lower maximal dis-
charge frequency than fast units, which probably is
related to the fact that the slow units reach fused
tensions at lower rates of activation than the fast
units. It would thus seem as if recruitment and
frequency modulation both play important parts in
the production of EOM force dynamics in the various
types of eye movements, but the relative contri-
bution of each of them for the different types of
motor units and eye movements have still to be
established. The possibilities of recruitment
disorders in different types of strabismus has not
52 G.LENNERSTRAND
yet been explored.
EOM properties in denervation and reinnervation

There is a renewed interest in the effects of
denervation and reinnervation on EOM structure and
function after the introduction of some new methods
of strabismus treatment. They include (i) botulinum
toxin injections in the EOM (Scott, 1981), (ii)
surgical neurectomy which is used to correct large
overaction of the inferior oblique muscle (Del
Monte and Parks, 1984), and (iii) attempts to
reinnervate a palsied EOM with a transplant from
another eye muscle (Aichmair, 1977).
The morphological changes that occur in the
denervated and reinnervated EOM muscle have already
been described by Spencer (this volume). During the
acute phase of denervation when the muscle shows
hypertrophy, the stiffness of a normal EOM is
increased quite markedly, in denervation either
from cutting the muscle nerve (Asmussen and
Gaunitz, 1981) or pharmacologically by injection of
botulinum toxin (King et al, 1986). However, the
effect of botulinum denervation in EOM with
contracture, e.g. an antagonist to a paralysed EOM,
is to reduce muscle stiffness (King et al, 1986).
Recovery of mechanical properties during
reinnervation has not been sufficiently studied in
EOM, but from work on hind limb muscles it is known
that contractile properties are restored succes-
sively when the muscle is reinnervated by its own
nerve (Bagust and Lewis, 1974). Reinnervation of
limb muscles through a different nerve can change
the contractile properties of the muscle fibers and
make a slowly contracting muscle to become a fast
one, and a fast muscle to contract slowly. The
changes also involve the histochemical properties
of the muscle fibers, i.e. the type of myosin they
contain (Buller and Pope, 1977). Since all EOM
consist of a mixture of slow and fast fibers,
equivalent experiments may be hard to perform, but
it is conceivable, although not very likely
(Spencer, this symposium) that muscle fiber
properties can be altered by reinnervation.In that
case one would have to postulate a different time
course for the reinnervation of fast and slow motor
units.
Experiments on the superior oblique muscle in
the goldfish have shown that this EOM can become
dually reinnervated over both the trochlear nerve
and the branch of the oculomotor nerve innervating
the antagonist, the inferior oblique musle. Thus,
the inappropriate synapses were not repressed
(Scott, 1977). A mixture of innervation is seen in

the humans with the congenital Duane type II
anomaly, where the lateral rectus muscle is
innervated from both the abducens and the oculo-
motor nerve (Huber, 1974b). Whether this can occur
in other forms of infantile strabismus has not been
sufficiently studied.
Polyneuronal innervation of individual muscle
fibers, which is seen in hind limb muscles of
newborn rats and kittens (Brown et al, 1976). The
excess nerve fibers retracted within 2 weeks of
birth, leaving each muscle fiber innervated by one
nerve fiber. It is not known if the same mechanisms
are active in EOM, but matters might be more
complicated here since the fibers are supplied with
multiple endplates in the adult stage. In fish EOM
polyneuronal innervation is the rule (Morgan and
Proske, 1984), but such innervation does not occur
in adult cat EOM (Bach-y-Rita, 1975), although the
evidence is not conclusive for the MINC units. If
similar postnatal changes from polyneuronal to
mononeuronal innervation occurs in EOM, malfunction
in neuron retraction could affect EOM function in
the infant and might be involved in early onset
strabismus.
EMG of EOM disorders in man

EMG studies have been of great assistance in
establishing the pathophysiology of different types
of motility disorders, mainly the myogenic types of
paralytic strabismus (see Huber 1974a also for
references). EMG has also been a decisive method in
determining the pathophysiology of a developmental
disorder (Duanes syndrome) which involves aplasia
of the abducens nucleus and nerve, and innervation
of the lateral rectus by branches of the oculomotor
nerve (Huber, 1974b). During EOM reinnervation,
single fiber EMG has demonstrated an increase of
fiber density indicating that the motor units
become larger and/or that grouping of fibers of a
particular type occurs (Huber and Schiller, 1982).
This is the characteristic fiber pattern in
reinnervation of other skeletal muscle, but it is
not known to what extend and in what types of
fibers it might occur in the EOM. EMG has also
proved useful in studies of EOM motor unit recruit-
ment as mentioned in section C.
With regard to etiology of concomitant
strabismus EMG studies with the present techniques
have been rather uninformative. Recently EMG has
been used in studies of possible role of EOM
proprioception in esotropia and exotropia (Mitsui
54 G.LENNERSTRAND
and Tamura, 1986), and this will be discussed in a
subsequent section of this paper. It is not clear
to what extent EOM disorders such as myopathy or
innervational disturbances contribute to con-
comitant strabismus, but the changes in EMG must be
rather subtle and more sophisticated electrophysio-
logical techniques would have to be developed in
order to demonstrate them. It would also be im-
portant to establish the motor unit firing patterns
in the different types of eye movements, both in
the normal and the strabismic EOM. Thus, EMG and
particularly the techniques of recording from
single motor units in the EOM, would seem the most
important method for functional studies of periphe-
ral motor control in the human oculomotor system,
and should be used more extensively in strabismus
research than has been done previously.
SENSORY FUNCTION OF EOM

Structure and functional properties of muscle
receptors in EOM
The muscle afferent signals derive from muscle
spindles with an intricate machinery of sensory
endings and specialized motor control, and from
tendon organs with a simpler structure and response
pattern. Functionally, the tendon organs respond to
muscle tension changes, since they are coupled in
series with the muscle fibres, while the spindles
react to changes in muscle length, due to their
position in parallel with the ordinary muscle
fibers. The major part of the information on
structure and function of muscle spindles and
tendon organs derive from studies on skeletal
muscle of the cat, but there is ample proof that
the conditions are very similarly in other species.
Structurally the spindles consist of a
collection of intrafusal fibers of three kinds;
nuclear bag 1 and 2 fibers and nuclear chain fibers
(Matthews 1972, 1981). Sensory endings are either
of the primary type with receptor sites on both
nuclear bag and chain fibers, or of the secondary
type with connection only to nuclear chain fibers.
Motor control is exerted over the fusimotor fibers
system with nerve fibers in the gamma or beta
diameter range. Each motor fiber innervates one of
the three types of intrafusal fibers. Human eye
muscle spindles have the same general character-
istics (Manni and Bartolami, 1982). The fusimotor
fibers are called dynamic or static, depending on
how stimulation of them change the responses of the
primary and secondary endings to muscle stretch.

From the studies of hind limb muscle spindles it is
known that the primary endings have a prominent
sensitivity to the rate of muscle stretch. The
sensitivity is further enhanced by stimulation of
the dynamic fusimotor fibres, but reduced by static
fusimotor activation (Lennerstrand, 1968). The
static fibers increase the static responses of the
spindle endings and can also increase the sensitiv-
ity to changes in muscle position. Secondary end-
ings have a much lower velocity sensitivity than
primary endings and are not influenced by dynamic
but only by static fusimotor fibers (Matthews,
1981). The difference in dynamic sensitivity
between primary and secondary endings is also
manifested in the responsiveness to vibration of
the muscle, which is much higher in primary than in
secondary endings (Matthews, 1972). This means of
differentiating the two groups of endings was used
in a study of muscle spindle responses in pig EOM
(Bach-y-Rita, 1975). In this study as well as in a
study of sheep EOM spindles (Browne, 1975) evidence
was found for primary and secondary endings and
dynamic and static fusimotor control in muscle
spindles of EOM. Thus, in animals with muscle
spindles in their EOM, the receptor machinary would
seem to function in very much the same way as in
limb muscles. This probably holds also for the
tendon organs in pig EOM, which respond in a manner
similar to those in limb muscles (Bach-y Rita,
1975).
Distribution of muscle spindles in EOM varies
with species. Human EOM contains a large number of
spindles, and the density per muscle weight is
actually the highest in the body (Cooper et al,
1955). Large numbers of spindles were also found in
EOM of pig, sheep, cattle and a few other animals
(Cooper et al, 1955). In the macaque and the baboon
the number of spindles is rather small. No typical
spindles have been observed in common laboratory
animals such as the cat, rat, mouse, rabbit, dog or
squirrel monkey (Manni and Bartolami, 1982). How-
ever, these is physiological evidence for stretch
receptors arranged in parallel with the ordinary
muscle fibers in e.g. cat and squirrel monkey
(Bach-y-Rita, 1971). When spindles are found, they
are mostly situated in the orbital zone of the
muscle and in the distal part (closest to the
tendon) (Cooper et al, 1955, Harker, 1972).
Tendon organs of different types (seldom
typical Golgi tendon organs) have been seen in the
EOM of almost all species (Ruskell, 1978). In man
the tendon endings have the shape of 'palisade
56 G.LENNERSTRAND
endings' (Steinbach, this symposium), and such

endings are also seen in cat EOM, where they are
situated on the extensions of multiply innervated
fibers in the orbital layer (Alvarado-Mallart and
Pincon-Raymond, 1979).
Afferent pathways and central actions of EOM

proprioceptors
In limb muscles the spindle primary endings
provide information on muscle position and rate of
length change, secondary endings information on
muscle position and Golgi tendon organs on muscle
force sensed at the tendon. The velocity sensitiv-
ity of spindle endings can be varied over the
dynamic fusimotor system and the position sensivity
over the static fusimotor system (Matthews, 1981).
The information is fed back to the motoneurons of
the same muscle as well as to those of synergistic
and antagonistic muscle, but it also reaches the
brain stem cerebellar complex. It is used in the
continuous control of limb movements, particularly
in the compensation time for unexpected changes in
load on the muscles, and in the adjustment of
individual muscle participation in complex move-
ments (Matthews, 1972).
Recently it has been shown that signals from
muscle spindles reach the cerebral cortex and that
they probably are involved in the conscious aware-
ness of the position of the limbs and the trunk. It
has been suggested that muscle spindles in the limb
muscles are more important than the joint receptors
in this respect (Matthews, 1982). In contrast our
knowledge on the function of the EOM receptors is
very limited, and this discussion will be re-
stricted to stretch reflexes and similar effects.
In most species including man the afferent
signals from EOM muscle receptors travel to the
brain stem through the trigeminal complex (Manni
and Bartolami, 1982; Porter and Spencer, 1982). The
cell bodies of the first order neurons are in the
semilunar ganglion and the second order neuron
somatas in the descending part of the spinal trige-
minal nucleus. From this structure projections have
been found to many different parts of the CNS:
superior colliculus, vestibular nuclei, pontine
nuclei, central gray matter dorsal to the third
nerve nucleus, cerebellum, and even to the frontal
cortex and the visual cortex (Bach-y-Rita, 1975;
Manni and Bartolami, 1982). However, there are very
few reports on any direct connections between
muscle receptor afferents and EOM motoneurons, and
the existence of an EOM stretch reflex is much
debated (Manni and Bartolami, 1982). Occasional

observations of excitatory or inhibitory responses
in cat and rabbit EOM have been reported but not
been verified in cat or monkey.
Activation of the primary endings in limb
muscles by vibration at rates above 200 Hz induces
a slowly increasing excitation of the homonymous
muscle and the contraction is well sustained
(Matthews, 1981). Stretches superimposed upon the
contracted muscle are counteracted by increased
muscle contraction. Actually this is the most
important way of demonstrating the stretch reflex
in human muscle under normal conditions. In cat
EOM with no spindles, Barbas and Dubrowsky (1981)
demonstrated an excitatory effect of vibration that
was produced very slowly with a latency of several
seconds. In the sheep, an animal with EOM spindles,
stretch or vibration induced an inhibition of EMG
and muscle tension in the same muscle (Petorossi
and Filippi, 1981). These actions have a short
latency of some 10 msec. Similar experiments could
be done in humans and the EOM efferent output
monitored with recordings of force and EMG.
Mitsui and Tamura (1986) have reported stretch
reflexes in EOM of humans with strabismus. Their
studies were based on the original finding by
Mitsui that small adductive forces applied to the
dominant eye in patients with manifest exotropia,
caused a corrective movement of the other eye,
which eventually reached an orthophoric position
(but never went beyond that point). The response
was of all-or-nothing character and it was ac-
companied by EMG changes (reduction of LR activity
and increase of MR activity) in the muscles of the
moving, non-dominant eye. The movement could be
elicited only in light, but was not considered a
convergence movement, induced by retinal slip or
some other type of visual stimulation of the
adducted, dominant eye. The phenomenon could be
produced also under general anesthesia if the drug
prifinium bromide had been administered beforehand,
in order to keep the muscles electrically active
under the anesthesia. The time to reach full effect
was rather long in the awake state (several sec-
onds) and even longer (minutes) during anesthesia.
The effect was abolished by retrobulbar anesthesia
of the manipulated eye. In esotropia it was found
that passive abduction could elicit a reduction of
EMG activity in the ipsilateral MR indicating
orthophorization of the manipulated eye. The effect
could be induced under general anesthesia, but only
under photopic conditions.
58 G.LENNERSTRAND
Since the effects of moving one eye, the

dominant one in exotropia and the deviated eye in
esotropia, depended upon light, Mitsui and Tamura
suggested that pathological optomotor reflexes
(Keiner, 1951) could be the basis for the mis-
alignement in strabismus. They further suggested
that exotropia was due to an exagerated contraction
of the LR of the non-dominant eye, caused by ab-
normal proprioceptive effects from the dominant
eye. In esotropia the proprioceptive influence is
from the deviated eye. Normalisation was considered
to take place when the appropriate muscles (LR in
exotropia, MR in esotropia) were streched. This was
thought to induce reflex actions and inhibition of
the contralateral LR in the case of exotropia and
of the ipsilateral MR in the case of esotropia.
However, this interpretation can be questioned
in several respects, as has already been pointed
out by Kommerell (1982). With regard to the stretch
reflex explanation one would argue that (i) the
anatomical correlate for direct connection are
missing, (ii) the time lag is too extensive for a
reflex pathway presumably involving only a few
synapses, and (iii) the EOM spindles would seem to
inhibit the homonymous and synergistic muscles wich
is contrary to the effects at the spinal and other
levels. The pathways for these long-latency reflex
actions have yet to be determined, although an
effect over the vestibular complex has been sug-
gested (Ashton et al, 1984).
The idea of abnormal proprioceptive reflexes
in strabismus lead Mitsui and Tamura to operate for
exodeviations on the dominant eye instead of on the
deviated eye. However, their rather excellent re-
sults of such operations (Mitsui et al, 1980) has
not been replicated, at least not in strabismus
surgery on children with exodeviations
(Lennerstrand 1986).
It should be noted that proprioceptive
reflexes are less well developed in neck muscles
than at other spinal levels. Possibly the afferents
in EOM and neck muscles are more important for the
global coordination of these muscles in eye-hand
interaction than for the control of the muscles of
origin. The possible role of EOM afferent signals
in eye position and movement control and the
effects on visual development under normal condi-
tions and in strabismus will be extensively covered
in a later part of this symposium by Maffei,
Steinbach and Campos.
SUMMARY
The aim of the review has been to supply
material for further discussion on extraocular
muscle (EOM) dysfunction as a cause or a conse-
quence of strabismus. In general there is much less
data on humans than on animals and the majority of
the animal findings are from normal, non-strabismic
muscle. The following areas were covered: (i) Motor
unit properties, unit distribution in EOM and the
recruitment in eye movements; (ii) peri- and post-
natal development of eye muscle function and the
influence of defects of binocular vision in ani-
mals; (iii) properties of EOM after denervation and
reinnervation; (iv) EMG of human EOM under normal
conditions and in neuromuscular disease. Very
little has been reported on EMG of strabismic
muscle; (v) muscle receptors of EOM, the types and
their occurance within different species. The func-
tional properties of spindles and tendon organs;
(vi) the central connections of EOM receptor
afferents, their reflex actions in EOM and the
possible role in strabismus.
ACKNOWLEDGEMENTS
The research reported from the authors:
laboratory has been supported by grants from the
Swedish Medical Research Council (No. 4751) and the
Karolinska Institute.
REFERENCES
Aichmair. H. (1977). Muscular neurotisation in
surgery of traumatic abducens paralysis. Jap. J.
Ophthalmol., ~. 477-487.
Alvarado-Mallart. R-M. and Pincon-Raymond. M.
(1979). The palisade endings of cat extraocular
muscles; a light and electron microscope study.
Tissue & Cell, !1. 567-584.
Ashton, J.A., Boddy, A. and Donaldsson, I.M.L.
(1984). Input from proprioceptors in the extrinsic
ocular muscles to the vestibular nuclei in the
giant toad, Bufo marinus.Exp.Brain Res. ,53,409-419
Asmussen, G. and Gaunitz, U. ( 1981 ) . Changes in
mechanical properties of the inferior oblique
muscle of the rabbit after denervation. Pfltigers
Arch., 392, 198-205.
60 G.LENNERSTRAND
Bach-y-Rita, P. (1971). Neurophysiology of eye

movements. In The control of eye movements.
(Eds. P. Bach-y-Rita, C.C. Sollins and J.Hyde). Pp
7-45, Academic Press, New York.
Bach-y-Rita, P. (1975). Structural-functional
correlation in eye muscle fibers. Eye muscle
proprioception. In Basic Mechanisms of Ocular
Motility and their Clinical Implications. (Eds. G.
Lennerstrand and P. Bach-y-Rita). Pp 91-109.
Pergamon Press, Oxford.
Bagust, J. and Lewis, D.M. (1974). Isometric con-
tractions of motor units in self-reinnervated fast
and slow twitch muscles of the cat. J. Physiol.,
237, 91-102.
Barbas, H. and Dubrovsky, B. (1981). Control and
peripheral effects of tonic vibratory stimuli to
dorsal neck and extraocular muscles in the cat.
Exp. Neurol., 74, 67-85.
Brown, M.C., Jansen, J.K.S. and van Essen,D. (1976)
Polyneuronal innervation of skeletal muscle in new-
born rats and its elimination during maturation. J.
Physiol., 261, 387-422.
Browne, J. (1975). The responses of muscle spindles
in sheep extraocular muscles. J. Physiol., 251,
483-496.
Buller, A.J. and Pope, P. (1977). Plasticity in
mammalian skeletal muscle. Phil. Trans. R. Soc.
Lond. B., 278, 295-305.
Burke, R.E. (1981). Motor units; anatomy, physio-
logy, and functional organization. In Handbook of
Physiology. Section I: The nervous system, Volume
II.Motor Control, part 1. (Eds. J.M. Brookhart and
V.B. Mountcastle). Pp 345-422, Williams & Wilkins,
Baltimore.
Chiarandini, D.J. and Davidowitz, J. (1979).
Structure and functions of extraocular muscle
fibres. In Current topics in eye research. Vol. 1.
Pp 114-120. Academic Press, New York.
Cooper, S., Daniel, P.M. and Whitteridge, D.
(1955). Muscle spindles and other sensory endings
in the extrinsic eye muscles; the physiology and
anatomy of these receptors and of their connections
with the brain stem. Brain, 78, 564-583.
Delgado-Garcia, J.M., del Pozo,,F. and Eaker, R.

(1986a). Behaviour of neurons in the abducens
muscles of the alert cat - I. Motoneurons. Neuro-
science, 12. 929-952.
Delgado-Garcia, J.M., del Pozo, F. and Baker, R.
(1986b). Behaviour of neurons in the abducens
nucleus of the alert cat II. Internuclear
neurons. Neuroscience, 12. 953-976.
Del Monte, M.A. and Parks, M.M. (1984). A compara-
tive study of denervation and extirpation for
marked inferior oblique overaction. In Strabismus
II. (Ed. R.D. Reinecke). Pp 567-578. Grune &
Stratton, Orlando.
Fuchs, A.F. and Binder, M.C. (1983). Fatigue
resistance of human extraocular muscles. J. Neuro-
physiol., 12. 28-34.
Gu~ritaud, J.P., Horchelle-Bossavit, G., Jami, L.,
Thiessen, D. and Tyc-Dumont, S. (1985). Resistance
to glycogen depletion of motor units in the cat
rectus lateralis muscle. Exp.BrainRes. ,£Q,542-550.
Harker, D.W. (1972). The structure and innervation
of sheep superior rectus and levator palpebrea
extra ocular muscles. II. Muscle spindles. Invest.
Ophthalmol., 11. 970-979.
Huber, A. (1974a). Elektromyographie der Augen-
muskeln. Ophthalmologica, 169, 111-126.
Huber, A. (1974b). Electrophysiology of the retrac-
tion syndromes. Brit. J. Ophthalmol., 58, 293-300.
Huber, A. and Schiller, H.H. (1982). Single fibre
EMG for investigation of extraocular eye muscles.
Electromyogr. Clin. Neurophysiol., 22, 303-314.
Keiner, G.B. (1951). New Veiwpoints on the Origin
of Squint. Martinus Nijhoff, The Hague.
King, W.M., Yeaple, R.N. and Metz, H. (1986).
Orbital stiffness changes induced by injection of
botulinum toxin into Simian extraocular muscles. In
Ada tive Processes in Visual and Oculomotor
Systems. eds. E.L. Keller and D.S. Zee). Pp 21-25.
Pergamon Press, Oxford.
Kiorpes, L., Boothe, R.G., Carlson, M.R. and Alfi,
D. (1985). Frequency of naturally occurring stra-
bismus in monkeys. J Ped. Ophthalmol. & Strabism.,
22, 60-64.
62 G. LENNERSTRAND
Lennerstrand, G. (1968). Dynamic Analysis of Muscle

Spindles in the Cat. Thesis, Karolinska Institute,
Stockholm.
Lennerstrand, G.(1975). Motor units in eye muscles.
In Basic Mechanisms of Ocular Motilit and their
Clinical Implications. Eds. G. Lennerstrand and P.
Bach-y-Rita. Pp 119-143. Pergamon Press, Oxford.
Lennerstrand, G. (1982). Postnatal development of
eye muscle function. In Functional Basis of Ocular
Motility Disorders. (Eds. G. Lennerstrand, D.S. Zee
and E.L. Keller). Pp 39-47. Pergamon Press, Oxford.
Lennerstrand, G. (1986). Effects of surgery on the
dominant eye in exodeviations. Acta Ophthalmol.,
64, 391-396.
Lennerstrand, G. and Baker, R. (1987). Motoneuronal
innervation and mechanical properties of extra-
ocular muscles in the catfish. Ictalurus Punctatus.
Acta Physiol. Scand., (in press).
Manni, E. and Bartolami, R. (1982). Proprioception
in eye muscles. In Functional Basis of Ocular
Motility Disorders. (Eds. G. Lennerstrand, D.S. Zee
and E.L.Keller). Pp 55-64. Pergamon Press, Oxford.
Matthews, P.B.C. (1972). Mammalian Muscle Receptors
and their Central Actions. Monographs of the
Physiological Society, No.23, Edward Arnold,London.
Matthews, P.B.C. (1981). Evolving views on the
internal operation and functional role of the
muscle spindle. J. Physiol., 320, 1-30.
Matthews, P.B.C. (1982). Where does Sherrington's
"muscular sense" orginate? Muscular, joints,
corollary discharges? Ann.Rev.Neurosci. ,Q,189-218.
Meredith, M.A. and Goldberg, J.S. (1986).
Contractile differences between muscle units in the
medial rectus and lateral rectus muscles in the
cat. J. Neurophysiol., Q£, no. 1.
Mitsui,Y., Tamura, D., Hirai, K-I., Akazawa, K.,
Ohga, M. and Masuda, K. (1980). Effect of master
eye surgery in exodeviations. Jap. J. Ophthalmol.,
24, 221-231.
Mitsui, Y. and Tamura, 0. (1986). Strabismus and
the Sensorimotor Reflex. Clinical Practice Series,
no. 30. Excerpta Medica, Amsterdam.
Morgan, D.L. and Proske, U. (1984). Vertebrate slow

muscle: Its structure, pattern of innervation, and
mechanical properties. Physiol.Reviews. ,64,103-169.
Nelson, J.S., Goldberg, S.S. and McClung, J.R.
(1986). Motoneuron electrophysiological and muscle
contractile properties of superior oblique motor
units in cat. J. Neurophysiol., QQ. 715-726.
Peterossi, V.E. and Filippi, G.M. (1981). Muscle
spindle autogenic inhibition in the extraocular
muscles of the lamb. Arch.Ital.Biol. ,119, 179-194.
Porter, J.D. and Spencer, R.F. (1982). Localization
and morphology of cat extraocular muscle afferent
neurons identified by retrograde transport of
horseradish peroxidase. J.Comp.Neurol. ,204, 56-64.
Ruskell, G.L. (1978). The fine structure of inner-
vated myotendenious cylinders in extraocular
muscles of rhesus monkeys. J. Neurocytol.,
708.
z,693-
Schiaffino, S., Pierobon Bormoli, S. and Aloisi, M.

(1974). Neural and non-neural control of muscle
differentation. In Exploratory Concepts in Muscle
Dystrophy. Internat. Congr. Serie no. 333.
(Ed. A.T. Milhorat). Excerpta Medica, Elsevier,
Amsterdam.
Scott A.B. and Collins, C.C. (1973). Division of

labor in human extraocular muscle. Arch. Ophthalmol
2.Q' 319-322.
Scott, A.B. (1981). Botulinum toxin injection of

eye muscles to correct strabismus. Tr. Am.
Ophthalmol. Soc., 79, 734-770.
Scott, S.A. (1977). Maintained function of foreign
and appropriate functions on reinnervated goldfish
extraocular muscles. J. Physiol., 268, 87-109.
5
KINETICS OF THE EYE
HARLEY E. A. BICAS
ORIGIN OF THE FORCES

An adequate understanding of the clinical problems concerning
eye positions and movemer.ts and, as a consequence, the way they are
to be treated, depends on how perfectly a theoretical model of the
ocular mechanics adjusts to them.In fact, any stable ocular position
may be defined as a consequence of balanced forces,while a movement
results from unequally acting torques upon the eye. But although re
duction of the oculomotor system to its physical manifestations may
be structurally understood, most of the functional correlations
among the forces and their origins are still to be discovered.
Energy for the system comes from the extraocular muscles (in-
ternal forces) and is expressed by a state of"tonicity".This repre-
sents the consecutive functions of a central command for the gener-
ation of a signal, of its neural transmission and of the muscular
response to it. Variations of the neuromuscular function (innerv-
ational impulses and, or tonic factors) are directly related to the
amount of the resulting ("active") muscular force, a relative in-
crease of it being named contraction and the reverse, relaxation.As
the ocular insertion of the muscle may move relative to its other
extremity (the fixed orbital insertion), this active force, or
tension, causes an ocular rotation, with its consequent and inverse
variation of the muscular length (isotonic contraction or relax-
ation). In fact, it is possible experimentally to obtain the varia-
tion of the muscular length between a known (initial) and an expected
(final) ocular position and the respective developed force for such
a rotation. Therefore, the resultant length-tension curve of the
muscle is important for the determination of how a stimulus (hence
innervation and then force) is related to a rotation (response).
Meaningful maps of length-tension curves of each muscle at least
for some (cardinal, diagnostic) of all possible ocular rotational di
rections are then needed to complement the data already known of th;-
horizontal recti at the horizontal plane (Robinson et al., 1969).
However, the relationship ~et~een innervation and ocular ro-
65
66 H. E. A. BICAS
tation is not closed: muscular tension may occur without change of

the muscular length (isometric contraction, or relaxation)Jherefor~
though active forces are usualy considered as propulsive (causing
movements), they can be resistive (preventing them) as in Duane's
syndrome. Furthermore, different active forces may occur for the same
muscular length. For instance, in abduction, at the end of a pursuit
movement, the activity of the medial rectus is low but may suddenly
reach higher values to initiate a saccadic movement in the opposite
direction. The concept of an "operational envelope" relating
innervation, muscular tension and eyeposition (Collins et al., 1975)
arises.
On the other hand, owing to its intrinsic elastic properties,

the muscle may have a passive shortening (i.e.,with no innervational
input) causing an ocular rotation. In this case (passive) forces are
directly proportional to the muscular length: a stretched muscle has
always a greater elastic tension (potential action) than when it is
shortened. Active and passive forces of a muscle pull the eye in the
same direction. However, while muscular activity(as measured by EMG)
increases as the ocular rotation progresses in the direction of the
muscular action (say, in a following pursuit movement), the elastic
pull decreases, and vice-versa.
Passive forces represent energetic economy: a movement, once

initiated, is self-controlled by the opposite and progressive (pote~
tial) forces of the stretched structures, which tend to pull the eye
back to the previous balanced position of "rest", when the original
(kinetic) muscular activity subsides. The system, however, is not
conservative. This is, again, very convenient: dissipation of forces
(both, active and passive) by friction, deformation of periocular
structures and translation, accounts for stability of the eye and
reestablishes conditions for starting another movement of fixation.
In summary, one energetic, agonis-tic, "active" input (volition

al or reflex) to an ocular rotation is partially spent to initiate the
movement (i.e., to overcome the starting friction), and maintain it
while heating the orbit (moving friction and deformation of
structures) and stretching elastic, antagonistic structures. Part of
it that is potentially accumulated (passive forces) is kinetically
recovered (a "spring-back" rotation), while the other part is also
dissipated (correct~on of previous deformations and more heat). A
simplified scheme to understand the variation of the oculomotor
balance of forces between a given ocular position (say ia) and the
primary position (say i 0 ) , may be proposed, according to there-
spective variables of a movement:
a) active (propulsive) forces of the agonist muscle (i = ia- i 0 ) ;
b) active (resistive) forces of the antagonist muscles (i'= i6 - ic);
c) passive (propulsive) forces of the agonist muscle (m = m0 - rna);
d) passive (resistive) forces of the antagonist muscles m'= -Cmc-m6);
e) passive (propulsive) forces of the ipsilateral (to the agonist)
periocular structures (p =Po- Pa);
f) id., ib. (resistive) of the contralateral side p' = -Cpc- p6);
KINETICS OF THE EYE 67
g) dissipation during the movement (s);

h) force due to the innervational stimulus, or resultant of the
active forces (I= i + i');
i) resultant of the passive muscular forces (J = rn + rn');
j) periocular elasticity (V = p + p');
k) resultant of the total passive forces (S = V + J);
1) resultant (propulsive) force of the agonist muscle (G = i + rn);
rn) resultant (resistive) force of the antagonist muscles (H=i'+ rn');
n) resultant of the total muscular forces (M = G + H I+ J);
o) total balance of forces (K= I+S = M+V) (hence S-V = M-I=J);
p) effective forces for the movement (T = K- s).
By definition, in the state of a perfect oculornotorbalance, the

passive forces self cancel (Si = p 0 + Ph+ rn 0 + mh = 0), which means
that at the beginning of a rotation from that (primary) position,
total (Ki) and active (Ii) forces can be taken as synonymous. But at
the end of the rotation, passive forces (Sf = -pa - Pc - rna - me)
counterbalance tonic activity (If = -Sf, hence Kf = 0) and if there
occurs a sudden annulment of tonus (If= 0), a reverse ocular ro-
tation is initiated by the potential energy related to the force Sf.
In other words, the energy for that spring-back rotation depends on
the previous balance of forces immediately before the beginning of
the movement, that is, on the value of Sf. This value is related to
the ocular position, i.e., the amount of stretching of the periocular
structures. Therefore it is not surprising that the angular velocity
with which a spring-back rotation is initiated is proportional to the
angular measurement of the position of the eye (peak velocities
greater than 400°/s may be reached from eye displacements of about
50° from the primary position).
Unfortunately, the knowledge of all previously listed variables

is not practical in each particular case. At best, simpler data have
to be taken and compared to those of a general solution of the model.
For instance, passive forces are ideally studied under deep general
anaesthesia (directly measured when balancing external forces during
the real "passive duction test" or indirectly by the spring back test.
when the balance of ocular positions and the angular velocity of the
spring-back rotations are considered). But in this condition, for
which no practical muscular tonus (internal forces) are present,
active forces can not be obtained. Even so, a complete isolation of the
passive forces of different origins is not feasible. On the other
hand, if the patient is awake, passive and active forces are super-
imposed, making analysis difficult. That happens when one measures
the increment of force directly (by preventing an ocular rotation
with external forces, the test of "generated forces"), or through its
indirect effect (the peak velocity) by electro-oculography.
Finally, the resistance to stretch gives direct values for the

elasticity of the studied muscle. This is important, since surgical
procedures alter both the elastic pull and the muscular responsiveness
to an innervational impulse: a resection shortens the muscle to obtain
its relative stretching in each gaze position. In that new length-
68 H. E. A. BICAS
-tension state, the passively stretched muscle rotates the eye, op-
poses to the action of the antagonist muscle (becomes more resistive)
and, since it is shortened, absorbs less energy to further con-
tractions, that is, reacts more promptly to innervation (becomes more
propulsive). A recession slacks the muscle, with reverse effects.
Pragmatical results may be then oriented by the model. For

instance, if one knows both the balance of forces (K), given by the
angle of deviation of the eyes of the awake patient, and the resultant
of the total passive forces (S), given by the angle of deviation under
deep general anaesthesia, the value of I is revealed. Then, if S = 0
(parallelism under anaesthesia) the basic deviation (K) is said to
be purely tonic (K =I). If S = K (no angular variation under anaes-
thesia), then the deviation is said to be purely anatomic or re-
strictive (I = 0). If K > S > 0, the squint is classified as being
caused by anatomic (S) and tonic (I = K - S) components. Obviously
it is expected that the proper correction of the deviation be of the
total anatomic defect (U 1 = S) and of a part (100 f%) of the tonic
component (U 2 = I. f, where f < 1, since the return of the tonus in
the awake patient has a multiplying effect 1/f. In fact, it is common
ly acceptr>d that, for the same surgical procedure of muscula"i="
weakening, "the greater the overaction, the greater the r:orrective
effect''). This can lead to the anticipation of the final angle which
has to be obtained at the end of the surgery (S - U 1 - U 2 ; Z). But,
asS= U1 , then Z = -U 2 = f.I, i.e., the angle of balance at which
the eyes have to be at the end of the surgery (Z), still during deep
general anar>sthesia, should be given by the value of the tonic com-
ponent (1) properly corrected by a factor (f).
Arbitrarily, a general value of f = 0.5 has been used for such

calculations, with empirical satisfactory results on the majority of
cases. However, more refined workup is necessary, not only to deter-
mine I (since part of its value may be masked by restrictive compo-
nents) but also the best fit of f for each case.
DIRECTIONS OF MUSCULAR ACTIONS

At each specific ocular direction of fixation, the effective ro-
tational components of the eye (vertical, torsional and horizontal)
given by each muscle, are determined according to the position of the
so-called plane of muscular action. Such a plane is defined as passing
through the center of ocular rotation and as containing the direction
of the force, i.e., the straight line between the point of its appli
cation and the point to where it is directed (muscular origin). How=
ever, the theoretical determination of this plane and of the com-
ponents of rotational actions of an extra-ocular muscle present some
difficulties owing to:
1) Proper conditions of the ocular mechanics. The eye is not a

rigid body (any force applied to it causes a slight but definite de-
formation), it is not perfectly spherical (especially in high
ametropias) or, even if it were spherical, it is not mechanically

symmetric (i.e., its center of mass does not coincide with its geo-
metric center). Furthermore the ocular inertia is not constant (the
mass of the eye) but varies with the friction (which, in turn, de-
pends on the velocity of a movement: starting sliding friction is
always greater than moving sliding friction). Finally, the resultant
of reaction forces of the periocular (orbital) tissues is not homo-
geneous.
2) Real translation. Even if one considers ideal mechanical con

ditions for the eye (a spherical rigid body, with its geometric ce~
ter at its center of mass, or its center of rotation, and sur=
rounded by an isomorphic medium), any isolated force applied to it,
will have a translational componen~i.e., evoke an ocular dis-
placement, in the direction of the force, along one orbital (fixed)
axis. Then, a force tangential to the eye, as that originated by a
muscular contraction or relaxation, is not purely rotational. The
only condition in which a pure rotation may occur is that of a
"conjugate system" (two simultaneous forces, with the same torque:
equal magnitudes and opposite directions, with diametrically op-
posed points of application). This is close to what happens in a
common eye rotation, by contraction of a muscle and equivalent re-
laxation of its antagonist but, in fact, all necessary conditions
of a conjugate system of forces do not occur. Therefore it is not
surprising that a fixed center of ocular rotation can not be found
(Park & Park, 1g33-)-.-
3) Eccentricity of the eye and of the ocular rotation. Because

the center of the eye is about 1.3 mm in front of the center of ro-
tation and 1.6 mm to its temporal side, even if a pure (mechanical)
rotation were possible, an ocular, "false", translation would have
to be considered. One of the consequences is that the traditional
concepts of an arc of (anatomic) contact between the muscle and the
scleral surface, as the condition of application of a tangential
force, are challenged. Firstly, the radius for the rotation Ci) is
not the radius of the eye (r) and depending on the point where the
force is applied, d > r, d ~ r or d < r (figure 1). At the primary
position, according to the original data (anatomic insertions) of
Volkmann (186g), d varies from ll,g mm (MR) to 14.2 mm (LR). The ef
fective angle of application of the force (a, considered relatively
to d, not to r) departs from goo, being smaller or greater according
to the direction of the force, which depends on the set of muscular
coordin~tes. Theoret~calJy, then, the torque (M) of a muscular
force (F), given by M = F d sin a, may increase even after the con-
tact is lost by the increase of sin a, or decrease even if the con-
tact is maintained (figure 1). On the other hand, a cushion under
the anatomic insertion of the muscle could also increase this angle
(a) up to goo and so maximize the muscular rotational torque.
As the structural and static view of a tangential contact

between muscle and eye has no practical meaning, the corresponding
measurement of this arc of contact is of small value. Better, a dy-
70 H. E. A. BICAS
(a l (b) (c l
FIGUP.E 1: Effects of rotations upon the (eccentric) eye. R is the

center of ocular rotation (no real translation of this
point is considered). E 1 , E2 , E 3 are the sucessive positions of the
center of the eye. On the left side of each graph, A1, A2 and A3
are the correspondent positions of the ocular anatomic insertion of
a muscle whose origin is 0 1 • On the right side, B 1 , B 2 and B 3 are
the positions of the ocular anatomic insertion of another muscle,
whose origin is 0 2 . For the muscle on the left side: (a) at the
initial position the physiologic insertion is P 1 , the arc of con-
tact is A1 P 1 (related to y 1 ) , the angle of application of the force
is a 1 and the radius for the rotc>.t-ion is RP 1 (the lever-arm is RL 1 ) ;
(b) after a rotation E the arc of contact becomes zero but although
the a.ngle of application of the force is the same (a 1 ) the radius
of rotation is greater than before (RA 2 > RP 1 ), so also is the
lever-arm (RL 2 > RL 1 ); (c) if the force is kept constant (isotonic
contraction) its torque further increases beyond the point where the
tangency is lost (A 2 ) because, now, though RA 3 = RA 2 it is the angle
of application of the force which increases (up to 900). The arc of
rotation (8) is A 1RA 3 . For the muscle on the right side, although
it r~mains in true contact with the eye from the beginning (B 1Q 1) to
the end of the rotation (8), the radius of rotation decreases (from
RQ 1 to RB 3 ), as well as the tangential component of the force
(a' 2 > a 2 > 90), so also its lever-arm.
namic "arc of rotation" must be considered (fig. 1 c). Note that the
relationship between the angle of the arc of rotation and its re-
spective length varies as function of d and therefore is not constant
for all muscles.
Consequently, apparent ambiguities may arise: the greatest arc

of rotation is that of the IO for angular measurements, i.e., eye
movements (79.34° against 72.850 of the LR) or of the LR for linear
values, i.e., variation of the muscular length (18.04 mm against
17.43 mm of the IO). An arc of rotation is usually much greater than
the respective arc of anatomic contact for the majority of the mus-
cles (respectively 79.34° and 49.26°, i.e., 161% for the IO, or
18.04 mrn and 9.85 mm, i.e., 183% for the LR) but depending on the set
of coordinates (especially for the cases of the MR and the SO), the
arc rotation may be smaller.
The "false" translation caused by eccentricity eventually super

imposes on the real translation (due to the application of forces).
The importance of the resultant ocular translation is not of its mag
nitude, which is relatively small compared with the normalrotationil
movements, but of the limiting effect on the rotation itself. By
translation, the eye is compressed against surrounding fat and tis-
sues, magnifying friction (Bicas, 1985) in addition to stretching
structures that restrict the rotation. Effects of translation are
also probably related to surgical results: an isolated recession of
a rectus muscle evokes an eye protrusion (which tends to stretch the
weakened wuscle) and an isolated resection evokes an eye intrusion
(which slacks the muscle), i.e., those techniques have contradictory
elements to diminish their results, mainly when compared to the
greater and more stable angular corrections of a combined recession-
-resection procedure.
4) Localization of the point of application of the force. The

plane in which the real point of application of the force (physi-
ologic insertion) lay, is not necessarily the same as that which
passes through the origin and the anatomic insertion, unless the mus
cle slides freely over the scleral surface. But intermuscular septa;
and check-ligaments prevent the muscle from occupying the shortest
path between its origin and ocular insertion. Then, it becomes poss!
ble that even a non-muscular surgical procedure may modify a mus-
cular action (figure 2). On the other hand, the integrity of such
structures supports the idea of mutual (physical) influences of
different muscles (i.e., one muscle may modify the plane of action of
another by simple stretching of intermuscular membranes between them).
However, the extent to which those fascial expansions of the muscle
sheaths influence the direction of the applied force has yet to be
determined.
5) Definition of coordinates. The muscle has not a single point

of application of force, but a broad (either the anatomic, or the
physiologic) insertion, nor a point to where it is directed, but a
conspicuous volume of muscular origin, i.e., of its orbital fixation.
Moreover, the coordinates of such points are stillnot determinable
in each particular case. The commonly used general values of
Volkmann (1869) in which analyses of ocular rotations have been based
(Krewson, 1950; Boeder, 1961) do not satisfy even the equation of a
spherical eye, but show a radius of ocular curvature varying from
12.77 mm (SO coordinates) to 11.35 mrn (LR coordinates), i.e., a
difference of 12.5%.
6) Variations during rotations. The central point of the or~g~n

and the midpoint of the ocular insertion are simplifications which can
not be valid even if the muscle adapts to the shortest possible path
for every gaze situation. That is because the effective insertion
72 H. E. A. BICAS
(a) (bl
FIGURE 2: Muscular action before (a) and after (b) section (CC') of
an intermuscular membrane. The direction of rotation changes
because of the displacement of the point where the force is applied
(from P 1 to P 2 ). R is the center of rotation (for the sake of simpli
fication made coincident with the center of the eye), 0 the orlgln
of the muscle and A the midpoint of its ocular (anatomic) insertion.
of the muscle, as well as the line at which the predominant force

acts, may be displaced relative to the ocular axes, during an eye ro
tation.· For instance, if an eye elevates, the inferior, stretched~
fibers of a horizontal rectus act more than the superior, shortened,
fibers of the same muscle, the opposite being true when the eye is
depressed. The effective insertion of a horizontal rectus tends,the~
to move in the direction opposite of that of the eye (down when the
eye elevates, and vice-versa), but the extent of such a displacement
is also dependent of the elasticity of the peri-muscular structures.
Because the passive length-tension curves of each muscle's margin are
not necessarily the same (unless the ocular rotation occurs exactly
at the plane of muscular action), and in view of the selective pro-
cedures which increasingly are being proposed (marginal tenotomies,
slanting of ocular insertions, etc.), it follows that an interesting
question to be solved is how the innervation is topographically dis-
tributed in the muscular fibers.
7) Definition of the angular measurements. Finally, it is to be

remembered the need of a standardization of the measurements. Since
it is possible to define horizontal, vertical and torsional rotations
around fixed (orbital) and/or movable (ocular) axes, many combi-
nations of systems may arise. For instance, Helmholtz proposed to
measure vertical angles (elevations) around a fixed (horizontal) axis
and horizontal rotations (azimuths) around a movable (vertical) axis,
while Fick took horizontal angles (longitudes) around a fixed axis
and vertical rotations (latitudes) around a movable axis. Obviously,
depending on the system which is used to measure rotations, the same
ocular position may be defined by quite different sets of coordinates.
8) Operationalization of the ocular position. Very strict con-

ceptual discussions concerning the definition of the fixation of an

object (with its optic, anatomic and physiologic considerations, as
that of the visual "axis", which is not a line, but a cone) would
lead to limits of precision of measurements, stability of the eye,
etc., but this is beyond the scope of this paper.
GENERAL BALANCE
The disposition of the six extra-ocular muscles makes reason-
able the assumption that they may be considered in pairs. So the hor
izontal muscles, laying exactly at the horizontal plane,are mutually
exclusive in their pure actions of adduction (the medial rectus) and
abduction (the lateral rectus). In primary position, the relative
values of the vectorial components for the vertical, torsional and
horizontal rotations may be taken respectively by the approximate
proportions of 8, 4 and 3 for the vertical recti and 6, 8 and 1 for
the obliques, such that all actions in each plane are mutually ex-
clusive for each pair of muscles (elevation and incycloduction of
the SR balanced by depression and excycloduction of the IR; ele-
vation, excycloduction and adduction of the IO balanced by depression,
incycloduction and abduction of the SO), except adduction due to the
vertical recti. This could be a possible hypothesis for the abduction
of the eye when all the tonus of the extra-ocular muscles is elimi-
nated. (Even if the oblique pair were considered to cause abduction,
that certainly would not compensate the adducting power of the ver-
tical recti).
On the other hand, it is not credible that anatomic symmetry

of the muscular coordinates be a necessary condition for the tonic
balance of the eyes. Therefore, slight anatomic asymmetries of the
muscle pairs and/or unbalance of (other) passive forces could ex-
plain, for instance, the vertical deviation which is also common when
all the muscular tonicity is absent.
Dynamic Balance
From the beginning to the end of an eye movement, vectorial com

ponents change in magnitude and, frequently, in direction as well~
Therefore, the rotational components with which each muscle con-
tributes to a whole rotation is between the initial and the final
states of the muscular action. For instance, while the vertical
action of the SR increases from the primary position (0°) to a po-
sition of abduction of about -26.5° and then decreases again sym-
metrically up to an abduction of about -53°, the horizontal and
torsional actions undergo similar symmetrical variations of actions,
though opposite in direction (from an adducting to an abducting
power and from an incycloducting to an excycloducting power). The
vectorial integration between the ocular positions 0° and -53° is
therefore nul for the case of horizontal and torsional actions. Obvi
ously, however, the influence of the muscle in each plane of action
74 H. E. A. BICAS
is not restricted to the vectorial component with which it can act,
but depends also of the applied force (stimulated by innervation).
That force may also vary with the ocular position. It results that a
combination of both factors is essential to the determination of the
dynamic performance of a muscle during a rotation.
Consequently, even in the case of a "simple" rotation, a complex

innervational correlation seems to be necessary. For a pure depression
of the eye, theoretically only two muscles could initiate the
movement: the IR and/or the SO. If the IR were the only muscle to
contract, its antagonist would have to relax "reciprocally"
(Sherrington's law). The resultant increase of the innervatjonof the
IR (evoking depression, excycloduction and adduction) and of the
relaxation of the SR (evoking depression, excycloduction and ab-
duction) would be of a desired movement (depression) and of a
parasitic one (excycloduction) since, for the sake of simplification,
the horizontal, now antagonic, actions are supposed to self-cancel.
However, as the movement (depression) progresses, the IR increasingly
loses its horizontal rotational torque, while the SR augments it and
becomes more stretched (which increases its passive pull). In fact,
although some variations promote adduction (contraction of the IR,
stretching of the SR, increase of the horizontal torque of the SR),
others induce abduction (by the opponent forces) since there is a
relative decrease of adduction (smaller passive action of the IR,
decrease of the horizontal torque of theIR, relaxation of the SR).
Therefore, depending on the values of each of such variables, ~
effect may result, i.e., this pair of "adducting" muscles may, in
depression, and compared with the primary position, promote a real
adduction, no horizontal action or even allow an abduction. This is
not surprising and provides the basis on which a centrifugal ro-
tation in the direction of action of a paralytic muscle may be res-
tored by the relaxation of the antagonist, sound muscle, previously
balanced by an increment of synergistic (opponent to the sound, an-
tagonistic, muscle) elastic forces (Bicas, 1984).
Similar reasonings may be used for the pair of obliques: then

the eye would present depression, incycloduction and abduction. Sup-
posing that the horizontal action could not happen (because of ana-
tomic equivalence and/or functional compensation),incycloduction of
the obliques and excycloduction of the vertical recti could be
neutralized by unequal variations of innervation to the vertical
recti and to the obliques (being already expected that that of the
vertical recti be greater than that of the obliques). Obviously, the
horizontal recti may be also considered to act, to avoid undesirable
rotations.
In summary, even for "simple" rotations, undesirable movements

originating from the action of one muscle or a pair of muscles, are
bound to happen and must be avoided by the function of other muscles.
Probably all muscles have to be considered in the study of any ro-
tation, which suggests that the distribution of innervation among
them is complex and not uniform in different cases: it has to be
specifically regulated according to the need to balance the multi-

ple factors involved.
REFERENCES
Bicas, H.E.A. (1984). Principios teoricos de substitui~ao de
a~ao rotacional de musculo extra-ocular. I - Generalidades.Arq.
Bras.Oftalmol., ~. 154-159.
Bicas, H.E.A. (1985). Principios teoricos de substitui~ao de

a~ao rotacional de musculo extra-ocular. VI - Efeitos de trans-
la~ao ocular e suas causas. Arq.Bras.Oftalmol., 48, 16-22.
Boeder, P. (1961). The co-operation of extraocular muscles. Am.

J.Ophthalmol., 51, 469-481.
Collins, C.C., O'Meara, D. & Scott, A.B. (1975). Muscle tension

during unrestrained human eye movements. J.Physiol., 245, 351-
369.
Krewson, W.E., III (1950). The action of the extraocular mus-

cles. A method of vector analysis with computations. Trans. Am.
Ophth.Soc., 48, 443-486.
Park, R.S.; Park, G.E. (1933). The center of ocular rotation in

the horizontal plane. Am. J.Physiol., 104, 545-552.
Robinson, D. A.; O'Meara, D.M.; Scott, A.B.; Collins, C.C. (1969).

Mechanical components of human eye movements. J.Appl.Physiol.,
26, 548-553.
Volkmann, A.W. (1869). Zur Mechanik der Augenmuskeln.Tr.Leipzig

Soc.Sc., 21,28-70.
ACKNOWLEDEMENTS
To Funda~ao de Amparo a Pesquisa do Estado de Sao Paulo (Medici

na 87/1182-9), and to Conselho Nacional de Desenvolvimento Cientifi=
co e Tecnologico (Proc. 402131/87. 8- CL).
I am also grateful to Dr.Sidney J.Faria e Sousa, to Dr. Robert

L.Zimmermann and to Mrs. Josefina Pisi de Queiroz for editorial as-
sistance in preparing the final manuscript.
6
PHYLETIC ORGANIZATION OF BRAINSTEM NEURONAL
CIRCUITS AND THE ETIOLOGY OF STRABISMUS
ROBERT BAKER
If I focused on established facts only, then the role of

brainstem neurons could be elaborated on this page; however, since
the conference intends to be contemplative, but in truth promises
more, this presentation will be of a contemporaneous nature.
Thus, the following commentary generally addresses versional and
vergent gaze pathways with enormous assumptions briefly, more
acerbically, stated. An evolutionary strategy is used to estab-
lish a rational structural basis for understanding genetic and
epigenetic factors that may be causal in the production of stra-
bismus and amblyopia. I can state that given present naiveties
about these particular issues and the clear limitations in obtain-
ing experimental solutions for mammalian CNS disorders, the out-
look for even positing a primary site, let alone its alleviation,
is not promising.
STATEMENT OF THE PROBLEM
Since I agreed to contemplate the relevance of brainstem motor
organization, I expect to pursue (challenge) the task to the end,
but I will not be able to provide a satisfying denouement just
yet. At the outset, I failed to realize that the literature was
so vast and empty. In spite of some great psychophysics and
behavior, there is little concept of the neurons or circuits
involved, either from the viewpoint of phylogeny (evolution) or
ontogeny (development), and each point could, but won't, be
elaborated fairly extensively. Strabismus is envisioned as a
cosmetically undesirable characteristic largely afflicting man-
kind. Given the incredulously disproportionate focus on our
species, the condition has clearly attracted great interest but
with little consequence for either understanding the phenomenon or
relieving the malady.
My first contemplation of this question coincided with evalua-
tion of a recent study (Tychsen and Lisberger, 1986) whose ration-
ale was particularly appealing. The answer appeared to be clearly
related to maldevelopment in the sensory rather than the motor
system. This notion led to another arguing that the deficit might
arise during maturation of the visual cortex and therefore be
related, inextricably, to visual experience. The sequelae of that
77
78 R. BAKER
hypothesis held that £nY misalignment of the visual axes, especi-

ally during the 'critical' period, conferring non-homologous re-
tinal images at cardinal CNS sites would produce permanent
sensory/motor damage. Similar arguments appear throughout the
literature and they typify a particular conundrum. Strabismus is
rarely thought of as a disease: "A condition of one part that
impairs the performance of a vital function." As a result, the
ubiquitousness of ideas offered for the factual interpretation of
its neuronal basis epitomises phenomenalism! Collectively, they
are directed at the question of ~. and not How, neural circuits
and/or behaviors become manifest. Without an inkling of the lat-
ter the former is nearly always unattainable. Given my predilec-
tions for establishing structural homology between species there
should be little ambiguity about either the direction or content
of this contribution.
GENERAL PHYLOGENETIC PRINCIPLES
In the absence of space for details, a background can be set
with a few broad assumptions formulated by others (summarized by
Gould, 1985). First, a major point of view argues that the
vertebrate CNS must be viewed concurrently as a static (finite)
and dynamic (prospective) entity. In a given species, the extant
sensory/motor machinery reflects a continuous and certainly unend-
ing evolutionary process captured momentarily in a single frame.
Phylogenetically, new and old (i.e., over evolutionary as opposed
to individual developmental time) neurons and their circuits are
so intimately congruent that resolving their functional composi-
tion is largely indeterminate from the viewpoint of our present
level of experimental analysis. Secondly, this organizational
embellishment is confounded by the likelihood that new behaviors
(e.g., disparity vergence) actually precede new neurons and/or
synaptic circuits specialized for a particular function in subse-
quent species (the unending process of natural selection). It is
worth emphasizing that neurons, as opposed to axonal pathways and
terminal arborizations, are not generally new, in the sense of
being unique for a species, as they are derived from neuronal pre-
decessors (i.e., homologous to a precursor phenotype). For exam-
ple, the mammalian accessory abducens nucleus is quite likely
comprised of motoneurons that separated from the abducens nucleus
in a pre-amphibian ancestor. Stated colloquially, before leaving
the motoneurons performed both eye rotation and retraction, but
only the latter after their departure (Baker, 1986). Thus, in-
distinguishable behaviors can arise from peculiar arrangements of
circuitry in which the neural organization may differ in each
descendent species from that conceived for a primitive ancestor
(in the strict sense of preceding phylogenetic age).
EVOLUTIONARY INTENTIONS FOR EYE MOVEMENT
The ability to perform symmetrical and conjugate movement of
the two eyes about any axis of three dimensional head rotation is
an ancestral vertebrate feature. More importantly, displacement
of one eye, or both simultaneously, occurs over an enormous range
of amplitude and velocity that is subject to extensive adaptive
control. In the 'early' vertebrate days, the purpose of conjugate
(i.e., symmetrical and parallel) eye movement was presumably
BRAINSTEM NEURONAL CIRCUITS AND STRABISMUS 79
related to both enhancing brightness (utilizing primitive retinal

streaks and possibly binocular overlap) as well as facilitating
lens accommodative reflexes permitting target recognition
(probably motion more than resolution of detail). A singularly
curious fact is that nearly all antecedent fishes generally
exhibit spontaneous, convergent saccades in which the adducting
eye adds the convergent angle. In the face of an ever changing
ocular alignment, the premise arises that the neuronal basis for
strabismus reflects a primitive (i.e., existed first) organi-
zation. The lack of alignment only became deleterious within the
context of a 'new' behavior!
Several lines of rationale need to be explored in order to
explain how natural selection united eye frontalization, retinal
specialization and ocular motility in a common endeavor. At what
juncture did enhanced sensory function allow for 'new' neuronal
circuitry in which accommodative and disparity vergence were
selected to, first, separate from circuitry producing conjugate
eye movement and second, generate their individual neuronal
centers. At issue is - When do new behaviors in fact acquire
their own peculiar set of neurons as opposed to merely continuing
to use those in existing neuronal nets that are capable of produ-
cing that behavior? At this point, emergent properties of
neuronal function, notably in the cerebral cortex, must be weighed
carefully in order to assess the extent, if any, of extrapolation
between existing animal models of strabismus. Since vergent
interocular angles are particularly common in all lateral-eyed
vertebrates, the study of neuronal mechanisms in antecedent
species is not automatically precluded from contributing to
experimental approaches that examine the role early acquistion of
accommodation, obviously the central circuitry, played in the
subsequent development of both mammalian accommodative vergence
and disparity vergence pathways (e.g., see argument in Miles,
1985). At the moment, we simply do not know how many times an
alignment behavior might have evolved. Nonetheless, I conclude
that transitions (i.e., evolution) in the oculmotor, clearly not
the sensory, system must be scrutinized in vertebrates, largely
mammals, if one wants to understand the etiology of strabismus.
HORIZONTAL EYE MOVEMENT IN MAMMALS
The neuronal history of conjugate eye movement is largely
conservative. The similarity of vestibular and internuclear
neuron arrangement in the brainstems of antecedent species to that
of mammals is striking; however, both degenerate and emergent
features have been conjectured (Baker and McCrea, 1979). In stark
contrast to accustomed 'oculomotor' dogma, there have never been
mirror symmetrical inhibitory/excitatory vestibular and reticular
pathways to medial rectus and abducens motoneurons (an exception
may be early fishes). Moreover, concomitant with the mechanical
separation of head from eye movement, the population of antecedent
(primitive) vestibular neurons dwindled as their behavior was
selected against, even while their axons were establishing new
targets (Baker, 1986). The latter event clearly resulted in two,
or more, types of horizontal eye movement related neurons in the
mammalian vestibular nucleus. By contrast, internuclear neurons
80 R. BAKER
arose early in evolution and from their inception shared every

brainstem afferent to the abducens nucleus (e.g., vestibular,
reticular, prepositus; see Baker and Spencer, 1981). Although
their conjugate, and indirectly vergent, task was initiated early
in phylogeny, an apparently sufficient morphological organization
requiring no modification was in the process of change (improve-
ment?). From a primitive peripheral location the internuclear
neurons distributed throughout the abducens nucleus in descendent
mammals (Baker, 1986). If this adaptive pattern paralleled
perfection of obligatory conjugacy in the mammalian horizontal
system then it simultaneously, and unwittingly, precipitated
constraints (fatal scenario?) in selecting for the neural organi-
zation responsible for 'alignment' behavior accompanying both
disparity and accomodative vergence. Why?
While popularity dictates moment to moment design on a gener-
ation time schedule, a structural/functional endpoint is virtually
molded in stone with extinction as the usual measure of correc-
tion. Consider the extensive symmetry in obligatory conjugate
linkage of the saccadic and vestibular reflexes. Innate adaptive
plasticity has be shown to be more than adequate to offset con-
siderable misalignment of optic (or visual) axes that appears
early in development or in the adult (Optican and Robinson, 1980;
Snow et al., 1985). A fundamental finding is that vergence-
related information is either absent or not found in a systematic
and/or predictably useful fashion in normal abducens internuclear
neurons (Mays and Porter, 1985; Delgado-Garcia et al., 1986).
Since this observation distinguishes between vergence and ver-
sional alignment at the neuronal level one can conclude that
symmetrical control of conjugate horizontal eye movement has been
structurally isolated during phylogeny. Given this theoretical/
experimental framework the critical, seemingly tautological,
extrapolation between CNS binocularity (motion and distance in the
cortex) and the overall adaptation (in an evolutionary sense) of
the existent sensory and motor machinery (i.e., the mammalian
repetoire of eye movements) is amenable to study in species that
address the strabismic condition.
SYMMETRY IN THE OCULOMOTOR SYSTEM
Our own eyes may be thought of as two limbs moving in mirror
symmetrical fashion thereby adhering to constraints specified by
principles of vertebrate bilateral, as opposed to radial, symmetry
(known more euphemistically as Herings' Law). Not only is sym-
metry a basic developmental design, but also it is intimately re-
lated to 'how' neuronal circuits are put together. Moreover, it
is an evolutionarily conservative mechanism relying on minimal
postnatal sensory experience (i.e., independent of epigenetic in-
fluence). In strabismus, the fundamental flaw is undoubtly at the
point that normally, and successfully, interfaces asymmetry with
the mirror symmetrical sensory (in this case vision and vestibu-
lar) and motor (extraocular muscle) organization. If evolution
had not been so parsimonious in the selection of coordinate
frameworks to specify the sensory/motor system for rotation
(direction), then the ensuing one for translation (distance) might
have avoided entanglement (coping with conjugacy).
Current knowledge suggests that during the course of verte-

brate evolution external visual and vestibular space became
vectorally overlapped onto common central components sharing the
same brainstem neuronal matrices. This idea suggests a precise
overlay of rotational coordinates onto neurons and/or circuits
that employ a common axis in which external (world) or internal
(self) rotation induces a common angular velocity in the rotating
body (the eye). By contrast, disjunctive eye movements in ver-
gence are responses to translational stimuli in which the new axes
are parallel to the old ones. The central concern is the neuronal
arrangement selected for through the course of evolution, which
interfaces the widely disparate linear and angular coordinate
systems. The near response utilizes its own intrinsic set of sen-
sory stimuli (disparity) without external reference to the conju-
gate network. Where is the neuronal linkage secured between these
two motion detecting systems? In one case, the design is to
ignore parallactic stimuli to the extent that the symmetry of
individual eye rotation is compromised except in binocular viewing
and, in the other, tracking of parallactic motion with each eye is
the major sensory feature. These radical, and diametrically op-
posed, behaviors coupled with separate, but co-adapting, develop-
ment in optokinetic and vestibular horizontal eye movement has
exacerbated a fairly straightforward basis for nasal to temporal
insufficiency into a primate-related causal reason for strabismus.
There is reason to assert that extraocular muscle coordinates
actually specify visual and vestibular geometry in the course of
ontogeny and natural selection. According to phylogenetic rules
not yet established (at least by me), the motor side remains con-
stant in evolution while central sensory paths are modified as the
eyes (location in the head) themselves are seemingly released to
adopt new axes of rotation as a consequence of neck co-adapta-
tion. This motor to sensory re-arrangement is further complicated
because the rotating bodies themselves have undergone spatial
transformation in two axes, torsion and vertical. Although move-
ment about z-axis rotation would be least affected, the addition
of a new joint (neck) required 'new' neurons (in the real sense)
to perform either independent eye version or vergence because the
original set of 'pre-neck' brainstem neurons synchronized both eye
and axial musculature by common branching. For this reason, one
might suspect that the additional degree of freedom added to
'gaze' further restricted the choice of an alignment mechanism.
THE STRABISMIC CONUNDRUM AND STRATEGY
Given spatial and temporal considerations, the final survey
needs to address actual behavior including circuitry (and/or lack
thereof) that might be the etiological factor in strabismic
models. At issue is - The proposal of a causal scenario that dis-
tinguishes the pre- from post strabismic condition (i.e., resolves
the chicken/egg analogy). Measurements of VOR and saccadic eye
movement metrics, while taking into consideration pursuit and
optokinetic asymmetry, suggest nearly normal function of the ver-
sional motor system (Kommerell, this volume). However, the
relevance of these findings are offset by the fact that many
(perhaps most) of the sensory and motor abnormalities reported in
82 R. BAKER
strabismus are not mutually precipitative for the disease ( Stark

et al., 1982). The corollary exists, of course, that certain
deficits may appear in the absence of strabismus (see Kommerell,
this volume).
Formulating an orderly ontogenetic relationship to explain
this disease has become an open ended task because the motor
performance can be acquired as opposed to genetically innate. The
status of eye movements in all models of strabismus and amblyopia
leads to an inescapable deduction that fits well under this sub-
heading. By and large, once manifest, misalignment of the optic
axis is permanent. This rationale suggests that a 'putative'
strabismic circuit is beyond 'structural' repair and as such
little more can be attempted than extraneous alignment correc-
tions. Accepting this idea implicity implies an unavoidable
symptomology in spite of restoration of ocular alignment during
development (this volume). Consequently, future efforts will
continue in two dichotomous directions. The first, in developing
orthoptic psychophysics for manifest conditions. The second,
formulating an experimental plan for understanding the central
morphological substrate.
The deficits reported in nasal/temporal pursuit and optoki-
netic responses, including the loss of stereopsis (and/or dispar-
ity vergence) are major reasons suggesting the cortex as neces-
sary, but not sufficient to produce strabismus. Given the status
in corrected conditions (Kommerell and Hoffmann, this volume),
these motility problems may be considered permanent secondary
acquisitions. Further exploration of the rationale and individual
circuit components intimates that the subcortical vergence
machinery offers the best warranty for dispelling the enigma.
ESTABLISHMENT OF A SUBCORTICAL LOCATION
How can a modest causal sequence be portrayed for the congeni-
tal condition? Surely, the fatal flaw is in the hardwiring (col-
loquially). If one begins with a small, but consistent, struc-
tural imperfection then visual axis misalignment becomes manifest
as the innate wiring diagram responds to developmental cues. The
undesirable behavior cannot be prevented. At least not yet.
Why? The plasticity 'software' purportedly available at the cor-
tical level cannot cope with the midbrain dilemma because the
capacity to do so is an emergent feature relying on the very same
midbrain alignment mechanism to establish the correct coordinate
framework. Even after the learning of binocular! ty, the role of
vision in motor feedback is uncertain. Lesion studies in both
infants and adults demonstrate remarkable recuperation in pursuit
and tracking eye movements. If there were any part of the mamma-
lian brain that should remain impervious to subcortical inade-
quacy, then it should be the cerebral cortex. By and large,
placing the primary deficit in a structure with remarkable adap-
tive plasticity extending so far beyond the critical period is
illogical. The reported deficits in cortical functioning (e.g.,
motion processing, pursuit, optokinetic) are so varied that it
would be farfetched to presuppose an evolutionary timetable sug-
gesting co-adaptation with the vergence system (i.e., they are
independent ontogenetic events!). In this regard, the question
that cannot be avoided is - Why can't motion processing and

pursuit be learned in both directions? Finally, if pursuit were
to be an emergent property of the cortex then co-adaptation for
optokinetic circuitry would have had to been selected for in order
to be directly related to the etiology of strabismus (and that
temporal sequence is very unlikely). I conclude that subcortical
motor levels must be the focus for conceiving a coherent
strabismic hypothesis.
VERGENCE BEHAVIOR AND SIGNALS
Obviously, the methods to discern epi-phenomenological from
causal observation are few in number, let alone demonstrate an
events' occurrence before, or after, the fact. The best place to
critically discriminate between a sensory/motor correlation and
hierarchial location in a neuronal pathway is at the level of the
single neuron. This determination is not a casual undertaking
because the neuronal response must be isolated, recorded during a
conflict vergence paradigm, studied with prism adaptation and then
correlated to cellular morphology. Although the above structure-
function correlates are obligatory, their sufficiency to elucidate
normal physiology, let alone establish the etiology of strabismus,
is not assured in mammals.
Recent behavioral and conceptual work in the vergence system
(Schor, this volume) coupled with neurophysiological correlates
(Judge and Cummings, 1986; Mays, 1984; Mays and Porter, 1984) con-
cludes that both velocity (phasic) and position (tonic) signals
are derived separately, and then subsequently combined, at the
motoneuronal level. Notably this parallel design for information
processing emulates the neuronal organization demonstrated to pro-
duce conjugate (e.g., saccadic) eye movement (Mays, 1984). The
kinematics of version and vergence are different; however, current
data suggests interdependence between saccades and vergence in
changes of gaze direction (rotation) associated with distance
(translation) (Erkelens, 1987; Schor, this volume). In various
combinations of monocular and binocular deprivation the conjugate
neuronal machinery becomes disassociated resulting in disjunctive
(vergent) saccadic-like eye movements (Sparks et al., 1986). Al-
though these observations imply that continuous higher order tun-
ing is required for proper performance of vergence and conjugate
eye movements they do not address the etiology of strabismus.
Experimental studies have separated sensory/motor attributes
of neural signals in vergence with success equal to that for
versional pathways. Even though midbrain neurons discharge in a
similar way to retinal blur, or disparity, recent data (Judge and
Cummings, 1986) shows that neuronal activity can be related to
motor output (accomodation or convergence) rather than to the
above sensory stimuli. Conflicting vergence and accomodation
tasks have revealed distinct groups of neurons for each behavior
suggesting the existence of 'multiple' integrators rather than the
singular one pictured in horizontal conjugate eye movement. In
the case of prism adaptation, the normal coupling between accomo-
dative and vergence tonic adaptable components are altered and
change the ratio cross linking these two parts of the motor system
(Schor and Kotulak, 1986). There is reason to assume that the
84 R. BAKER
acconnnodative vergence machinery (AC/A) is not ahmys equipped with

sufficient balance in adaptive control to respond appropriately
under certain viewing conditions (Shor, this volume). Since the
cross-link interactions between accomodation and vergence are
learned (i.e., temporarily modified by various environmental
manipulations; Miles, 1985) the misalignment process, by default,
must include the 'so-called' tonic adaptable component of vergence
(i.e., a slow integrator in the contemporary Schor model). If the
neuronal compromise were assumed to be directly at this point then
the 'state' of alignment would be equivalent to the 'state' of the
adaptive response as visualized by the magnitude of the coupling
ratio (all more or less independent of sensory cues). By selec-
tively situating the lesion at this hypothetical point, induction
of accomodative or disparity vergence would be least affected and,
even in corrected viewing conditions, the serious fatality would
be in the role oculomotor vergence plays in the initiation of
sensory fusion (stereopsis). In fact, distinct asymmetries in
adaptive responses are normal (Miles, 1985) however, when
excessive for vergence and inadequate for accomodation, mis-
alignment is the most likely, if not inevitable, outcome (Schor,
this volume).
CONSIDERATION OF A NEURONAL MECHANISM
In light of the rationale presented to this point, one feels
relieved, if not comfortable, to consider only the brainstem for
the hypothetical and elusive morphological flaw. According to the
phylogenetic timetable applicable to this scheme a 'new' motor
behavior appeared that first, coupled accomodation to vergence in
the near-response. Parceling of function, i.e. division of labor,
within that neuronal circuitry gave rise to 'specialization' of
purpose in individual midbrain oculomotor internuclear neurons.
In fact, extraocular motoneurons, themselves, have become
differentiated between version and vergence (Mays and Porter,
1984). Within the context of this progressive temporal and
spatial neuronal re-organization arose the 'alignment' predica-
ment. There is reason to argue that the mistakes lie either
within the intrinsic or extrinsic projections from, not to,
midbrain circuitry and, more specifically, involve misguided
axonal pathways rather than deviant terminal arborizations.
By and large, one can conclude that strabismus does, and
should have, a genetic determinant because the developmental
imperfection lies in construction of circuits. This reasoning
does not exclude involvement of epigenetic factors in the ensuing
etiology, but distinguishes between behavior as the experience-
related manifestation of circuit operation and the genetic clues
available at a particular ontogenetic moment. Behavior is
dependent on both the embryology and genetics of neuronal
organization. In the latter context, the favorite, but not
necessarily exclusive, hypothesis centers on aberrant axonal
pathways in the motor rather than sensory system.
The literature regarding albinism, innate crossed eyes and
role of visual sensory experience are presented in this volume;
however, there are clear cut correlates in motor systems (e.g.,
the pyramidal, oculomotor) that have not yet been studied with
equal emphasis. In the latter case, axonal projections initiated

during development persist to maturation and maintain abnormal
physiological responses (Baker, 1986). During the critical
period, this simple scenario cascades to influence all synaptic
contacts temporally related to the maturational process. This
rationale alone describes the causal sequence for cortical control
of subcortical optokinetic pathways in the nasal/temporal assym-
metry (Hoffman, this volume). Disparity vergence never really had
a chance because the constraints for cortical binocularity
(motion) required visual axis alignment as the initial specifying
condition for learning.
Obviously little is known yet about the extent of synaptic
relationships between midbrain vergence circuitry and pontine-
medullary horizontal eye movement pathways. Nevertheless vergence
signals obviously access conjugate centers under stimulus
dependent circumstance (Erkelens, 1987). However, even the most
elaborate scheme will probably not include more than the
mesencephalic vergence center extending to either the abducens
and/or to the prepositus nucleus. Indirectly, the cerebellar
flocculus should be mentioned because of its established
involvement in fixation nystagmus and eye position (i.e.,
alignment). Interestingly, an inordinant amount of vergence
information has been observed in the flocculus (67% of Purkinje
cells). Even so, the synaptic modification of the AC/A ratio does
not reside in this part of the cerebellar cortex, because
flocculectomy had no role in vergence adaptation (Judge, 1987).
Obviously, the detailed brainstem organization for vergence is
still surmise but, in essence, only those neurons and circuits
discussed in the next section are candidates.
OCULOMOTOR, ABDUCENS AND PREPOSITUS INTERNUCLEAR NEURONS
In descendent mammals, there is a progressive shift from an
ipsi- to contralateral predominance of the oculomotor internuclear
termination in the abducens nucleus. The projection is well
correlated with acquisition of vergence behavior. Moreover this
source of afferent input to the abducens nucleus is the only one
to exhibit any change in synaptic pattern. All others are
phylogenetically conserved conjugate pathways. If the oculomotor
internuclear neurons represent alignment error, not alignment,
then the primary flaw resides directly within the midbrain even
though manifestation of the defect is elsewhere (e.g., the
abducens nucleus).
Since this scenario could easily be thought of as interfering
with a 'corollary discharge' of central alignment (i.e., efferent
copy) the susceptible loop could also be located in the caudal
brainstem at the level of the prepositus nucleus. For the
moment, this suggestion is structurally attractive. The fascina-
ting point is citing an etiological mechanism in the caudal
brainstem (medulla); however, the prepositus is a phylogenetically
new structure with appropriate afferent and efferent connections.
The horizontal integrator comprises the caudal part and the
rostral end is intimately concerned with vertical eye movement
(i.e., a basis for the vertical deviation). The prepositus
receives visual motion information directly from midbrain
86 R. BAKER
optokinetic centers and, in turn, acts as the major interface

between the vestibular nuclei and the cerebellum. Most signifi-
cantly, the prepositus generates and distributes ubiquitously
throughout the brain, both eye position and velocity information
(efferent copy). Establishing a broader role for the prepositus
nucleus in horizontal eye movement, especially in relationship to
vision (optokinetic and smooth pursuit) is inevitable.
There is no experimental evidence for an alignment message in
the prepositus, but like for vergence, it hasn't been tested yet.
Any egregious eye position signal begins a vicious congenital
cycle of maldevelopment in motion processing thereby disrupting
binocular correspondancy and compromising completion of a pursuit
system capable of opposing the advancing momentum. The whole
ritual could easily unfold in the preposi tus nucleus and therein
might lie the essential 'strabismic' structural circuit. The
corollary of this idea links the pursuit system with strabismus
and nystagmus from birth unless rectified by proper brainstem
development.
THE CONCLUDING SCENE
The final common denominator (i.e., bottom line) is that the
initial design for the oculomotor system did not perceive the
other end of a phylogenetic tree (if that is us) in which natural
selection would be so aggressive as to combine ocular alignment,
foveal development (for pursuit) in conjunction with, but in dis-
tinction to, global motion detection (the optokinetic system) and
then add on the circuitry for disparity vergence (distance)
permitting stereopsis. Every one of these new behaviors carved an
imprint into extant neuronal networks distributed throughout the
CNS, from cortex to medulla. Both eyes could no longer simply
look at the world and each see it in the direction it appeared,
because the 'binocular' visual system demanded precise alignment
to see only one target (a unified world). Immense pressure was
placed upon temporal establishment of appropriate connections
especially those related to experience and, therefore, not
genetically determined spatial cues. Although a question obvious-
ly resolved by natural selection, it is surprising that the inci-
dence of visual motor problems is not higher given the realization
that the system is indeed constructed upon such a phylogenetically
fragile basis. Can an incongruous visual input due to simple
alignment mismatch early in ontogenesis produce the variety of
mistakes reflected by the loss of acuity, stereopsis and motor
control? Of course. In the end one must decide how to interpret
the general, but paradoxical, message that the cortex itself
actively promulgates, rather than corrects, some perverse
brainstem developmental abnormalities.
REFERENCES
Baker, R. (1986). Brainstem neurons are peculiar for oculomotor
organization. In Oculomotor and Skeletalmotor Systems: Dif-
ferences and Similarities. (eds. H.-J. Freund, U. Buttner, B.
Cohen and J. Noth) pp. 257-272, Elsevier, North Holland.
Baker, R. and McCrea, R.A. (1979). The para-abducens nucleus. In
Integration of the Nervous System. (eds. H. Asanuma and V. Wil-
son) pp. 97-122, Igakaku Shoin, New York.
Baker, R. and Spencer, R. (1981). Synthesis of horizontal con-

jugate eye movement signals in the abducens nucleus. Jap. J.
EEG EMG Suppl., Jl, 49-59.
Delgado-Garcia, J.M., DelPozo, S. and Baker, R. (1986). Behavoir
of neurons in the abducens nucleus of the alert cat. II. In-
ternuclear neurons. Neuroscience 17, 953-973.
Cumming, B.G. and Judge, S.J. (1986). Disparity-induced and blur-
induced convergence eye movement and accomodation in the mon-
key. J. Neurophysiol. 55, 896-914.
Erkelens, C.J. (1987). Adaptation of ocular vergence to stimula-
tion with large disparities. Exp. Brain Res. 66, 507-516.
Judge, S.J. (1987). Optically-induced change in tonic vergence and
AC/A ratio in normal monkeys and monkeys with lesions of the
flocculus and ventral paraflocculus. Exp. Brain Res. 66, 1-9.
Judge, S.J. and Cumming, B.G. (1986). Neurons in the monkey mid-
brain with activity related to vergence eye movement and ac-
comodation. J. Neurophysiol. 55, 915-930.
Gould, S.J. (1985). The Flamingo's Smile. W.W. Norton & Co., New
York, NY.
Mays, L.E. (1984). Neural control of vergence eye movements: con-
vergence and divergence neurons in midbrain. J. Neurophysiol.
51, 1091-1108.
Mays, L.E. and Porter, J.D. (1984). Neural control of vergence eye
movements: Activity of abducens and oculomotor neurons. J.
Neurophysiol. 52, 743-761.
Miles, F.A. (1985). Adaptive regulation in the vergence and accom-
modation control systems. In Adaptive Mechanisms in Gaze Con-
trol: Facts and Theories. (eds. A. Berthoz and J. Melvill
Jones) pp. 81-94, Elsevier, North Holland.
Optican, L.M. and Robinson, D.A. (1980). Cerebellar-dependent
adaptive control of primate saccadic system. J. Neurophysiol.
44, 1058-1072.
Schor, C.M. and Kotulak, J.C. (1986). Dynamic interactions between
accomodation and convergence are velocity sensitive. Vision
Res. 26, 927-942.
Schor, C.M. and Kotulak, J. (1986). Mutual interactions between
accomodation and convergence are reduced by tonic adaptation.
In Adaptive Processes in Visual and Ocular Motor Systems (eds.
E.L. Keller and D.S. Zee) pp. 135-142, Pergamon Press, Oxford.
Snow, R., Hore, J. and Vilis, T. (1985). Adaptation of saccadic
and vestibula-ocular systems after extraocular muscle
tenectomy. Inves. Ophth. and Vis. Sci. 26, 924-931.
Sparks, D.L., Gurski, M.R., Mays, L.E. and Hickey, T.L. (1986).
Effects of long-term and short-term monocular deprivation upon
ocular motor functions in the rhesus monkey. In Adaptive Pro-
cesses in Visual and Ocular Motor Systems (eds. E.L. Keller
and D.S. Zee) pp. 191-200, Pergamon Press, Oxford.
Stark, L.W., Ciuffreda, K.J. and Kenyon, R.V. (1982). Abnormal
eye movements in strabismus and amblyopia. In Functional
Basis of Ocular Motility Disorders (eds. G. Lennerstrand, D.S.
Zee and E.L. Keller) pp. 71-82, Pergamon Press, Oxford.
Tychsen, L. and Lisberger, S.G. (1986). Maldevelopment of visual
motion processing in humans who had strabismus with onset in
infancy. J. Neuroscience Q, 2495-2508.
7
NEURAL BASIS FOR CHANGES OF THE OPTOKINETIC
REFLEX IN ANIMALS AND MEN WITH STRABISMUS
AND AMBLYOPIA
K.-P. HOFFMANN
1 Introduction
Numerous experiments dating back to the last century have shown
that visual structures in the midbrain of vertebrates play an
important role in visually guided behaviour. In addition to the
oculomotor nuclei themselves there are two other midbrain structures
with clearly defined visuomotor functions. The superior colliculus
(SC) is important for saccadic eye movements as well as head- and
body- movements in the visual grasp reflex and the nucleus of the
optic tract (NOT) is essential for slow eye movements in the
optokinetic reflex (OKR) (Akert 1949; Wurtz and Albano 1980; Precht
1982; Hoffmann 1983a; Vanegas 1984). Comparative neuroanatomical
studies of mammals representing an ascending sample from the
phylogenetic tree have shown that both the SC and the NOT in the
pretectum receive a continually increasing proportion of their input
from cortical visual areas as opposed to the well established direct
retinal projection. In relation to the well established rule that
ontogeny recapitulates in part mechanisms of this phylogeny
developmental studies of these structures may offer the opportunity
to examine the rules by which information from these different
sources is used and integrated to create the specific properties of
nerve cells in SC and NOT of modern mammals. In this chapter
examples will be reviewed to show that the specific filter
characteristics of retinal recipient midbrain nuclei are established
early in ontogeny by the retinal input and before the arrival of
cortical afferents. Thereafter information flow via cortical
connections is accepted in the midbrain only if it agrees with the
complements of the retinal imprint. Nevertheless, in adult cats and
monkeys analyses of the response properties of SC and NOT cells have
revealed the strong influence of the cortical input and the
dependence of some specific properties on that input. For example,
visual cells in the SC of the cat show a high degree of binocularity
and direction specificity. These properties are lost when the
cortical input is disrupted (Wickelgren and Sterling 1969). Visual
cells in the NOT of cats or monkeys are all direction specific over
89
90 K.-P. HOFFMANN
a large range of velocities and mostly (in cats) or always (in

monkeys) binocular. NOT binocularity and responsivity to high
stimulus velocities are lost after decortication whereas direction
selectivity for lower velocities is maintained by the direct retinal
input (Hoffmann 1981).
In this review a model is proposed that could explain the
differences in the optokinetic reflex (OKR) of various mammals as
well as the impairment of OKR with strabismus and amblyopia by the
specific interactions between retinal and cortical projections in
the NOT. The model is based on the following assumptions.
1. A genetically prespecified retinal input reaches the
contralateral NOT first during ontogeny. This connection
develops independently of early visual experience.
2. Thereafter information flow via cortical connections is accepted
in the NOT only if it agrees with the complements of the retinal
input. The development of this connection is critically
depending on normal binocular visual experience.
3. After the cortico-pretectal connections have been established,
the retino-pretectal connections gradually lose their influence
and are replaced by cortical afferents.
This model explains why after the loss of visual cortex the OKR
is much weaker and asymmetric in cats and monkeys or absent in man
and why amblyopia caused by visual deprivation or strabismus leads
to a loss of binocularity in the NOT and as a consequence to an
impaired and asymmetric OKR.
2 Properties of neurons in the nucleus of the optic tract and their

relation to OKR --
In the NOT of all mammals tested so far (for a review see
Simpson (1984)) neurons with very large (up to 90° horizontal width)
receptive fields and with direction specific responses to stimulus
movement over a very broad velocity spectrum ( 0.1°/s- 100°/s)
have been found. There is also an internal structural segregation
according to the preferred stimulus direction. Almost all cells in
the left nucleus prefer stimulus movements to the left and those in
the right nucleus prefer stimulus movements to the right in the
visual world (Collewijn 1981; Hoffmann and Schoppmann 1981; Hoffmann
and Distler 1986). The output of these neurons goes to at least 3
sites, i.e. the dorsal cap of the inferior olive, the nucleus
prepositus hypoglossi and the area of the nucleus reticularis
tegmenti pontis (Precht et al. 1980; Magnin et al. 1983; Lannou et
al. 1984). All these areas have been shown to contribute to the
control of reflex or voluntary eye and head movements in different
animals. Interestingly, recordings from NOT neurons in awake cats
with implanted search coils to measure eye movements showed that the
discharge rate is entirely dependent on the retinal slip of the
stimulus and is not modulated during optokinetic afternystagmus or
by vestibularly induced nystagmus in the dark (Hoffmann and Huber
NEURAL BASIS FOR CHANGES OF THE OPTOKINETIC REFLEX 91
1983). Electrical stimulation through NOT recording electrode

(pulse width 1ms, frequency 60Hz, maximal amplitude 0.5mA) elicits
clear nystagmus with the slow phase towards the stimulated site.
Therefore, independent of whether visually or electrically driven,
an increase in neuronal activity in the left NOT over that in the
right NOT leads to optokinetic nystagmus (OKN) with slow phase to
the left whereas higher activity in the right NOT leads to OKN with
slow phase to the right. Slow phase eye velocity during OKN was
always slower than the stimulus velocity and cells in the left NOT
would discharge at a rate above spontaneous activity as long as the
stimulus moved leftward and vice versa (Hoffmann and Huber 1983).
In normal cats or monkeys each eye can activate NOT cells on
either side of the brain because many or all NOT cells are
binocular. It directly follows that monocular OKR is symmetrical
with nearly equal gain for the two horizontal directions. The well
known asymmetry of OKR seen so often in strabismic patients can be
explained by the loss of binocular responses of these NOT cells. If
in strabismic patients the retina is connected only to cells in the
contralateral NOT monocularly tested OKR has to be weaker for
stimuli moving from nasal to temporal in the visual field in
comparison to stimuli moving from temporal to nasal.
~ Retinal input to the nucleus of the optic tract

In the cat the conduction time along retinal axons from the
chiasm to the NOT is 3 - 7ms. By this measure, the axons are slowly
conducting and must originate from W-cells in the retina (Hoffmann
and Schoppmann 1981). In primates like macaque monkeys the
conduction time along retinal axons from the chiasm to the NOT is
3 - 7ms which leads to a conduction velocity of 3 - 7m/s assuming a
conduction distance of 21mm from the chiasm to NOT (Hoffmann et al.
in press). This clearly is in the low conduction velocity range of
retinofugal axons in the monkey. Such axons originate from so
called "rarely encountered cells" which have special receptive field
properties (Schiller and Malpeli 1977). In the cat we sought to
directly identify the type of ganglion cells in the retina
projecting to NOT by recording in the retina and identifying
ganglion cells which could be antidromically activated by low
current stimulation in the contralateral NOT (Hoffmann and Stone
1985). From a sample of 558 retinal ganglion cells only 11 W-cells
met both of the criteria for putative retinal afferents to the NOT:
appropriate low conduction velocity and low threshold of antidromic
stimulation. Of interest is the presence of 5 on-center direction
selective cells among the 11 putative NOT afferent cells because
this is much higher than the proportion of these cells among W-cells
generally. Although direction selective retinal ganglion cells
still await identification in the monkey retina we would like to
suggest that direction selective on-center cells may form the major
input to the NOT in all mammals (Collewijn 1981).
92 K.-P. HOFFMANN
i Cortical input to the nucleus of the optic tract

Cortical cells in area 17 and 18 projecting to the NOT in the
cat were identified by antidromic stimulation of the terminals of
such cells in the NOT (Schoppmann 1981). In agreement with many
anatomical and physiological studies these cells were shown to be
layer V pyramidal cells. The response properties of antidromically
activated cells were very similar to the so called cortico-tectal
layer V pyramidal cells. All units had (for the visual cortex)
quite large receptive fields (up to 5° in diameter) and could be
equally well activated by oriented light bars as well as by large
area random dot patterns moved across their receptive field. All
the antidromically activated units were direction selective to some
degree with a preference for horizontal movements on the average. A
wide range of stimulus velocities was effective in driving these
cells and most of them responded well to speeds greater than 20°/S.
All cells but one were binocular and the binocular responses were
always stronger than those to monocular stimulation. Electrical
stimulation in area 17 and 18 in the cat or in area V1 and the
middle temporal (MT) as well as middle superior temporal (MST) area
in macaque monkeys activated all cells recorded in the NOT of these
animals. Neurons in MT and MST are particularly sensitive to the
movement of visual stimuli and are mostly direction specific.
Lesion studies of Wurtz and his co-workers have recently shown that
areas within the superior temporal sulcus may be involved in the
control of the optokinetic reflex (Duersteler et al. 1986).
The comparison of retinal and cortical input to the NOT very
clearly shows that the typical response profile of cells in the NOT
of adult cats and monkeys reflects a very strong cortical input. As
has been elaborated in previous reviews (Hoffmann 1983a, 1986)
bilateral lesions of the visual cortex in the cat result in NOT
cells that are only weakly modulated by visual input and respond
only to stimuli moving at speeds less than 20°/s. Without a visual
cortex all cells are exclusively driven by the contralateral eye.
Clearly the functional role of the cortical visual input to the NOT
added to the retinal input, is to contribute binocularity and
increased responsiveness to higher stimulus velocities.
~Effects of visual deprivation on the ontogeny of the receptive

field properties in the nucleus of the optic tract
To study the role of visual experience for the development of
normal receptive field properties in the NOT of cats we examined the
effect of monocular lid suture for 6-24 months, beginning the first
week after birth. In these cats binocularity in the NOT was always
reduced. In most cases all NOT neurons lacked an ipsilateral input
regardless of whether the deprived or the non-deprived eye is the
ipsilateral one. The neurons driven by the deprived as well as by
the non-deprived eye exhibit a clear direction specificity for
temporo-nasal movements of optokinetic stimuli. They differ,
however, significantly in their capacity to respond to high stimulus

velocities. Only cells which can be stimulated through the
non-deprived eye can respond when the stimulus moves faster than
20°/s. Direction selectivity in NOT cells of the cat is thus
independent of visual experience whereas responses to high
velocities and binocularity do not develop without visual
experience. This is consistent with the view that responses to high
velocities and binocularity are mediated through visual cortex and
that this structure is plastic during early life and is dependent on
visual experience whereas the development and organization of the
direct retinal inputs to the NOT is not. Concurrent changes can be
observed in the OKR (Markner and Hoffmann 1985). When testing the
deprived eye OKR is totally asymmetric. Only stimuli moving at low
velocity from temporal to nasal elicited OKN. Also the OKR of the
non-deprived eye can be asymmetric. Again OKN responses to nasa-
temporal stimulus movement are abnormally reduced or impaired.
A simple explanation can be given for the changes ipsilateral
to the deprived eye. The deprived eye appears to have lost its
cortical connections (Hubel and Wiesel 1970) and all neurons in the
NOT are controlled by the contralateral non-deprived eye and show
normal properties. The changes in the NOT contralateral to the
deprived eye are quite different. Here all cells are driven by the
deprived eye. The ipsilateral cortical input which should be able
to mediate the influence of the non-deprived (ipsilateral) eye has
not developed its normal function. The anatomical substrate for
this cortico-pretectal projection is still present and can be
demonstrated with electrical stimulation of the corticofugal fibers
as well as with retrograde tracing methods. Electrical stimulation
of the cortical areas leads to postsynaptic action potentials in NOT
cells. But this cortical projection to NOT cannot change the rate
of action potentials in response to visual stimulation (Hoffmann
1983b). As we have to accept that cortical cells driven by the
non-deprived eye send axons or axon collaterals to the NOT, the
disruption may be at the cortico-pretectal synapse. We suggest that
a mechanism following Hebb's paradigm (Hebb 1949) controls the
placement of cortical terminals on NOT cells in accordance with
visual experience. If, however, the retinal influence is blocked by
a closed lid during the sensitive period and NOT cells discharge all
the time at their spontaneous rate, the cortical axon terminals may
make inappropriate or subthreshold connections in the NOT. Although
the non-deprived eye remains able to drive cells in the ipsilateral
cortex, the cortex is unable to exert a functional influence on the
ipsilateral NOT.
94 K.-P. HOFFMANN
~ Effects of strabismus on the ontogeny of the receptive field

properties in the nucleus of the optic tract
Rendering kittens strabismic by surgical section of the medial
rectus results again in severe disruptions of binocular connectivity
in the NOT (Cynader and Hoffmann 1981). In the normal cat nearly
one half of the units encountered could be driven by visual stimuli
presented through either eye (Hoffmann and Schoppmann 1981) with the
remaining units influenced only by visual stimuli presented through
the contralateral eye. In the strabismic animals virtually all the
units encountered could be driven only via stimulation of one eye,
namely the eye contralateral to the NOT under study. Binocular
input onto single cells was virtually abolished. Each eye became
the sole source of input to the NOT on the opposite side of the
brain. In these strabismic cats OKR tested monocularly is
asymmetric and elicited only by slow stimulus velocities in tempore-
nasal direction (Cynader and Harris 1980).
Apart from the disruptions of binocular connectivity described
above, response properties of NOT cells in normal and strabismic
cats appeared similar. As in normal cats, all cells recorded in the
left NOT of the strabismic cats responded best with horizontal
movement from right to left and vice versa. Their activity was
below spontaneous rate when the stimulus moved in the direction
opposite to the preferred one. Maximum excitatory response occurs
at lower stimulus velocities (near 1°/S) than normally and the
units' ability to distinguish the two directions of motion declines
steadily once stimulus velocity increases beyond a few degrees per
second.
The loss of input from the ipsilateral eye in strabismic cats
appears to be attributable again to a partial or complete functional
suppression of the pathway from the visual cortex to the NOT. This
suppression of ipsilateral eye response may be a consequence of
unequal strength of input from the two eyes in normal cats. Input
from the contralateral eye reaches the NOT via a direct pathway from
the retina and via the visual cortex while that of the ipsilateral
eye takes only a route passing through the visual cortex and the
strength of these two functional pathways is unequal. In the
strabismic cats, the different directions of view of the two eyes
must cause a lack of spatial and temporal congruity of inputs
reaching the NOT via these two pathways. It appears as though
incongruity of input via two pathways of unequal strength is a
sufficient condition for the functional suppression of the weaker
pathway. Thus even though each eye reaches adequate stimulation,
the relatively weaker pathway from the ipsilateral eye to the NOT
via the visual cortex becomes non-functional in the strabismic
animals.
7 Shaping of the receptive field properties in the NOT by selection

of correctretTnal and cortTCalinput during early Tiifancy
The proposed model follows closely the summary made recently by
Fawcett and O'Leary (1985) of the role of electrical activity in the
formation of topographic maps in the nervous system (Changeux and
Danchin 1976; Cowan and O'Leary 1984). Basically we assume that
survival of retinal or cortical projections or consolidation of
their synapses is dependent on the availability of a critical amount
of a trophic survival factor. This survival factor is located in
the target cells and the amount released is strongly dependent on
the degree of depolarization of the target cells. Many inputs with
the same properties releasing their transmitter at the same time
will depolarize the target cell more strongly and as a consequence
receive more trophic survival factor than inputs which discharge
asynchronously.
During early development, axons of direction selective retinal
ganglion cells which respond to horizontal temporo-nasal movements
grow towards and terminate within the NOT. This growth is
genetically specified or due to as yet unidentified influences. The
postsynaptic cell is subsequently strongly direction selective for
global movement or retinal slip in temporo-nasal direction, i.e.
the cell will be strongly depolarized with movements in
temporo-nasal direction presented to the contralateral eye and
remain uninfluenced or become hyperpolarized with movements in the
nasa-temporal direction. At birth, direct retinal axons from the
ipsilateral eye have a much weaker influence on NOT-cells. Also, as
has been shown for the rabbit (Oyster and Barlow 1967), most retinal
on-center direction selective ganglion cells prefer the opposite
horizontal direction as the ipsilateral NOT (the temporo-nasal
direction for the ipsilateral eye is equivalent to nasa-temporal in
the contralateral eye). Thus retinal terminals from the ipsilateral
eye will mostly discharge when the target cell is not depolarized by
its major input from the contralateral eye and the ipsilateral
retinal input present at birth may even be weakened or eliminated
during this shaping process. The retinal projection from the
contralateral eye does not depend on experience because deprivation
does not alter this connection. Rotation of an eye will also rotate
the preferred direction of the cells in the NOT contralateral to the
rotated eye by the same amount (Hoffmann and Cynader unpublished).
Thus a genetically determined retino-pretectal specificity is
already present before the cortico-pretectal projection matures
(about 4- 6 weeks after birth). The selection of the correct
cortical axon terminals may occur according to the following rule:
Cortical cells strongly activated by the contralateral eye and
projecting to the ipsilateral NOT will discharge their terminals at
a higher probability in synchrony with retinal axons than cells
strongly activated by the ipsilateral eye. In addition only those
cortical cells qualify whose preferred direction is identical to the
preferred direction of the retinal axons terminating on the target
cells. These cortico-pretectal terminals receive enough survival
96 K.-P. HOFFMANN
factor because this substance is released by the target cell only

upon strong depolarization due to the retinal input. This process
will select cortical axons carrying the same direction specific
signal from the contralateral eye as the retinal axons. Other cells
will be less successful depending on how much their preferred
direction deviates from that of the retinal axons or how strongly
they are influenced by the ipsilateral eye. Under the assumption
that the activity of axons with information from the same retina is
more correlated than the activity from different retinae, the
ipsilateral retina will be connected to NOT-cells only as an
accompanist of the contralateral retina, i.e. through binocular
cortical cells (Hoffmann 1987). After the cortico-pretectal
connections have been established, the retino-pretectal connections
may gradually lose their influence and be replaced by cortical
afferents. This process may occur to a small extent in cats, to a
greater extent in monkeys, and almost completely in humans.
8 Conclusions
This model could explain why after the loss of visual cortex
the optokinetic reflex is much weaker and asymmetric in cats or
monkeys and absent in man. This model could explain why in
monocularly deprived cats suffering an occlusion amblyopia OKR is
asymmetric. The non-deprived eye can make no connections to the
ipsilateral NOT. The NOT cells are never sufficiently depolarized
by the sparse ipsilateral retinal input and the activity normally
relayed by the more massive contralateral retinal input is blocked
by the lid suture. The deprived eye (amblyopic eye) has lost its
cortical connections. This model could also explain why OKR is
asymmetric in strabismic cats. Only contralateral input controls
the NOT responses. The ipsilateral eye is connected to the NOT only
through binocular cortical cells and binocular cells are rare in
strabismic cats. All these inferences of course could also apply to
monkeys and man.
So far we have no information on the development of this
pathway in primates. The similarities in the ontogeny of the
optokinetic reflex and in the NOT response properties in adult cats
and monkeys suggests, however, that similar developmental mechanisms
may exist in all higher mammals, including man.
9 References
~ert, k. (1949) The visual grasp reflex. Helv. Physiol. Acta,
7. 112-134.
Ballas, I., Hoffmann, K.-P., Wagner, H. J. (1981) Retinal
projection to the nucleus of the optic tract in the cat as
revealed by retrograde transport of horseradish peroxidase.
Neurosci. Lett., 26, 197-202.
Changeux, J.-P., Danchin, A. (1976) Selective stabilization of
developing synapses as a mechanism for the specification of
neuronal networks. Nature, 264, 705-711.
Collewijn, H. (1981) The oculomotor system of the rabbit and its

plasticity. Studies of brain function, Springer, 5, 75-106.
Cowan, W. M., O'Leary, D. D. M. (1984) Cell death and process
elimination: The role of regressive phenomena in the
development of vertebrate nervous system. Medicine, science
and society, Wiley, 643-668.
Cynader, M. (1979) Competitive interactions in postnatal
development of the kitten's visual system. In Developmental
neurobiology of vision (Freeman, R.D. ed.), pp. 109-120,
Plenum Press, New York.
Cynader, M., Harris, L. R. (1980) Eye movement in strabismic cats.
Nature (Land.) 286, 64-65.
Cynader, M., Hoffmann, K.-P. (1981) Strabismus disrupts binocular
convergence in cat nucleus of the optic tract. Dev. Brain
Res. 1, 132-136.
Duersteler, M.R., Wurtz, R. H., Yamasaki, D.S. (1986) Pursuit and
OKN deficits following ibotenic acid lesions in the medial
superior temporal area (MST) of monkey. Soc. Neurosci.
Abstr., 12, 1182.
Fawcett, J. W., O'Leary, D. D. M. (1985) The role of electrical
activity in the formation of topographic maps in the nervous
system. Trends Neurosci., 8, 201-206.
Hebb, D. 0. (1949) Organization of behaviour, Wiley, New York.
Hoffmann, K.-P. (1982) Cortical versus subcortical contributions to
the optokinetic reflex in the cat. Functional basis of ocular
motility disorders, Pergamon Press, Oxford, 303-310.
Hoffmann, K.-P. (1983a) Control of the optokinetic reflex by the
nucleus of the optic tract in the cat. Spatially oriented
behavior, Springer, New York, 135-153.
Hoffmann, K.-P. (1983b) Effects of early monocular deprivation on
visual input to cat nucleus of the optic tract. Exp Brain
Res., 51, 236-246.
Hoffmann, K.-P. (1986) Visual inputs relevant for optokinetic
nystagmus in mammals. Progress in Brain Research, Vol. 64,
Elsevier, 75-84.
Hoffmann, K.-P. (1987) The influence of visual experience on the
ontogeny of the optokinetic reflex in mammals. In: Imprinting
and cortical plasticity, Wiley, New York.
Hoffmann, K.-P., Distler, C. (1986) The role of direction selective
cells in the nucleus of the optic tract of cat and monkey
during optokinetic nystagmus. Adaptive processes in visual and
oculomotor systems, Pergamon, Oxford, 261-266.
Hoffmann, K.-P., Distler, C., Erickson, R.G., Mader, W. (1987)
Physiological and anatomical identification of the nucleus of
the optic tract and dorsal terminal nucleus of the accessory
optic tract in monkeys. Exp. Brain Res, in press.
Hoffmann, K.-P., Huber, H. P. (1983) Responses to visual
stimulation in single cells in the nucleus of the optic tract
(NOT) during optokinetic nystagmus (OKN) in the awake cat.
Soc. Neurosci. Abstr., 9, 1048.
98 K.-P. HOFFMANN
Hoffmann, K.-P., Schoppmann, A. (1981) A quantitative analysis of

the direction specific response of neurons in the eat's nucleus
of the optic tract. Exp. Brain Res., 42, 146-157.
Hoffmann, K.-P., Stone, J. (1985) Retinal input to the nucleus of
the optic tract of the cat assessed by antidromic activation of
ganglion cells. Exp. Brain Res., 59, 395-403.
Hubel, D.H. & Wiesel, T. N. (1970) The period of susceptibility
to the physiological effects of unilateral eye closure in
kittens. J. Physiol. (Land.) 206, 419-436.
Lannou, J., Cazin, L., Precht, W., LeTaillanter, M. (1984)
Responses of prepositus hypoglossi neurons to optokinetic and
vestibular stimulations in the rat. Brain Res., 301, 39-45.
Magnin, M., Courjon, J. H., Flandrin, J. M. (1983) Possible
visual pathways to the cat vestibular nuclei involving the
nucleus prepositus hypoglossi. Exp. Brain Res., 51, 298-303.
Markner, C., Hoffmann, K.-P. (1985) Variability in the effects of
monocular deprivation on the optokinetic reflex of the
non-deprived eye in the cat. Exp. Brain Res., 61, 117-127.
Oyster, C. W., Barlow, H. B. (1967) Direction selective units in
rabbit retina: distribution of preferred directions. Science,
155, 841-842.
Precht, W. (1982) Anatomical and functional organisation of
optokinetic pathways. Functional basis of ocular motility
disorders, Pergamon, Oxford, 291-302.
Precht, W., Montarolo, P. G., Strata, P. (1980) The role of the
crossed and uncrossed retinal fibers in mediating the
horizontal optokinetic nystagmus in the cat. Neurosci. Lett.,
17, 39-42.
Schiller, P. H., Malpeli, J. G. (1977) Properties and tectal
projections of monkey retinal ganglion cells. J.
Neurophysiol. 40, 428-445.
Schoppmann, A. (1981) Projections from areas 17 and 18 of the
visual cortex to the nucleus of the optic tract. Brain Res.,
223, 1-17.
Simpson, J. I. (1984) The accessory optic system. Ann. Rev.
Neurosci., 7, 13-41.
Stone, J., Fukuda, Y. (1974) Properties of cat retinal ganglion
cells: a comparison of W-cells with X- and Y-cells. J.
Neurophysiol., 37, 722-748.
Vanegas, H. (1984) Comparative neurology of the optic tectum,
Plenum Press, New York.
Wickelgren, B. G., Sterling, P. (1969) Influence of visual cortex
on receptive fields in the superior colliculus of the cat. J.
Neurophysiol., 32, 16-23.
Wurtz, R. H., Albano, J. E. (1980) Visual-motor function of the
primate superior colliculus. Ann. Rev. Neurosci., 3,
189-226.
8
OCULAR MOTOR PHENOMENA IN
INFANTILE STRABISMUS
Asymmetry in optokinetic nystagmus and pursuit, latent nystagmus, and
dissociated vertical divergence
GUNTRAM KOMMERELL
ABSTRACf
The so-called infantile strabismus syndrome consists of (1) strabismus, (2) a

defect of temporally-directed pursuit and optokinetic tracking in monocular viewing, and
(3) latent nystagmus (LN). The following causal relationship between these three
phenomena is suggested. Firstly, infantile strabismus impairs the development of
binocularity in the visual cortex. Secondly, the reduced binocularity prevents oculomotor
maturation: The nasal-temporal asymmetry in pursuit and optokinetic nystagmus (OKN)
that is a normal feature in the first few months of life remains as a permanent defect.
Finally, the asymmetry in the smooth tracking systems produces LN.- The impairment of
temporally-directed tracking cannot be due to a defect in the retino-cortical pathway
because the patients can perceive temporally-directed object motion, and distinguish
various velocities. Rather, the visual cortex seems to lack the ability to transmit
temporally-directed object motion to the premotor structures of the brainstem if
binocularity has failed to develop in the first months of life. - The nystagmus of patients
with infantile strabismus often has a gaze-paretic component. The pathophysiological
relationship of this component with the above-mentioned three signs of infantile
strabismus is not yet clear. - Dissociated vertical divergence (DVD) is another
phenomenon that frequently occurs together with early onset strabismus. As is the case in
LN, DVD also depends on the balance of inputs coming through the right and left eyes.
Otherwise, nothing is known of the pathogenesis of DVD.
INTRODUCTION
The occurrence of early onset strabismus, asymmetry in pursuit and optokinetic

nystagmus, latent nystagmus, and dissociated vertical divergence is highly correlated
(Doden 1961; Keiner and Roelofs 1955; Kornhuber 1960; Loewer-Sieger 1962; Mein
1983; Nicolai 1959; Roelofs 1928; Schor 1983). Therefore, the combination of these
phenomena has been defined as a syndrome, the so-called "congenital squint syndrome"
(Lang 1968). Because strabismus is rarely present at birth and usually becomes manifest
during the first six months of life, the term "infantile strabismus syndrome" may be more
appropriate.
99
100 G. KOMMERELL
Although a common cause for all four phenomena should be given consideration,
a causal interdependence between them appears to be more likely. Indeed, a wealth of
evidence indicates that strabismus is the primary abnormality which subsequently leads to
an asymmetry in the pursuit and OK systems, and it is a plausible hypothesis that the
asymmetry in these smooth tracking systems ultimately leads to LN. Less clear is the
causal relationship between DVD and strabismus, but DVD also seems to occur as a
consequence of the strabismus.
ASYMMETRY IN PURSUIT AND OPTOKINETIC NYSTAGMUS
The asymmetry in pursuit and OKN is defined as a reduction of gain for uniocular
stimuli directed to the temporal side, while the gain for nasally-directed stimuli is normal
or only moderately reduced. Healthy infants show such asymmetry in the smooth
tracking systems (Atkinson 1979; Atkinson and Braddick 1981; Hainline et al. 1984;
Naegele and Held, 1982), but the asymmetry disappears by about six months of age if
signs of normal binocularity appear (Atkinson 1979; Naegele and Held 1982). In adults,
a slight nasally-directed preponderance is only observed if optokinetic stimulation is
confined to the temporal hemifield; this asymmetry is counterbalanced by a temporally-
directed preponderance of the nasal hernifield (Ohrni et al1986).
The reduced, though not necessarily absent (Flynn et al 1984, Sorsby 1931),
binocularity caused by strabismus could prevent maturation of the smooth tracking
systems (van Hof- van Duin 1978). This hypothesis is supported by the persistence of
nasal-temporal asymmetry in each eye in cats which had been deprived of binocular
vision by unilateral lid suture early in life (Hoffmann 1979; van Hof- van Duin 1976).
The crucial factor in producing asymmetry in pursuit and OKN appears to be the
reduced binocularity rather than monocular or binocular deprivation. Certainly as far as
strabismic humans are concerned, amblyopia is not a prerequisite for the asymmetry in
the smooth tracking systems (Mohn et al1986; Schor and Levi 1980; Tychsen et al1985;
Tychsen and Lisberger 1986; van Hof- van Duin and Mohn 1986). Cats rendered
exotropic, but not amblyopic, by early surgery, show a reduced OKN, predominantly of
the temporally-directed slow phases (Cynader and Harris 1980). In monkeys an adequate
model still seems to be lacking, although deprivation of binocular vision by alternating lid
suture may be promising (Tusa et al. 1987). In the monocularly deprived (i.e.,
amblyopic) monkey, nasal-temporal asymmetry in the OKN is present when the deprived
eye is stimulated with a rotating drum, whereas stimulation of the non-deprived eye
results in normal OKN (Sparks et al. 1986).
The neural mechanism of how the loss of binocularity leads to the asymmetry in
the smooth tracking systems could be that the cortical projection to the nucleus of the
optic tract (NOT) has not developed its normal function (Hoffmann, this symposium).
The NOT which is located in the pretectum is an important relay station of the optokinetic
system that receives direct input from the contralateral eye and indirect input from both
eyes via both occipital lobes (Hoffmann 1982, 1983). Lesions of the visual cortex
drastically reduce the optokinetic response to temporally-directed motion under conditions
of monocular viewing in the cat (Hoffmann 1982; Strong et al. 1984) and in the monkey
(Zee 1986). Dark rearing of cats also results in asymmetry in the smooth tracking systems
and may be equivalent to a surgical lesion of the visual cortex (Harris and Cynader 1981;
van Hof - van Duin 1978). The relative preservation of responses to nasally-directed
motion appears to be due to the direct connection from the retina to the contralateral NOT
both in the cat (Hoffmann 1982) and in the monkey (Hoffmann and Distler 1986). In the
normal adult human, the subcortical projection alone seems to be insufficient to drive the
NOT, as most cortically blind patients do not show any optokinetic response (Jung and
Kornhuber 1964). But the relative preservation of responses to nasally-directed stimuli in
patients with incomplete bilateral occipital lobe destruction (Mehdorn 1982) could be due
OCULAR MOTOR PHENOMENA IN INFANTILE STRABISMUS 101
to remnants of the subcortical projection to the NOT which might have been released from
cortical control.
Although a cortical defect of binocularity induced by infantile strabismus may be
responsible for the impairment of temporally-directed tracking, this defect does not imply
difficulties in motion perception. Indeed, patients are able to differentiate between various
stimulus velocities. We ascertained this in a patient with infantile esotropia who had a
marked asymmetry in the smooth tracking systems (Fig. 1). Using optokinetic stimuli
and Stevens' (1957) magnitude estimation, the patient could clearly distinguish between
fast and slow slip velocities of the retinal image. The performance was the same
regardless of whether the stimuli were nasally or temporally-directed (Fig. 2A and B).
Similarly, there was good velocity discrimination independent of whether the optokinetic
stimulus was applied in the nasal or temporal hemifield (Fig. 2C).
LE stimulation
temp. ~I n j a .temp.l +fit.- nas.
~- ~/~\.
./ ':1'\(:/V
N/"1;]~o·
, L
RE stimulation
nas. ~I temp.
1\J\ i\~'\ :\]~o·

nas. I~ temp. Fig. 1. Full-field temporally-
directed optokinetic stimulation
of either eye at 60 deg/s does not
I I \ .' I
overcome the nasally-directed
'I I • ~ l slow phases of the LN in patient
\J ~ CW130360. Nasally-directed
L___l stimulation evokes a strong
1s OKN. DC-Electro-oculogram.
This result is compatible with recent work conducted by Tychsen and Lisberger
(1986) who presented a nasally or temporally moving single target to patients with
infantile strabismus whilst they were fixating a central stationary target (technique of
McKee and Welch 1985). The ability to discriminate differences in velocity was normal
when nasally and temporally-directed motion were considered separately. Only when the
patients compared target speed in the two directions, did they judge temporally-directed
stimuli to be slightly slower than nasally-directed stimuli. The authors regard this
perceptual asymmetry as an indication of a defect in the cerebral pathways responsible for
velocity perception, but an alternative interpretation is possible. Patients may have
underestimated temporally-directed stimuli because of adaptation to unidirectional slip of
the retinal image in every-day life: their "latent" nystagmus may have been partly manifest
under natural viewing conditions, i.e., these patients may have had so-called manifest
latent nystagmus (Dell'Osso et al1979).
Schor and Levi (1980) measured the contrast sensitivity function for perceived
motion in strabismus patients with nasal-temporal asymmetry of the OKN. Contrast
thresholds for nasally and temporally-directed stimulus movements were the same. This
finding along with the patients' unimpaired ability to distinguish between fast and slow
slip velocities in both horizontal
102 G. KOMMERELL
nystogmus w•th .l•xot•on·· of left eye

slow phose wloc1ty 1'/sec
nystagmus w1lh .. flxOIIon" of lfoft eye
nystagmus with .. f•~at•on" of right r;;ye
slow phase velocity 1'/sec
slow phase velOcity 6'/sec
i
-; 150
0 0
X @ X <I <I <I
0
li> X 0 0 0 E 100
-~
Q X
~ <I
0 0
~
·o
l!i> 1i> ijSO <I~IJJ>
......~ modulus
X X
X X
\modulus x 0 '-modulus ~
<l <I ..,
X li>
<I
&0
20 40 60 20 40 60 80 100"/sec
20 40 60 80 100"/sec
stripe-velocity in front of right eye stripe-velocity in front of left eye stripe-velocity in front of left eye
from nasal to temporal
x = from temporal to nasal x = from temporal to nasal
o = from nasal to temporal 0 = from nasal to temporal
• stimulus in ~~ hemifield
<l stimulus in nasal hemifield
Fig. 2. Unimpaired velocity perception of contours slipping across the retina.

The patient also represented in Fig. 1 was asked to look monocularly at a
fixation point in the middle of a full-field optokinetic stimulus. Under this
condition his eyes were nearly stationary, except for a persisting LN, as
indicated on the top of each panel. The stripes were moved either nasally or
temporally at various velocities in random order. Exposure time was 10 s.
The patient had to estimate the velocity relative to a "modulus" which was
presented three times at the beginning of the test series for each eye (A and
B). In (C), only one hemifield of the left eye was stimulated. Velocity
estimation was equally good independent of whether the stimulus was
presented in the temporal or nasal hemifield.- The method is similar to the one
employed by Komer and Dichgans (1967) in normal subjects.
directions argues against a defect in the retino-cortical pathway which could be
responsible for the asymmetry in the smooth tracking systems. Rather, the visual cortex
seems to loose the ability to transmit temporally-directed object motion to the premotor
structures of the brainstem if binocularity fails to develop in the first few months of life.
LATENT NYSTAGMUS (LN)
LN is defined as a jerk nystagmus whose rapid phases are directed to the side of
the visually dominant eye. With the left eye occluded the slow phases are directed to the
left, and with the right eye occluded the slow phases are directed to the right. We
suggested the hypothesis that the asymmetry in pursuit and OKN typically present in
patients with infantile strabismus might be the cause of LN (Kommerell 1978; Kommerell
and Mehdorn 1982). According to this hypothesis, the nasally-directed vector of the
smooth tracking systems would be preponderant, even in the presence of stationary
patterns. This directional preponderance could drive the slow phases of LN if the visual
input is unbalanced in favour of one eye. Total occlusion of one eye causes the greatest
imbalance in visual input, but as most patients with LN also have strabismus, part of the
directional preponderance becomes manifest with spontaneous suppression of the

squinting eye. This results in the above mentioned "manifest latent nystagmus"
(Dell'Osso et al. 1979; Kestenbaum 1961; Roelofs 1928). However, with binocular
viewing, the well-functioning nasally-directed smooth tracking systems of both eyes
complement each other and largely prevent drifting of the eyes.
Stimulation of LE Stimulation of RE
on~ -ttt-7t •
Fig. 3. Asymmetry in the
OKN in 6 patients with
LN. A full-field stimulus
with seven-degree wide
stripes was presented
monocularly for 15 s. The
average velocity of the ten
fastest slow phases that
occurred during this time
was taken as the response.
The velocity of the slow
0
E 100
phases of LN was
-~ measured when the patient
~ looked at stationary stripes
~
(response to zero
·o
ijSO velocity). The responses
0
E 100
to temporally-directed
-~ stimulus motion (filled
~ circles) are much weaker
~
than the responses to
·o
ijSO nasally-directed stimulus
motion (open circles), but
the responses to nasally-
directed stimulus motion
are also subnormal,
particularly at stimulus
velocities above 20° /s.
Responses with a gain of
1.0 would have appeared
on a 45-degree-line, either
above or below the zero
line. Patient codes:
AZ020874, UD110671,
0 15 30 60 90 120 0 15 :l) 60 90 120 KH170470, MK201072,
I I SP050469, AS291070.
LN LN The patient represented in
the uppermost panel is the
Stimulus velocity [•Js] same as in Fig. 4.
104 G. KOMMERELL
There are two analogues of LN in which a directional preponderance of the

smooth tracking systems may also cause spontaneous nystagmus in humans. The first is
downbeat nystagmus (Zee et al. 1974), and the second is a lesion of the cerebral
hemispheres which leads to slow phases directed to the contralateral side (Sharpe et al.
1979). Lesions of one cerebral hemisphere in the monkey also produce nystagmus (Tusa
1987) which may be due to directional preponderance of the smooth tracking systems.
The hypothesis that the asymmetry in the smooth tracking systems might be the
cause of LN is supported by a high correlation between the intensity of LN and the
magnitude of pursuit asymmetry. Tychsen and Lisberger (1986) found this in seven
patients by comparing the velocity of the slow phases of the LN with the eye acceleration
in response to a ramp stimulus, and with sinusoidal tracking.
Comparing LN with the asymmetry in full-field OKN, a clear correlation is not so
obvious (Mehdom and Kommerell1983; Fig. 3). But it may well be that the asymmetry
in pursuit or small-field OKN (Schor 1981; Schor and Levi 1980) is more relevant to the
LN than large-field OKN. Moreover, it is questionable whether a very high correlation
between the intensity of LN and the magnitude of the asymmetry in the smooth tracking
systems should be anticipated, because the intensity of LN can vary greatly (Sorsby
1931). Particularly during occlusion therapy for strabismic amblyopia, the nystagmus in
the viewing amblyopic eye can decrease considerably in a few days (unpublished
observation).
Cognitive factors also modify LN. For instance, the slow phases can be reversed
if the patient alternately occludes his right and left eye in total darkness (Hain et al1985;
Jung and Kornhuber 1964; Kommerell and Mehdom 1982; Schor 1981; van Vliet 1973).
Similarly, drift bias of the eyes in darkness depends on whether the patient used his right
or left eye for fixation prior to darkness (Schor and Westall 1984).
R
[ 10'
L
wiggle! stop! occlusion LE
Fig. 4. "Manifest latent" nystagmus evoked at will. The patient had an

infantile convergent strabismus with a slight amblyopia in the right eye. When
he was asked to "wiggle" his eyes, a strong left-beating nystagmus appeared.
It is remarkable that the rapid phases overshot the target. With both eyes
open, the patient was unable to evoke a nystagmus corresponding to a
dominance of his squinting right eye, i.e., he could not reverse the slow
phases at will. Right-beating nystagmus only appeared when the leading left
eye was occluded. Infrared reflection recording. The same patient is also
represented in the first panel of Fig. 3. The "manifest latent" nystagmus
evoked by this patient at will should not be confused with the well-known
"voluntary nystagmus" that occurs in otherwise normal subjects. Voluntary
nystagmus shows a high frequency (often above 10 Hz) of to and fro
saccades without intersaccadic intervals, and convergence is often
superimposed. In the patient depicted in Fig. 4, there was no change of
vergence or pupil size when he "wiggled" his eyes.
The cognitive influence is further exemplified by an unusual patient who was able
to manifest his LN at will. When asked to "wiggle" his eyes he evoked a strong
nystagmus with the fast phases directed to his leading left eye (Fig. 4). Binocular vision
was tested with Bagolini's striated glasses. In the every-day condition when the
nystagmus was not obvious (only occasionally was there a slight "manifest latent"
OCULAR MOTOR PHENOMENA IN INFANTILE STRABISMUS !OS
nystagmus), the patient reponed the streak of the squinting right eye to be faintly visible.
When he evoked the left-beating nystagmus at will, he completely suppressed the streak
of the squinting right eye. This result suggests that some residual binocularity was
available to the patient under normal circumstances, but that he was also able to fully
concentrate on his leading left eye. Concentrating on the left eye may have offset his
pursuit balance so much that a strong nystagmus appeared. Our patient required visual
contours to evoke the nystagmus at will. He could not produce it in darkness or when he
was looking at a contourless light screen.
A remarkable cognitive influence on LN in the dark was demonstrated by Abel et
al. ( 1986) in a patient who had been blind in the right eye from birth due to a
malformation of the eye. The right eye had been enucleated. In darkness, the patient's
"latent nystagmus" beat as though his right prosthesis were viewing (similar to the case
described by Ohm in 1928); he could, however, "look" with either eye at will, producing
the appropriate reversals of direction.
We do not believe that cognitive influences are at variance with our hypothesis
that LN is related to the asymmetry in the smooth tracking systems. One has to realize that
the smooth tracking systems of normal observers can also be influenced by cognitive
effort. Zikmund (1966) has shown this very convincingly: Trained subjects, studied in
total darkness, were able to evoke OKN by visual imagery of moving stimulus patterns.
Nobody would deny that OKN is a visual oculomotor phenomenon because it can be
evoked also in the dark. An analogous argument should apply to LN.
Summing up the arguments presented so far, we have suggested the following
causal relationship: Infantile strabismus impairs the development of binocularity in the
visual cortex. The reduced binocularity prevents oculomotor maturation in that the nasal-
temporal asymmetry in the smooth tracking systems which is a normal feature in the first
few months of life remains a permanent defect. Finally, the asymmetry in the smooth
tracking systems produces LN.
The reverse relationship suggested by Tychsen and Lisberger (1986), that LN
might constitute a tonic drive which leads to convergent strabismus appears unlikely to
us. It has to be borne in mind that the drift directed to the nose refers only to the viewing
eye. The drift is always conjugate, i.e., at the same time, the non-viewing eye drifts
temporally. It is hard to see how such a conjugate drift could lead to convergent
strabismus. Moreover, asymmetry in the smooth tracking systems and LN also occur in
patients who have had divergent strabismus from infancy (Roelofs 1928, present writers
experience).
Although early deprivation of binocular vision appears to be an important
pathogenetic factor of optokinetic asymmetry and LN, a lack of binocular vision is not an
absolute determinant of LN, and, conversely, the presence of binocular vision does not
preclude LN absolutely. We have seen a patient (SW250672) with a unilateral hypoplasia
of the optic nerve who had congenital squint and very likely never had binocular vision.
Nevertheless, we found only a subtle nasal-temporal asymmetry in the OKN, and LN
was absent on careful ophthalmoscopic examination. On the other hand we have seen
exceptional cases with LN who had binocular vision with only slightly reduced stereopsis
(JK051067). These observations show that factors other than defective binocular vision
play a role in the pathogenesis of LN. These factors can compensate, or aggravate the
condition.
In many patients with infantile strabismus, the strength of the nystagmus
increases with lateral gaze, in a few it does not. This means that a gaze-paretic component
of the nystagmus can be associated with the latent component, but there is no strict
quantitative correlation between the two. The latent component depending on the visual
dominance of the right or left eye can be marked while the gaze-paretic component is
subtle, and vice versa. Clinical impression suggests that the gaze-paretic component
decreases during the first four years of life, but oculographic data are lacking. The
106 G. KOMMERELL
pathogenetic link between the gaze-paretic component and infantile strabismus is not yet
clear.
Two alternative hypotheses on the pathogenesis of LN have been proposed
recently. Lang (1982) suggests that LN may be due to a dominance of the nasal half of
the retina over the temporal half. However, this hypothesis fails to explain why there are
slow drifts towards the allegedly dominant half of the retina. We would rather expect
rapid eye movements, in analogy to acquired lesions where the preserved area of the
retina is typically reached by saccades. Dell'Osso et al. (1979) advanced the hypothesis
that switching the egocentric localization from one eye to the other may cause the eyes to
drift in patients with LN. This explanation appears rather unlikely to us because there are
no other conditions, such as pastpointing in patients with an eye muscle palsy, where a
change of egocentric localization is associated with a drift of the eyes.
DIS SOCIATED VERTICAL DIVERGENCE (DVD)
DVD occurs only rarely in subjects with otherwise normal binocularity, but is
much more frequent in patients with early onset strabismus. There is one similarity
between LN and DVD: Both of these ocular motor abnormalities depend on the balance of
visual inputs coming through the right and left eyes. A balance with real scales can be
used to illustrate the clinical findings in DVD (Kommerell and Mattheus 1984). If the
right eye is fixing a bright picture and the left eye is occluded, the balance of visual inputs
shifts strongly to the right eye. This imbalance forces the left eye up. In the classical
Bielschowsky test, the imbalance is decreased by placing a dark filter in front of the
fixing right eye, thus reducing the dominance of the input through the right eye and
allowing the left eye to move slightly down (Bielschowsky 1930). If the right eye is
completely occluded and the left eye takes over fixation, the imbalance is reversed to give
a dominance of the input through the left eye forcing the right eye up.
The balance of visual inputs can also be influenced by conscious effort. The more
the patient tries to resolve a difficult acuity task, the more the squinting eye is forced up
(Rlissmann 1986).
Obviously, scales are only a metaphor to demonstrate the phenomena of DVD.
They do not tell us very much about the neural mechanism which remains obscure. The
changes in vertical deviation on the alternate cover test are brought about by disjunctive
movements, and they are not combined with vertical nystagmus (Helveston 1980).
Therefore, we cannot see any relationship to a vertical asymmetry in the conjugate smooth
tracking systems as has been suggested by Tychsen and Lisberger (1986), although
preference for upward tracking does indeed occur in patients with DVD (Tychsen and
Lisberger 1986). Preference for downward tracking was found by Schor and Levi (1980)
in patients with strabismic amblyopia. It is possible, though not specifically reported, that
their patients had also DVD. We assume that asymmetries in the vertical smooth tracking
systems are most likely a coincident abnormality, not causally related to DVD.
DVD is probably not a variant of the so-called ocular tilt reaction (Westheimer and
Blair 1975) because the cyclorotation that accompanies the vertical deviation is opposite in
the two conditions. In DVD, there is excyclorotation in the upward moving eye, whereas
in the ocular tilt reaction, there is incyclorotation in the upward moving eye.
REFERENCES
Abel LA, Grossman G, Dell'Osso LF, Thurston SE, Daroff RB (1986) Abstract. In:
Thurston SE (ed) Ocular motor minisymposium. Neuro-ophthalmology 6: 137-138
Atkinson J (1979) Development of optokinetic nystagmus in the human infant and
monkey infant. In: Freeman RD (ed) Developmental neurobiology of vision.
Plenum, New York, pp 277-287
Atkinson J, Braddick 0 (1981) Development of optokinetic nystagmus in infants: An
indicator of cortical binocularity? In: Fisher DF, Monty RA, Senders JW (eds) Eye
movements: Cognition and visual perception. Erlbaum, Hillsdale, NJ, pp 53-64
Bielschowsky A (1930) Die einseitigen und gegensinnigen ("dissoziierten")
Vertikalbewegungen der Augen. Albrecht von Graefes Arch Ophthalmol125: 493-
553
Cynader M, Harris L (1980) Eye movement in strabismic cat. Nature 286:64-65
Dell'Osso LF, Schmidt D, Daroff RB (1979) Latent, manifest latent and congenital
nystagmus. Arch Ophthalmol97: 1877-1885
Doden W (1961) Latenter Nystagmus bei Strabismus concomitans alternans. Ber Dtsch
Ophthalmol Ges 63: 486-490
Flynn IT, Pritchard C, Lasley D (1984) Binocular vision and OKN asymmetry in
strabismic patients. In: Reinecke RD (ed) Strabismus II, Proceedings of the fourth
meeting of the international strabismological association, Grune & Stratton,
Orlando, pp 35-43
Hain TC, Kelman SE, Zee DS (1985) Pursuit and saccade asymmetries in latent
nystagmus. Soc Neuroscience Abstract
Hainline L, Lemerise E, Abramov I, et al (1984) Orientational asymmetries in small-field
optokinetic nystagmus in human infants. Behav Brain Res 13: 217-230
Harris LR, Cynader M (1981) The eye movements of the dark-reared cat. Exp Brain Res
44:57-70
Helveston EM (1980) Dissociated vertical deviation. A clinical and laboratory study.
Trans Amer Ophthalmol Soc 78: 734-779
Hoffmann KP (1979) Optokinetic nystagmus and single cell responses in the nucleus
tractus opticus after early monocular deprivation in the cat. In: Freeman RD (ed)
Developmental neurobiology of vision. Plenum Press, New York, pp 63-73
Hoffmann KP (1982) Cortical versus subcortical contribution to the optokinetic reflex in
the cat. In: Lennerstrand G, Zee DS, Keller EL (eds) Functional basis of ocular
motility disorders, Pergamon Press, Oxford, pp 303-310
Hoffmann KP (1983) Neuronal responses related to optokinetic nystagmus in the eat's
nucleus of the optic tract. In: Fuchs A, Becker W (eds) Progress in oculomotor
research, Elsevier North Holland, New York, pp 443-454
Hoffmann KP, Distler C (1986) The role of direction selective cells in the nucleus of the
optic tract of cat and monkey during optokinetic nystagmus. In: Keller EL, Zee DS
(eds) Adaptive processes in visual and oculomotor systems. Pergamon Press,
Oxford, pp 261-266
Jung R, Kornhuber HH (1964) Results of electronystagmography in man: The value of
optokinetic, vestibular, and spontaneous nystagmus for neurologic diagnosis and
research. In: Bender MB (ed) The ocular motor system. Harper and Row, New
York, pp 428-488
Keiner GBJ, Roelofs CO (1955) Optomotor reflexes and nystagmus. In: Niihoff M (ed)
La Haye, Vol I, p 224. Cited after Godde-Jolly D, Larmande A (1973) Les
nystagmus, Vol!. Masson and Cie, Paris, Chap VIII, p 632
Kestenbaum A (1961) Clinical methods of neuro-ophthalmologic examination. Grone and
Stratton, New York, London, p 366
108 G. KOMMERELL
Kommerell G (1978) Beziehungen zwischen Strabismus und Nystagmus. In: Kommerell

G (ed) Augenbewegungsstorungen, Neurophysiologie und Klinik, Bergmann,
Miinchen, pp 367-373
Kommerell G, Mattheus S (1984) Reversed fixation test (RFT); a new tool for the
diagnosis of dissociated vertical deviation (DVD). In: Reinecke RD (ed) Strabismus
ll. Grune & Stratton, Orlando, pp 721-728
Kommerell G, Mehdorn E (1982) Is an optokinetic defect the cause of congenital
nystagmus? In: Lennerstrand G, Zee DS, Keller EL (eds) Functional basis of
ocular motility disorders, Pergamon Press, Oxford, pp 159-167
Korner F, Dichgans J (1967) Bewegungswahrnehmung, optokinetischer Nystagmus
und retinale Bildwanderung. Albrecht v. Graefes Arch klin exp. Ophthalmol174:
34-48
Kornhuber HH (1960) Uber Begleitschielen und latenten Nystagmus aus neurologischer
Sicht. In: Reiser KA (ed) Sitzungsbericht 102. Versammlung Verein Rheinisch-
Westfalischer Augenlirzte, pp 45-48
Lang J (1968) Squint dating from birth. First International Congress of Orthoptists.
Kimpton, London, pp 231-237
Lang J (1982) A new hypothesis on latent nystagmus and on the congenital squint
syndrome. In: van Balen AThM, Houtman W A (eds) Strabismus Symposium
Amsterdam 1981. Doc Ophthalmol Proc Ser 32:83-88
Loewer-Sieger DH (1962) Amblyopie. Een Studie over de Kenmarken en de
Behandeling. J Ruysendaal, Amsterdam
McKee SP, Welch L (1985) Sequential recruitment in the discrimination of velocity. J
Opt Soc Am A 2:243-251
Mehdorn E (1982) Nasal-temporal OKN-asymmetries after bilateral occipital infarction in
man. In: Lennerstrand G, Zee DS, Keller EL (eds) Functional basis of ocular
motility disorders, Pergamon Press, Oxford, pp 321-324
Mehdorn E, Kommerell G (1983) Beziehungen zwischen latentem Nystagmus,
asymmetrischem optokinetischem Nystagmus und defektem Binokularsehen.
Fortschr Ophthalmol80: 281-283
Mein J (1983) The asymmetric optokinetic response. Br Orthoptic J 40: 1-3
Mohn G, Sireteanu R, van Hof- van Duin J (1986) The relation of monocular optokinetic
nystagmus to peripheral binocular interactions. Invest Ophthalmol Vis Sci 27: 565-
573
Naegele JR, Held R (1982) The postnatal development of monocular optokinetic
nystagmus in infants. Vision Res 22: 341-346
Nicolai H (1959) Differenzen zwischen optokinetischem Rechts- und Linksnystagmus bei
einseitiger Schielamblyopie. Klin Mbl Augenheilk 134: 245-250
Ohm J (1928) Der latente Nystagmus im Stockdunkeln. Archiv fiir Augenheilkunde 99:
417-437
Ohmi M, Howard IP, Eveleigh B (1986) Directional preponderance in human optokinetic
nystagmus. Exp Brain Res 63: 387-394
Roelofs CO (1928) Nystagmus latens. Archiv fiir Augenheilkunde 98: 401-447
Riissmann W, Albrecht J (1986) Aufmerksamkeitsverlagerung und dissoziiertes
Hohenschielen. Klin Monatsbl Augenheilk 188: 245-247
Schor CM (1981) Directional anisotropies of pursuit tracking and optokinetic nystagmus
in abnormal binocular vision. In: Fender D, Cooper B (eds) Ocular motor
symposium. Cal-Tech Press, Pasadena, Cal, pp 83-104
Schor CM (1983) Subcortical binocular suppression affects the development of latent and
optokinetic nystagmus. Am J Optom Physiol Opt 60: 481-502
Schor CM, Levi DL (1980) Disturbances of small field horizontal and vertical optokinetic
nystagmus in amblyopia Invest Ophthalrnol Vis Sci 19: 668-683
Schor CM, Westall C (1984) Visual and vestibular sources of fixation instability in
amblyopia Invest Ophthalmol Vis Sci 25: 729-738
OCULAR MOTOR PHENOMENA IN INFANTILE STRABISMUS !09
Sharpe JA, Lo AW, Rabinovitch HE (1979) Control of the saccadic and smooth pursuit
systems after cerebral hemidecortication. Brain 102: 387-403
Sorsby A (1931) Latent nystagmus. Brit J Ophthalmol15: 1-18
Sparks DL, Gurski MR, Mays LE, Hickey TL (1986) Effects of long-term and short-
term monocular deprivation upon oculomotor function in the Rhesus monkey. In:
Keller EL, Zee DS (eds) Adaptive processes in visual and oculomotor systems.
Pergamon Press, Oxford, pp 191-197
Stevens S (1957) On the psychophysical law. Psycho! Rev 64: 153-184
Strong NP, Malach R, Lee P, van Sluyters RC (1984) Horizontal optokinetic nystagmus
in the cat. Recovery from cortical lesions. Dev Brain Res 13: 179-192
Tusa RJ, Smith CB, Herdman SJ (1987) The development of nystagmus in infant
monkeys following visual deprivation. Neuroscience abstract
Tychsen L, Hurtig RR, Scott WE (1985) Pursuit is impaired but the vestibulo-ocular
reflex is normal in infantile strabismus. Arch Ophthalmol103: 536-539
Tychsen L, Lisberger St G (1986) Maldevelopment of visual motion processing in
humans who had strabismus with onset in infancy. J Neuroscience 6: 2495-2508
van Hof- van Duin J (1976) Early and permanent effects of monocular deprivation on
pattern discrimination and visuomotor behaviour in cats. Brain Res 111: 261-276
van Hof- van Duin J (1978) Asymmetry in optokinetic nystagmus observed in normal
kittens and light deprived cats. In: Kommerell G (ed) Augenbewegungsstorungen,
Neurophysiologie und Klinik, Bergmann, Miinchen, pp 363-366
van Hof- van Duin J, Mohn G (1986) Monocular and binocular optokinetic nystagmus in
humans with defective stereopsis. Invest Ophthalmol Vis Sci 27: 574-583
van Vliet AGM (1973) On the central mechanism of latent nystagmus. Acta Ophthalmol
(Copenh) 51:772-781
Westheimer G, Blair SM (1975) The ocular tilt reaction- a brainstem oculomotor routine.
Invest Ophthalmol 14: 833-839
Yamazaki A (1979) Abnormalities of smooth pursuit and vestibular eye movements in
congenital jerk nystagmus. In: Shimizu K (ed) Ophthalmology, vol 2, Excerpta
Medica, Amsterdam, pp 1162-1165
Zee DS, Tusa RJ, Herdmann SJ, Butler PH, Gucer G (1986) The acute and chronic
effects of bilateral occipital lobectomy upon eye movements in monkey. In: Keller
EL, Zee DS (eds) Adaptive processes in visual and oculomotor systems. Pergamon
Press, Oxford, pp 267-274
Zee DS, Friendlich AR, Robinson DA (1974) The mechanism of downbeat nystagmus.
Arch Neurol30: 227-237
Zikmund V (1966) Oculomotor activity during visual imagery of a moving stimulus
pattern. Stud Psychol8: 254-272
9
PHASIC-TONIC ORGANIZATION OF ACCOMMODATION
AND VERGENCE
CLIFTON M. SCHOR
INTRODUCTION
Stereoscopic depth discrimination is stimulated by binocular

disparities that are computed from the two retinal images. This
computation relies upon a stable correspondence or linkage between
retinal regions, which when stimulated simultaneously, yield the
perception of identical visual directions. The task of the
oculomotor vergence system is to obtain and maintain a registration
of homologous portions of the two retinal images with the
retinotopic organization of binocular correspondence (horopter). To
facilitate this task, the vergence system responds both voluntarily
and reflexively to a variety of distance cues.
RANGE DEPENDENT STIMULI FOR VERGENCE
Large changes in vergence can be willfully initiated but they

are not sustained in response to perceived distance. Monee ular
depth cues such as changing size (Erkelens and Regan, 1986) and
binocular cues such as large-brief disparities (Ogle 1950;
Westheimer and Mitchell, 1969; Jones, 1980) evoke large transient
vergence response. These vergence responses will decay unless they
are refined and maintained by motor responses to small retinal image
disparities and blur. These two latter stimuli are nulled by the
negative feedback control systems of disparity vergence and optical
reflex accommodation respectively. These motor systems, which
refine and maintain binocular eye alignment, are frequently abnormal
in strabismus wherein vergence stimulated by accommodation is
excessive and vergence stimulated by disparity is lacking.
* This project was supported by grant #EYO 3532-06 from the

National E.¥e Institute of the National Institutes of Health,
Bethesda l'H. U.S. A.
111
112 C.M. SCHOR
PHASIC-TONIC ORGANIZATION OF ACCOMMODATION AND VERGENCE
Both disparity vergence and optical reflex accommodation have

phasic and adaptable tonic components. Hlasic or rapid responses of
these motor systems are characterized by reflex responses by
vergence to small disparities (<1-2 degrees), (Westheimer and
Mitchell, 1969)and by reflex responses by accommodation to small
errors of focus (<2 D)(Fincham, 1951). In addition to these rapid
reflex responses there are sustained or long-term responses which
are manifest as motor aftereffects that appear when the stimulus for
vergence or accommodation has been removed. For example, if
convergence is stimulated by 1-2 degrees for more than 15 seconds,
then when one eye is occl uied, the vergence response can continue
for several minutes in the absence of a disparity stimulus.
However, if the vergence stimulus is too brief to initiate
adaptation(< 15 seconds), then upon occlusion of one eye, the
response decays in 5-15 seconds (Schor, 1983). Similarly prolonged
accommodative aftereffects are manifest either with a pinhole pupil
or an empty field after stimulating 1-4 D of accommodation with
minus lenses (Schor, Kotulak and Tsuetaki, 1986). This accommodative
aftereffect is greatest if, after opening the accommodative loop,
the eye views a photopic or mesopic field. In darkness, or scotopic
illumination the accommodative aftereffect decays rapidly (in less
than 10-15 seconds) however it will return when the lighted field
reappears (Schor, Kotulak and Tsuetaki, 1986).
AFTEREFFECTS OF CROSS-LINK INTERACTIONS BETWEEN ACCOMMODATION AND

VERGENCE
In addition to the direct adaptation of accommodation to blur

and adaptation of vergence to disparity there are indirect
adaptation responses that result from the cross coupling between
these two motor systems. Vergence aftereffects can result from
stimulation of accommodative vergence, and likewise accommodative
aftereffects can result from stimulation of vergence accommodation
(Schor and Kotulak, 1986; Schor Kotulak and Tsuetaki, 1986).
INFLUENCE OF ADAPTATION UPON CROSS COUPLING BETWEEN ACCOMMODATION

AND VERGENCE
Adaptable states of accommodation and vergence have a direct

influence on accommodative vergence and vergence accommodation
respectively. 'Ihe AC/ A ratio is inversely related to the
adaptablility of accommodation and the CA/C ratio is inversely
related to the adaptability of vergence (Schor and Kotulak, 1986)
These inverse relationships are based upon two properties of both
accommodation and vergence. First, there are two separate processes
that are believed to underly accommodation, a fast (phasic) optical
reflex, component, which is followed serially by a slower (tonic)
adaptable component (Schor and Kotulak 1986). Servoanalytical
PHASIC-TONIC ORGANIZATION OF ACCOMMODATION AND VERGENCE 113
theory predicts that the linear sun of the responses by these two
mechanisms accounts for the total accommodative response (Kotulak
and Schor 1986). Similarly two separate processes are proposed to
underly vergence control, a fast (phasic) disparity vergence
component followed serially by a slower (tonic) adaptable component
(Schor 1983). The linear sun of the responses of these phasic and
tonic mechanisms is believed to account for the total vergence
response (Schor 1983). The second property underlying the effect of
adaptation on cross coupling interactions is that only phasic
components stimulate accommodative vergence and vergence
accommodation whereas the adaptable processes do not stimulate
cross-link interactions. This is demonstrated by aftereffects of
accommodation and vergence which decay independently of one another
in darkness (Kotulak and Schor, 1986). Additional evidence for this
comes from observed increases of the AC/ A and CA/ C ratios with the
speed or temporal frequency of the driving stimulus. Because
adaptation is a slow process, it responds more to low than high
temporal frequency stimuli. Consequently the adaptable tonic
response decreases with temporal frequency while phasic responses
ani their cross-link stimulation increase. At lower temporal
frequencies, more of the total response is controlled by the
adaptable component, which results in lower cross-link interactions.
At higher frequencies, adaptation is unable to keep pace with the
stimulus and the fast process controls the response. If only the
phasic process interact between the two motor systems then the
cross-link interactions become temporal frequency dependent. Thus,
both accommodative vergence and vergence accommodation are mainly
associated with rapid changes in their respective stimuli.
A MODEL OF DYNAMIC INTERACTIONS BETWEEN ACCOMMODATION AND VERGENCE
+
TARGET ACCOMMODATION
DISTANCE
VERGENCE
+_
Input Fast Cross Tonic Adaptation Plant Output

Element Links with limited input
Figure 1. System model representation of synchenetic interactions

between accommodation and vergence. The two motor systems are
interconnected at points in their feed-forward paths located between
phasic and tonic neural integrators.
114 C.M. SCHOR
These :rhasic and tonic characteristics of accommodation and

vergence are sunmarized in the model smwn in Figure 1. Cross
coupling between accommodation and vergence systems emerge after the
phasic components but prior to adaptable tonic elements of the two
systems. These cross-links merge with other stimuli at the
adaptable element of the receptive motor system. For example, the
phasic innervation to accommodation is proposed to merge with
disparity vergence innervation and cause aftereffects of the
vergence system. Similarly the phasic innervation of vergence is
proposed to merge with optical reflex innervation for accommodation
and cause aftereffects of accommodation. The serial arrangment of
phasic and tonic elements produces a reciprocal relationship. High
tonic activity reduces the need for phasic control of motor
responses. Consequently robust adaptation reduces both phasic
innervation and its input to cross-link interactions. Accordingly
cross coupling between accommodation and vergence becomes reduced
with their degree of adaptability.
IMBALANCED ADAPTATION OF ACCOMMODATION AND VERGENCE
Frequently, the amplitu:ies of the AC/ A and CA/C ratios are

inversely related. That is, one ratio is very high and the other is
very low. Invariably this reciprocity is associated with unequal
adaptation of the two motor systems (Schor and Tsuetaki 1987). In
cases where accommodation is more adaptable than vergence, there is
a high CA/C ratio and a low AC/ A ratio. The reverse occurs i f
vergence is more adaptable than accommodation. In these cases, a
high cross-link interaction need not originate from an unadaptable
motor system. It can become elevated by a robust adapter which
receives cross-link stimulation. For example, vergence aftereffects
can result from accommodative vergence. If vergence adaptation is
extremely robust compared to accommodative adaptation then the AC/A
ratio will appear to be elevated. HOwever, i f both accommodation
and vergence are equally adaptable, (either low or high) then both
the cross-link interactions described by the AC/ A and CA/C ratios
will have moderate values. This interpretation suggests that
abnormally high and low AC/ A and CA/ C ratios can result from an
imbalance between adaptation mechanisms of accommodation and
vergence (Schor, 1987).
CAN THE AC/A AND CA/C RATIOS BE CHANGED ?
In cases of extreme or abnormal interactions between

accommodation and vergence that are associated with imbalanced
adaptation, the extreme ratios may be corrected by restoring the
balance of adaptable accommodation and adaptable vergence. This can
be demonstrated temporarily be fatiguing the adaptability of
accommodation and vergence with repeatative ramp tracking tasks.
~CCOMMOOATIVE VERGENCE
C. H.
I
L
LEFT EYE
10 P. d.
RICHT [¥£
ACCOiol.
SliM.
<0. OS Hzl
15 s.c •.
~CCOMMOOATIVE VERGENCE
(, M.
I
l
LEfT [Y(
10 P.d.
•
RIGHT 0£
ACCOM.
<,I /M.
<0. OS Hzl
,__ _ _____,
15 Seca.
Figure 2. Accommoiative vergence responses to a 2D, 0. 05 Hz

monocular accommoiative step stimulus are illustrated prior to (top)
and following (bottom) the accommoiation ramp tracking task (2D at
0. 75 Hz for 4 minutes). The AC/ A ratio increased from <2 pd/D to 4
pd/D as a result of the ramp tracking task.
116 C.M. SCHOR
VERGENCE ~CCOMMODAT!DN
J. T.
I•o
lEFT EYE
P.d.
•
R!GHT EYE
ACCOM. I 1.5 D
STIH.
<0. 05 Hz)
15 s.c•.
VERGENCE ~CCOMMODAT!ON
J. T.
I
LEFT EYE l
8 P.d.
•
RIGHT EYE
r~.~o
ACCDM.
ST JM.
(Q_ 06 Hz)
Figure 3. Vergence accommodation responses to an 8 pd BO, 0. 05 Hz

disparity-step stimulus are illustrated prior to (top) and following
(bottom) the vergence ramp tracking task (5 pd BO at 0. 75 Hz for 4
minutes). The CA/C ratio increased from <0.2 D/MA to 0.6 D/MA as a
result of the ramp tracking task.
PHASIC-TONIC ORGANIZATION OF ACCOMMODATION AND VERGENCE ll7
Maptability of tonic accommodation becomes reduced following 4

minutes of accommodative ramp tracking (2D at 0.75 Hz) (Schor and
Tsuetaki, 1987). Figure 2 illustrates that once accommodative
adaptation has been fatigued, a low AC/ A ratio ( <2pd/D) can increase
by more than 100 percent (>4pd/D). Similarly, the duration and
magnitude of vergence aftereffects become reduced following 4
minutes of disparity ramp tracking (3 degrees at 0.75 Hz). Before
the tracking task vergence aftereffects of a 30 second stimulus can
last for up to 10 minutes. Following vergence tracking, vergence
aftereffects decayed quickly to the resting position in 5 seconds
while the vergence loop was opened by monocular occlusion. Similar
changes in vergence adaptation are obtained with convergence and
divergence tracking ramps. Figure 3 illustrates that once vergence
adaptation has been fatigued, a low CA/ C ratio ( < 0. 2D/MA) can
increase by 300 percent(> 0.7D/MA)(Schor and Tsuetaki, 1987).
While these fatigue effects on tre AC/ A and CA/C ratios are
only temporary, they suggest a possible means of altering the cross
link interactions by methods which will produce prolonged changes in
adaptability of accommodation and vergence. Such long term changes
have been reported by North and Henson (1982) in the form of
enhanced adaptability of vergence following a standard orthoptic
exercise program. If this therapeutic approach were to bring about
a balance in adaptation of accommodation and vergence it would
perhaps return extreme AC/A and CA/C ratios to within normal ranges.
NEUROPHYSIOLOGICAL CORRELATES OF PHASIC AND TONIC COMPONENTS OF THE

VERGENCE SYSTEM
Comparative studies in monkey of premotor neurones in the

mesencephalic reticular formation near the oculomotor nucleus reveal
cells that respond to pure vergence signals and that are independent
of versional eye movements (Mays 1984; Judge and Commings, 1986).
Firing rate of these cells is a monotonic function of vergence angle
or accommodation response. Evidence for two classes of cells in the
categories of phasic and tonic was obtained by comparing cell firing
rate as a function of vergence angle before and after adapting
animals to prism vergence. The firing rates of some cells were
reduced by an amount proportional to the vergence adaptation whereas
the firing rates of other cells remained mchanged by the prism
adaptation response (Mays and Tello, 1986; Morley Lindsey and Judge,
1986). Cells whose firing rate decreased with adaptation correspond
to the phasic category whereas cells whose firing rate was
unaffected by adaptation could receive the sum of both phasic and
tonic sources of innervation (Mays, Porter, Gamlin, and Tello,
1986). In addition, there is a subgroup of cells whose firing rate
is correlated with accommodative responses (Morley, Lindsey, and
Judge, 1986) Future research will reveal if there are also phasic
and tonic classes of cells that are associated with optical reflex
accommodation and adaptable tonic accommodation.
118 C.M. SCHOR
In addition to cells whose firing rate increases with vergence

angles, there is a class of burst cells which encode instantaneous
vergence velocity (Mays, Porter, Gamlin, and Tello, 1986). The
combined effect of burst and tonic cells is to produce gradual
changes in firing rate of medial rectus motoneurones during step
vergence responses (Mays and Porter, 1984). The burst cells could
improve the frequency response of the vergence system without
influencing its steady-state behaviour. The influence of a
velocity signal has been modeled as a bypass proportional gain
element in recent behavioral studies of the vergence system (Schor
and Kotulak, 1986). This organization of burst and tonic cells is
similar to that proposed for the saccadic control system even though
the kinematies of the two systems differs markedly.
IMPLICATIONS FOR STRABISMUS
In some classes of esotropia, the angle of strabismus increases

anomalously in response to corrective prism (Hallden 1952, Bagolini
1976). These anomalous movements indicate a negative prognosis for
cosmetic surgical correction (Bagolini 1986) and are screened for
clinically with the prism adaptation test (PAT) (Jampolsky 1971).
The process underlying anomalous movements is suggested by the
current investigation since vergence responses by esotropes to prism
bear a striking resemblance to the normal slow fusional processes
described above as prism adaptation. Both are evoked by disjunctive
prism and both are slower than disparity vergence responses.
Vergence adaptation or gradual changes in heterophoria can be

stimulated binocularly with disparity vergence and monocularly with
accommodative vergence. Either of these stimuli could evoke
anomalous movements in strabismus. Constant strabismics lack robust
fast fusional responses to disjunctive prism (Alpern and Hofstetter
1948, Pisano and Maraini 1966, Bagolini 1976) however they may still
exercise very small disparity vergence responses which become
integrated over time (0.5 -2 hrs) to produce a large adaptation of
tonic vergence. In addition to disparity vergence, accommodative
responses to lenses and perceived distance (proximity) can also
produce changed in the vergence phoria. This interaction is
exacerbated by a high accommodative/vergence ratio which when
combined with hypermetropia can percipitate an esotropia.
Adaptive responses of accommodation and vergence can also play

a major role in determining the magnitude of the AC/ A ratio. When
the AC/A ratio is abnormally high, such as in accommodative
esotropia, the current model suggests there is an imbalance between
adaptation of accommodation and vergence, where adaptation is
excessive for vergence and insufficient for accommodation. If this
prediction can be verified in some classes of esotropia, then a
therapeutic approach may be suggested which restores the balance of
adaptation between accommodation and vergence. These threaputic
procedures should either increase the weak adapter (North and Henson
1982) or reduced the excessive adapter (Schor and

Tsuetaki 1987).
I f these changes in adaptability are prolonged then the AC/ A ratio
could be modified in the treatment of esotropia. Some long term
changes in the AC/ A ratio have already been reported following
orthoptics therapy (Flom, 1960) and drug therapy (DFP) however the
influences of these procedures upon adaptability of accommodation
and vergence has yet to be examined.
REFERENCES
Al~rn, M. and Hofstetter, H.W. (1984). The effect of prism on

esotropia. A case report. Am. J. Optom. Arch. Am. Acad. Optom.,~,
80-91.
Bagolini, B. (1976). Part II: Sensoria-motorial anomalies in

strabismus (Anomalous movements). DJcumenta Ophthalmol. ,±1_, 23-41.
Bagol ini, B. , Zanasi, M. R. and Bol zani, R. ( 1986). Surgical

correction of convergent strabismus: its reltionship to prism
com~nsation. D:lcumenta Ophthalmol. ,62, 309-324.
Erkelens, C. and Regan, D. (1986). Human ocular vergence movements

induced by changing size and disparity. J. Physiol. ,379, 145-169.
Finchsm, E.F. (1951). The accommodative reflex and its stimulus.

Brit. J. Ophthalmol.,35, 381-393.
Flom, M. C. ( 1960). On the relationship between accommodation and

accommodative vergence. III. Effects of orthoptics. Am. J. Optom.
Arch. Am. Acad. Optom., 37, 619-632.
Hallden, U. (1952). Fusional phenomena in anomalous correspondence.

Acta. Ophthalmol.,37(suppl), 1-93.
Jampolsky, A. (1971). A simplified approach to strabismus diagnosis.

In Symposium on Strabismus. Trans New Orleans Academy Ophthal, CV
Mosby, st l.Duis, 34-92.
Jones, R. (1980). Fusional vergence: sustained and transient

components. Am. J. Optom. and Physiol. Optics., 57, 40-644
Judge, S.J. and Cumming, B.G. (1986). Neurons in the monkey midbrain
with activity related to vergence eye movement and accomodation. J.
Neurophysiol. ,.2.2_, 915-930.
Kotulak, J. and !:bhor, C.M. (1986). The dissociability of

accommodation from vergence in the dark. Invest. Ophthalmol. and
Vis. !:bi. ,27, 544-551.
Mays, L. E. ( 1984). Neural control of vergence eye movements:

convergence and divergence neurons in mid brain. J.
Neurophysiol. •21.• 1091-1108.
120 C.M. SCHOR
Mays, L. E. and Porter, J. ( 1984). Neural control of vergence eye

movements: activity of abducens and oculomotor neurons. J.
Neurophysiol. ,52, 743-761.
Mays, L.E. and Tello, C.A. (1986). Neurophysiological correlates of

convergence and its tonic adjustment. In Adaptive Processes in
Visual and Oculomotor Systems( eds. Keller EL and Zee, DS). Pergamon
Press, New York pp 143-149.
Mays, L.E., Porter, J., Gamlin, P. and Tello, A. (1986). Neural

control of vergence eye movements: Neurones encoding vergence
velocity. J • Neurophysiol. ,56, 1007-1 021.
Maraini, G. and Pasino, L. (1964). Variations in the angle of

anomaly and fusional movements in cases of small-angle convergent
strabismus with harmonious anomalous retinal correspondence. Brit.
J. Opthalmol. ,48, 439-443·
Morley, J.W., Lindsey, J.W. and Judge, S.J. (1986). Changes in

activity of brainstem near-response neurones induced by
prism-adaptation. Soc. Neurosci. Abstr. # 126.21,2_, 460.
North, R. and Henson, D. (1982). Effect of orthoptics upon the

ability of patients to adapt to prism-induced rnterophoria. Am.
Acad. Optom. and Physiol. Optics., 59, 983-986.
Ogle,K.N. (1950). Researchers in binocular vision. Philadelphia: WB

Saunders.
Schor, C.M. (1983). Fixation disparity and vergence adaptation, In

Vergence eye movements: basic and clinical aspects. ( eds. Schor,
C.M. and Ciuffreda, K.), Butterworths, Boston 465-516.
Schor, C.M., Kotulak, J, and Tsuetaki, T. (1986). Adaptation of

tonic accommodation reduces accommodative lag and is masked in
darkness. Invest". Ophthalmol. and Vis. Sci.,27, 186-193.
Schor, C.M. and Kotulak, J.C. (1986). Dynamic interactions between

accommodation and convergence are velocity sensitive. Vision
Res. ,26, 927-942.
Schor, C.M. and Tsuetaki, T. (1987). Fatigue of accommodation and

vergence modifies their mutual interactions. Invest. Ophthalmol.
and Vis. Sci.,28, in press.
Schor, C.M. (1987). Imbalanced adaptation of accommodation and

vergence produces opposite extremes of the AC/ A and CA/ C ratios. Am.
J. Optom. and Physio. Optics. ,64, in press.
Westheimer, G. and Mitchell, D. (1969). Tre sensory stimulus for

disjunctive eye movements. Vision Res. ,2_, 749-755.
10
CLINICAL ASPECTS OF VERGENT MECHANISMS
BRUNO BAGOLINI
Maddox suggested the existence of different types of

convergent movements: l)tonic convergence, 2)accommoda-
tive convergence, 3)fusional convergence. Furtherly, 4)
voluntary convergence and 5)"proximal convergence" were
recognized. Voluntary convergence is the only disjunc-
tive movement influenced by will power.
Most of these types of disjunctive movements may have

variable importance in various types of oculomotor imbal-
ance. They will be briefly reviewed, though in different
order, mainly underlining some new clinical aspects of
this old, but not fully understood subject.
l)Accommodation_through its association with conver-

gence plays an important role by contributing totally or
partially to determine the so-called "accommodative stra-
bismus" occurring frequently in clinical practice. The
AC/A ratio, instead, appears not to behave differently in
both the normal and the strabismic population (Hofstetlen
though in esotropia this ratio was found frequently high
(Parks) and in exotropia at times high (von Noorden) at
times low (Jampolsky).
The reason why only some hyperopic children develop

esotropia is poorly understood. Worth believed that
these children have a fusion defect that predisposes them
to convergent strabismus. It is reasonable to think that
a high AC/A ratio accompanied by hyperopia or a low AC/A
ratio accompanied by myopia adds a further predisposition
respectively to eso- or exotropia. Park's work throws
121
122 B. BAGOLINI
some light onto this problem. He found that average hy-

peropia in accommodative esotropia was +4.75Sph while in
patients with high AC/A ratio it was +2.25Sph. It should
be added that even a small amount of anisometria facili-
tates esotropia in high hyperopic children (roughly above
+4 Sph) through a weaker motor fusion and because of fre-
quent occurrence of amblyopia in the more hyperopic eye.
2)Proximal convergence does not appear to play any

role in strabismus and it occurs both in normal and stra-
bismic people.
3)Voluntary convergence has been recently discovered

to play an important role in developing convergent stra-
bismus in congenital nystagmus (or in sensorial nystagmus
developed for bilateral visual handicaps in early life).
It is well known that nystagmus decreases or is complete-
ly blocked when convergence is exerted for near vision or
by inducing convergence using base-out prisms in patients
with good binocular vision. Many patients looking at a
distance progressively converge by using progressively
strong base-out prisms while nystagmus gradually disap-
pears. This type of prism-induced nystagmus block be-
comes possible only when patients enjoy binocular vision
with good motor sensorial amplitude. Many children learn
to spontaneously block nystagmus by exerting a convergent
movement even when looking at distant vision (Adelstein &
Clippers). They fix with one eye and overconverge the con-
tralateral one; hence, they develop an esotropia that may
be permanent or, more rarely, intermittent. This is a
voluntary movement which is frequently learned rather
spontaneously to improve vision for distance by blocking
nystagmus. This is proven by the fact (Bagolini, Penne &
Zanasi) that voluntary convergence may be taught to block
nystagmus for distant vision by overconverging and becom-
ing esotropic in some form of congenital nystagmus.Figure
1 shows an eleven-year-old girl with congenital nystagmus
(A) to which voluntary convergence was taught. Her visum
acuity for distant vision was 4/10 with each eye separat~
ly, and 6/10 with both eyes open. Visual acuity of her
fixing right eye reached 8/10 while the contralateral eye
was squinting but the patient was seeing double (B). She
was then operated on to regain straight eyes; her visual
acuity after surgery reached 10/10 in binocular vision.
This type of treatment can be performed only in milder
CLINICAL ASPECTS OF VERGENT MECHANISMS 123
Fig. 1 A Fig. 1 B
Pr 1• .., D• op\.e ,.•

.. lot.•t "'-ounl or ,,. ,. ,.,.\.IC" COfi'IP• "'••llon
v z,,. '- ••1 ""Gl • or sv·• b• •""-'•
•• • Po•t.~•t. • ""• Alnijl a cr Sl,..O • ellll.le
"
••
so ....
lo
lO
10
·10
Fig. 2. The p re-operative prism compensation (~ ) is in-

dicated in the vert i cal axi s . In the h o rizont a l axi s a r e
repo r t ed Gl cas e s of eso t ropia going to the ones that had
n o compe n s ation at t h e lef t side , to the ones h a ving the
stronge r c omp e n sa t ion t o t h e igh
r t . The s ymb o l ( V ) in-
dicates the pre-operative angle, the symbol ( ~ ) i ndi-
cates the post-operative angle after recession of o ne
medial rectus. The line connecting the two symbols indi-
cates the a moun t o f surgical co rrecti on in diopte r s .
A s tatistical e v alu a tion has s hown a pos it ive co rrelat i on
(P . 001 ) b etwee n pri s m c ompensati o n a n d surgica l r esu l t s .
124 B.BAGOLINI
cases of nystagmus, with good binocular vision, and in

cooperative subjects. Further proof that the convergent
nystagmic block is a voluntary act is related to the fact
that esotropia in a few patients increases by increasing
visual demand. When these patients are asked to read pr~
gressively smaller letters of the optotype at a distance,
some patients with a nystagmic block strabismus increase
convergence to decrease nystagmus in accordance with the
visual demand requested. Strabismus increases according-
ly. The nystagmic block may, however, be acquired by any
mechanism determining convergence. Accommodative conver-
gence, provided by negative lenses, may also block nys-
tagmus for distant vision when fusion is interrupted by
occluding one eye and the patient freely converges under
the accommodative impulse (open loop). Therefore, if fu-
sion is maintained (closed loop) and convergence is not
exerted, negative lenses and accommodation are unable to
block nystagmus for distant vision. Fusional convergence
in patients with nystagmus and normal binocularity fre-
quently blocks nystagmus for distance if we make the pa-
tient overconverge in a closed loop by base-out prisms.
It is the convergent impulse, whatever its origin, that
appears to superimpose onto the nystagmic impulses exer-
ting a contractive effect on the medial recti, which may
nullify the weaker impulses of nystagmus. The shaking in
horizontal nystagmus should continue at the level of the
lateral recti which are apparently unaffected by the con-
vergent movement. The lateral recti very probably stop
the shaking because when the antagonistic medial recti
contract, there is an inhibitory effect (Sherington's La~
on the tonus of the lateral recti which involve nystagmic
impulses,too. The inhibiting effect on the nystagmus due
to convergence is, therefore, an effect which, at least
in part, is not peripheral but at mid-brain level. This
should be better investigated by electromiography.
4)Tonus is the first factor considered in Maddox'

classification. Its importance in eye misalignment is of
great relevance. The rest position of the eyes is more
or less one of orthophoria in normal subjects. This rest
position is, nevertheless, supported by a basic muscle
activity demonstrated by electromyography (Breinin) which
is responsible for the tonicity of the muscles. The or-
thophoria is supported by a balance between opposing
forces in the innervation of antagonistic muscles arising
from numerous sources only partly known. These antago-

nistic sources sum up algebraically leading usually to
orthophoria but an imbalance of these sources of tonus
leads to a variation of the rest position. This is the
cause of the various types of strabismus; unfortunately,
we have no way of measuring these tonic forces in a cli-
nically acceptable way.
5)Fusional movements are a third factor recognized

in Maddox' classification. They are elicited when suffi-
ciently equal images in the retina are formed over dis-
parate retinal elements. Disparate retinal elements bear
different spatial directional localization and bear a
certain fusional motorial value which increases with the
increase of retinal disparity. Corresponding retinal
elements have the same directional localization and 0 fu-
sional motorial value. The fusional movements are elici-
ted by a reflex process (psycho-optic reflex of Hofmann &
Bielschowsky). This reflex activity is evident if the
fusional stimuli fall inside (and to a certain extent,
outside) Panum's area. For stimulations of quite dispa-
rate retinal elements, however, attention and a voluntary
effort are necessary to bring the two retinal images over
relatively disparate retinal elements so as to elicit the
reflex part of the fusional movement. Fusional movements
are much slower than saccadic movements (see Burian in
von Noorden). When a fusional effort is made, due to pro-
longed wearing of opposite based prisms (or in etheropho-
ric patients) to maintain fusion, an "after-effect" be-
comes evident. It lasts for a period of several minutes,
hours or even more if the prisms are worn for a prolonged
period of time (Carter). The after-effect consists in a
change in the rest position in the direction of the fu-
sional movements and the patient becomes eso-exo-hyper or
hypophoric according to the prism orientation. Maddox
had already observed this phenomenon and felt it is an
expression of an "adaptation" to relieve convergent stress
The motor fusional effort becomes, in fact, less stress-
ing; there is less demand for innervational impulses to
support this effort if the tonus of the muscles increases
in the direction that motor fusion is exerted.
This phenomenon bears practical importance for allevi-

ating stress in etherophoric patients. It is also very
126 B. BAGOLINI
important if we have to operate on these patients to re-

lieve their symptoms. The constant fusional effort of an
exophoric patient to keep the eyes straight generates an
adaptational after-effect that partially masks the basic
deviation upon which we want to operate. The angle of the
rest position that we can clinically detect, appears to be
less than what it really is. This angle decreases of an
amount that varies from patient to patient for an indivi-
dual variation of this adaptational after-effect. A pro-
longed occlusion of one eye to eliminate the fusional sti-
muli and consequently the adaptational after-effect, fre-
quently reveals a greater angle of exophoria which is the
one to be eliminated by appropriate surgery.
Schor has investigated the adaptation after-effect. In

a model he has developed, he postulates two fusional ver-
gence systems. A fast fusional system (fast integrator)
responds immediately to the stimulus over retinal dispa-
rate points to eliminate the disparity. The second system
is a slow fusional vergence system (slow integrator) which
does not seem to be directly under the influence of the
stimulus but it receives its imput from the fast fusional
system. According to Schor's model, the actual vergence
angle is determined by the sum of the outputs of these
two systems.
Mais and Tello found in the mesencephalic reticular

formation, near the oculomotor nucleus in monkeys, cells
which elicit a pure convergent signal. Their work leads
to confirming Schor's model. It can also be inferred from
the work of Mais and Tello that not only the fast system,
but also the slow one has a mesencephalic origin. The af-
ter-effect is, therefore, not due to a sort of intrinsic
contraction of the muscles which take time to relax.
At this point, it may be interesting to compare this

tonic adaptational after-effect just described with a slow
increase of tonus occurring in many esotropic patients
whose angle deviation is corrected by prisms. This
increase of tonus of the medial rect cause-s an increase
of the angle of strabismus (prism compensation).
Figure 2 shows the behaviour of 61 esotropic patients with
an esotropia not over 25 ~ . On the right side, we see
patients capable of compensating for base-out prisms which
were progressively increased up to 60 and more prism di-

opters. Going toward the left, patients were compensa-
ting for a decreasing amount of prisms till no compensa-
tion could be found in some of them. This phenomenon was
sporadically described (Travers, Burian 1947, Halpern &
Hofstetter) and under experimental conditions was studied
in a few cases by Hallden. This increase of tonus of the
medial recti which develops only in some esotropic pa-
tients wearing a prismatic correction is important not
only from a physiopathological point of view, but also
because it implies poor surgical results for the cases
capable of compensating for a great amount of prisms (Ba-
golini 1986). This phenomenon is a slow compensation of
prismatic application. The increase of tonus detectable
as an increase in the angle of squint under the prismatic
effect, takes minutes or hours to develop. After prism
remotion, the increase of angle may take minutes or hours
to reach the initial squint angle. It is logical to think
that the convergent impulse derives from the same mesen-
cephalic cells responsible for the slow fusional vergence
system mentioned before in normal subjects. The puzzling
fact is that the fast fusional phase is lacking in these
patients and it is not clear, therefore, what gives the
imput to this increase of tonus. A few theories have
been presented by ophthalmologists, but do not stand
criticism (see Bagolini 1986).
Bagolini (1976) supports the opinion that they are

anomalous fusional movements tending to bring equal ima-
ges over anomalously corresponding retinal areas. The
anomalous correspondence is, in fact, not a point to
point correspondence as in normal correspondence, but in-
volves more or less large areas in the two eyes. This
implies that a single point on the retina of the fixing
eye may acquire a correspondence not with a single reti-
nal point, but with a large retinal area in the deviated
eye. This has been demonstrated by Bagolini & Capobianco,
Hallden (1973), Johnston & others. Most of these experi-
ments in esotropic patients were done with an horopter
apparatus suitably modified in such a way that binocular
cooperation inside the areas of binocular vision (anoma-
lous Panum's area or pseudo-Panum's area) was studied by
using a non-dissociating device (striated glasses - Bago-
lini, 1961). This device does not disturb (does not dis-
sociate) this weak type of binocularity which would dis-
!28 B. BAGOLINI
appear when elements disturbing this anomalous sensorial

fusion are introduced (dissociating tests). Dissociation
would stop this abnormal binocularity and suppression or
diplopia is the common reaction.
Apparently, the slow convergent movements, compensa-

ting for prism correction of the angle of strabismus,
have the aim of maintaining some binocularity bringing
equal images over anomalous corresponding areas. Anoma-
lous retinal correspondence (a.r.c.) is an adaptational
phenomenon to the strabismic deviation (see, for example,
Burian, von Noorden, Bagolini 1967 & 1976 Part I). The
slow tonic movement (anomalous movement= a.m.) just des-
cribed appears also to be an adaptational phenomenon to
the angle of deviation as can be inferred from the fact
that they are most of the time not present or are weak
in recent strabismus. A.r.c. and a.m. usually parallel
each other in strength (Bagolini 1985), but they may ap-
pear (and disappear after treatment) at different times.
The a.r.c. appears to be a variation of spatial localiza-
tion of the retinal elements. The a.m. appear to be a
variation of the motorial value of the retinal elements.
The Q fusional motorial value of normally corresponding
retinal points appears to be shifted into newly corres-
ponding retinal areas in binocular vision. Their fusion-
al nature appears to be proven by Campos and Zanasi. They
noticed that these slow movements are convergent with
base-out prisms but small divergent or vertical movements
can be elicited, respectively, with base-in or vertical
prisms.
The resemblance of these slow anomalous movements

with the after-effect previously described is striking.
A work of Schor's who demonstrates a convergent after-ef-
fect triggered by accomodation, suggests that accomodation
could also possibly trigger the convergent mouvement
in esotropia. It is conceivable that a slow fusional
phase to maintain an anomalous binocularity is not trig-
gered by a fast fusional phase which is lacking in squin-
ting patients, but is triggered by accommodation (as sug-
gested by Schor). Studies to explore this possibility
are underway. The work of Campos and Zanasi, however,
?howing some vertical movements when applying vertical
(l~he strength of the a.m. can be inferred from the a-
mount of prismatic overcorrection they can compensate for.
(see Bagolini 1986).
prisms cannot be explained by a triggering of accommoda-

tion.
REFERENCES
Adelstein, F.E., Clippers, C. (1966). Zum Problem der ech-

ten und scheinbaren AbducensHimung ( das sogenante "Bloki-
erungssyndrome"). In Augenmuskellahmungen, Elich. Augenartz
Enke Stuttgart. 46:271-278.
Alpern, M., Hofstetter, H.W. (1968). The effect of prism

in esotropia. A case report. Am. J. Optom. Arch. Am.Acad.
Optom. 25: 80-91.
Bagolini, B. (1961). Diagnostic et possibilite de traite-

ment de l'etat sensoriel du strabisme comitant avec des
instruments pour dissociants (test du verre strie et barre
des filtres). Ann. d'Ocul. Paris, 146: 236-258.
Bagolini, B. (1976). Part I. Sensorial anomalies in stra-

bismus (suppression and a.r.c.). Doc. Ophthal. 41: 1-22.
Bagolini, B. (1976). Part II. Sensoriomotorial anomalies

in strabismus (anomalous movements). Doc. Ophthal. 41:
23-42.
Bagolini, B. (1985). Objective evaluation of sensorial

and sensoriomotorial status in esotropia: their importance
in surgical prognosis. British J. Ophthal. 69: 725-728.
Bagolini, B., Capobianco, N.M. (1965). Subjective space in

comitant squint. Am. J. Ophthal. 59: 430-442.
Bagolini, B., Penne, A., Zanasi, M.R. (1983). Ocular nys-

tagmus: some interpretational aspects and methods of
treatment. Inter. Ophthal. 6: 37-48.
Bagolini, B., Zanasi, M.R., Bolzani,K. (1986). Surgical

correction of convergent strabismus: its relationship to
prism compensation. Doc. Ophthal. 62: 309-324.
Breinin, M.G. (1962). The electrophysiology of extraocu-

lar muscle. Toronto Press, 41.
130 B.BAGOLINI
Burian, H.M. (1947). Sensorial retinal relationship in

comitant squint. Arch. Ophthal. 37: 336-504-618.
Burian, H.M. (1985). In von Noorden, G.K. Burian-von Noor-

den's binocular vision and ocular motility. Third edition.
Mosby, St. Louis, 78.
Campos, E.C., Zanasi, M.R. (1978). Die anomalen Fusionbe-

wegung. Der sensorische Aspekt des anomalen Binocularse-
hens. Graefes Arch. Clin. Exp. Ophth. 205: 101-111.
Campos, E.C., Chiesi, C. (1982). Binocularity in comi-

tant esotropia and exotropia. Docum. Ophth. Proc. Series.
32: 55-63.
Carter, D.B. (1963). Effects of prolonged wearing of

prism. Am. J. Optom. 40: 265-275.
Hallden, V. (1952). Fusional phenomena in anomalous cor-

respondence. Acta Ophthal. 37 (suppl. 1): 1-93.
Hallden, V. (1973). The longitudinal horopter in a case

of comitant strabismus with anomalous correspondence.
Acta Ophthal. 51: 1-11.
Hofmann, F.B., Bielschowsky, A. (1900). Ueber die Willkuer

Entzogenen fusionsbewegungen der Augen. PflUg. Arch. ges.
Physiol. 80: 1-14. Quoted by Alpern, M. (1962). The eye.
Vol. 3, Muscular Mechanism (ed. H. Davson). Academic
Press, London, 63-151.
Hofstetter, H.W. (1946). Accommodative convergence in

squinters. Am. J. Optom. 25: 417-437.
Jampolsky, A. (1970). Ocular divergence mechanism. Trans.

Am. Ophthal. Soc. 68: 730-822.
Johnston, H.V. (1973). Clinical horopter determination

and the mechanism of binocular vision in anomalous corres-
pondence. Ophthalmologica 163: 102-119.
Keiner, G.B. (1951). New viewpoints on the origin of

squ~nt. Martinus Nijhoff, The Hague.
Maddox, E.E. (1983). The clinical use of prisms and the

decentering of lenses. 2nd ed., Wright & Sons Press,
Bristol, England.
Mais, L.E., Tello, C.A. (1986). Neurophysiological corre-

lates of convergence and its tonic adjustment. In Advan-
ces in Biosciences, 57. Adaptive processes in visual and
oculomotor systems. (eds. E.L. Keller & D.S. Zee). Perga-
mon Press, London, 143-149.
Noorden, G.K. von. (1969). Divergence excess and simula-

ted divergent excess: diagnosis and surgical management.
Documenta Ophthalmologica. 26: 719-728.
Parks, N.M. (1958). Abnormal accommodative convergence

in squint. Arch. Ophthal. 59: 364. Quoted by Burian-von
Noorden, Binocular vision and ocular motility. 3rd ed.,
(ed. G. von Noorden). Mosby Press, St. Louis, 1985, 282.
Schor, M.G. (1983). Fixation disparity and vergence adap-

tation. In Vergence Eye Movements: Basic and Clinical
Aspects. 'eds. M.G. Schor & J. K. Ciuffreda). Butterworth,
Boston, 465-516.
Travers, T.B. (1936). The comparison between the results

obtained by various methods employed for the treatment of
comitant squint. Pulman & Sons, London.
Worth, C. (1906). Squint: its causes, pathology and treat-

ment. Blakiston's Son, Philadelphia.
11
NORMATIVE OCULOMOTOR DEVELOPMENT
IN HUMAN INFANTS
RICHARD N. ASLIN
INTRODUCTION
A complete understanding of infantile strabismus has been

hindered by uncertainties regarding the development of oculomotor
control in normal infants. Normative oculomotor development cannot
be understood without information about the sensory stimuli that
drive eye movements and about the adaptive mechanisms that fine-tune
the synergy between sensory stimuli and oculomotor commands. In the
past decade, much has been learned about the characteristics of eye
movements in normal infants (Aslin, 1987b). However, it remains
unclear precisely why oculomotor control is initially deficient and
what mechanism leads to oculomotor maturity.
The etiology, diagnosis, and treatment of strabismus would be

simplified if oculomotor control were uniformly mature throughout
the life span. However, we know that virtually all aspects of
oculomotor control in normal infants are immature. Moreover,
because young infants are frequently sleepy, fussy, or inattentive,
their average oculomotor performance is much less mature than their
best oculomotor performance. Despite these problems, reliable data
on oculomotor control can be obtained from young infants (Aslin,
1985), although rarely with the precision typical of recordings
obtained from adults. Unfortunately, data from infants are not easy
to gather, particularly under the time constraints faced by the
clinician. As a result, if one hopes to characterize a given
infant's oculomotor capacity, then brief samples of eye movements
are very likely to overestimate suspected deficiencies. Thus, data
from normal infants may provide useful "landmarks" for comparison to
the observed oculomotor characteristics of infants with manifest or
suspected strabismus.
1
Supported by NIH research grant (EY-05976) to R.N.A. and by NIH
center grant (EY-01319) to the Center for Visual Science.
133
134 R.N. ASLIN
SENSORY REQUIREMENTS FOR BINOCULAR ALIGNMENT
To maintain accurate binocular alignment, the visual system

must detect retinal non-correspondence and convert that interocular
error signal into appropriate oculomotor commands. Detection of
binocular misalignment is limited by the quality of both monocular
and binocular sensory information. Psychophysical and electro-
physiological data gathered over the past two decades have provided
compelling evidence that virtually all aspects of spatial and
temporal vision are deficient in normal infants until several months
after birth (Banks & Dannemiller, 1987). As a result, the effective
stimulus for triggering accommodative and fusional vergence is
limited not only by poor acuity and contrast sensitivity but also by
the apparent absence of functional stereopsis until approximately 4
months after birth (Held et al., 1980; see also Held, this volume).
The neural locus mediating these sensory deficits in young

infants has generally been assumed to be the visual cortex. The
poor disparity tuning and poor contrast sensitivity that char-
acterize the response properties of immature cortical neurons (see
Blakemore, Crawford, and von Noorden, this volume) are consistent
with this assumption. However, anatomical developments at the level
of the retina may account for a large proportion of the sensory
deficits observed in young infants. Both the immature morphology of
cone outer segments and the reduced packing density of cones in the
central fovea (Yuodelis & Hendrickson, 1986) suggest that as much as
5 deg of the central retina may be only marginally functional at
birth. In contrast, extrafoveal cone morphology appears to be very
similar to that observed in adult retina. These anatomical data are
certainly consistent with the large dispersion of eye position
observed in young infants as they fixate a small target, although
other factors (e.g., inattention) may also contribute to high
fixational variances. In addition, because the postnatal migration
of photoreceptors toward the fovea alters the relation between
visual direction and oculomotor commands (Aslin, 1987a), some
recalibration of the sensory signals for eye movements must occur
after birth.
The foregoing argument does not imply that all sensory deficits
observed in normal infants can be attributed to retinal development.
There is compelling evidence that response properties of neurons
beyond the retinal level undergo significant postnatal development.
However, if we view the retina as a preliminary filter on visual
input, significant postnatal retinal developments would limit the
sensory information available at higher neural levels.
OCULOMOTOR CONTROL IN NORMAL INFANTS
The preceding summary suggests that sensory developments

account for part of the deficiency in oculomotor control observed in
young infants. Moreover, monocular sensory deficits can contribute
INFANT OCULOMOTOR DEVELOPMENT 135
to binocular misalignments in patients with amblyopia or strabismus

(see Levi, this volume). The key question is whether these sensory
deficits alone are sufficient to account for inaccurate binocular
alignment in young infants.
Despite widespread clinical evidence that normal newborns

appear to be somewhat exotropic (e.g., Rethy, 1969), there have been
only four detailed studies of the accuracy of binocular alignment in
normal infants (Aslin, 1977; Maurer, 1975; Slater & Findlay, 1975;
Wickelgren, 1967). All four studies used the corneal reflection
technique, but only two studies varied target viewing distance to
induce a change in vergence angle. A portion of the apparent exo-
tropia could be attributed to measurement errors associated with
uncalibrated use of corneal reflection data (Bronson, 1983).
As shown in Fig. 1, steady-state errors in binocular alignment

are present until approximately the third postnatal month. Younger
infants show systematic evidence of convergence insufficiency when
fixating a near target. Under these binocular viewing conditions,
in which a real target is positioned at different distances, both
accommodative (blur) and fusional (disparity) stimuli are present to
trigger changes in vergence angle. Thus, although these data show
that binocular alignment in young infants is frequently inaccurate,
they do not document whether this inaccuracy is the result of
deficits in accommodative vergence, fusional vergence, or both.
--.,
OJ
Q)
"0
5
'-
~ 0
OJ
c: 10
"(
Q)
(.,)
c:
Q)
2' 15 o Newborns
~ 0 1-month-olds
0 2- month -aids
• 3-month-olds
20
•
50 33 25 15 12.5
Target distance (em)
Figure 1. Vergence angle as a function of target viewing
distance in newborns (Slater & Findlay, 1975) and in older infants
(Aslin, 1977). Dotted line indicates expected value.
136 R.N. ASLIN
Accommodative vergence
A deficiency in accommodative vergence could contribute to the

inaccuracies in young infants' binocular alignment shown in Fig. 1.
Although accommodative accuracy in young infants is quite poor,
Banks (1980) has provided compelling evidence that these accom-
modative inaccuracies, as well as their reduction with age, can be
accounted for solely by the degraded blur signal resolvable by the
infant visual system. The obtained data on accommodative accuracy
fall within the predicted range of the depth of focus of the infant
eye (based on eyeball size, grating acuity, and contrast discrim-
ination). Thus, the improvement with age in the accuracy of accom-
modation can apparently be accounted for solely by the degraded
sensory signal available for triggering an accommodative response.
Insufficient accommodation could account for the inaccuracy of

binocular alignment shown in Fig. 1 if there were an innate linkage
between the accommodative and vergence systems. That is, because
the newborn visual system cannot detect changes in stimulus blur
associated with moderate changes in target viewing distance, the
vergence response to changes in viewing distance may be incomplete
because of a reduced contribution from accommodative vergence. The
only study that has directly measured accommodative vergence in
young infants (Aslin & Jackson, 1979) reported that by 2 months of
age a 7 diopter accommodative stimulus elicits a convergence
response. Younger infants were not tested, but it seems likely that
the linkage between the accommodative and vergence systems is
present at birth. Unfortunately, the accuracy of the corneal
reflection technique used to measure changes in vergence angle was
not sufficient to provide an estimate of the AC/A ratio. However,
because of the postnatal increase in interocular separation, it
seems likely that the AC/A ratio (expressed in visual degrees rather
than meter angles) changes during postnatal development. In
summary, although accommodative vergence may provide a partial
explanation for the inaccuracies in binocular alignment observed in
young infants, once spatial resolution improves to near-adult levels
by the end of the first postnatal year, accommodative vergence alone
will not be sufficient to align the two visual axes within the range
required for sensory fusion.
Fusional vergence
A deficiency in disparity detection may also contribute to the

inaccuracies in young infants' binocular alignment shown in Fig. 1.
In normal adults, fixation disparities as small as several minutes
of arc are sufficient to trigger a vergence eye movement (Rashbass &
Westheimer, 1961), even if diplopia is absent (Duwaer & van den
Brink, 1981). Moreover, large disparities can initiate a vergence
response even when the two half-images differ in shape or
orientation (Westheimer & Mitchell, 1969). Of course, under these
conditions binocular rivalry rather than sensory fusion results once

the vergence response has been completed.
Fusional vergence has not been measured directly in young

infants. This is largely the result of the difficulties involved in
recording binocular alignment while infants view a target under
dichoptic conditions, thereby eliminating the contribution of
accommodative vergence. However, as shown in Fig. 2, when a small
wedge prism is placed in front of one eye while a target is fixated
binocularly, refixation eye movements are absent in normal infants
until the fourth postnatal month. Thus, the estimated size of the
minimum stimulus for motor fusion prior to 4 months of age is
greater than 10 prism diopters. In normal adults this amount of
binocular misalignment would lead to diplopia, followed by a
vergence response to realign the target onto corresponding retinal
loci. Apparently, either young infants are not capable of making a
vergence response under these conditions in which disparity and blur
signals are dissociated, or the sensory stimulus triggering a
fusional vergence response in young infants is degraded.
The hypothesis that the sensory stimulus for fusional vergence

is degraded in young infants is consistent with two lines of
evidence. First, the rapid improvement in stereoacuity at 4 months
of age suggests that cortical neurons tuned to small disparities are
initially absent or deficient. Second, normal adults' sensitivity
(/)
•• 105 prism diopters
...J 100
<{ 0 0 (control)
a::
I-
I-
2
w
~~
wl-
>2 50
ow
~~
w
2a..
o~
!;:{
X
LL.
w
a::
0 o••
2 3 4.5 6
AGE (MONTHS)
Figure 2. The probability of a re-fixation response to prism-
induced binocular misalignment. Data replotted from Aslin (1977).
138 R.N. ASLIN
to diplopia is proportional to the presence of high spatial fre-

quencies in the fixation stimulus (Schor et al., 1984), and high
spatial frequencies are not detected by young infants. Thus, the
data presented in Fig. 2 are consistent with the hypothesis that
sensory signals triggering fusional vergence are degraded in early
infancy. These data do not clarify whether sensory fusion is
present in young infants-.--However, Birch et al. (1985) and Shimojo
et al. (1986) have provided evidence that binocular rivalry and
sensory fusion emerge at 4 months of age (see Held, this volume).
In addition to deficiencies in these steady-state mechanisms that
may underlie binocular misalignments, the dynamics of converting
suprathreshold disparity information into an appropriate vergence
response may be immature in young infants. These dynamic properties
of infant fusional vergence have not been measured.
Conjugate eye movements
Although accurate binocular alignment is determined primarily

by the accommodative and fusional vergence systems, non-conjugate
saccadic and pursuit movements could lead to transient binocular
misalignments that might prevent sensory fusion and lead to per-
manent deficits in disparity processing. Normal infants appear to
make conjugate saccades, but no quantitative data with sufficient
resolution have been provided to determine the degree of conjugacy.
Recent evidence from normal adults (Enright, 1986) suggests that
changes in vergence are accomplished in part by non-conjugate
saccades. It is important to note that saccades in normal infants
are grossly hypometric (Aslin & Salapatek, 1975). Although there
are several potential explanations for these deficient infant
saccades, it is possible that coordinated control of saccadic gain
in the two eyes requires a minimum amount of postnatal visuomotor
experience. In monkeys, both monocular patching (Vilis et al.,
1985) and unilateral extraocular muscle weakness (Optican &
Robinson, 1980; Snow et al., 1985) lead to non-conjugate saccades
that become recalibrated only after binocular visual experience
lasting hours, days, or even weeks. An inefficient interocular gain
control mechanism, therefore, could lead to lengthy periods of
binocular misalignment.
The pursuit and optokinetic systems in normal infants are also

characterized initially by deficiencies that could lead to binocular
misalignments. As summarized by Hoffmann and by Kommerell (this
volume), cats and humans show asymmetrical OKN under monocular
viewing conditions. A similar superiority of nasalward over
temporalward smooth pursuit under monocular viewing conditions has
been observed in normal 2-month-olds in my laboratory. It is
important to note, however, that infants' tracking responses to low
velocity smooth movement is entirely saccadic until the end of the
second postnatal month (Aslin, 1981). As shown in Fig. 3, when
smooth pursuit emerges, it is characterized by low gain with
frequent saccadic interruptions.
&-WEEK-OLD
z
0
i= 10-WEEK-OLD
Ui
0
a.
w
>-
w
-'
<(
~
z
2
a:0
:I:
ADULT
TIME
Figure 3. Sample eye movement tracings illustrating the

developmental emergence of smooth pursuit to horizontal target
excursions (represented by the smooth sinusoidal line). Data
replotted from Aslin (1981).
Smooth pursuit may be absent initially in normal infants

because of sensory deficits associated with foveal immaturity, and
this may explain why whole-field motion is an effective elicitor of
the slow phase of OKN. Alternatively, smooth pursuit may be absent
because of deficiencies in the processing of velocity information.
A nasal-temporal asymmetry in smooth pursuit gain is present in
congenital esotropes, and this asymmetry is correlated with nasal-
temporal differences in velocity perception (Tychsen & Lisberger,
1986). Under binocular viewing conditions, these nasal-temporal
asymmetries in OKN and pursuit are eliminated or greatly reduced,
even for infants younger than 4 months of age who lack stereopsis,
fusion, and rivalry. If an initial nasalward drift bias were
suppressed under binocular viewing conditions by alternation of
fixation, then pursuit might be largely saccadic.
140 R.N. ASLIN
An additional factor that may contribute to the absence of

smooth pursuit in early infancy, as well as the presumed
deficiencies in fusional vergence dynamics, is the adaptive control
of oculomotor gain. The inherent delay (approximately 160 msec) in
initiating a pursuit or fusional vergence response requires the
eye(s) to accelerate to "catch up" to the displaced target. If this
initial acceleration is too great, repeated overshoots
(oscillations) occur. Optican et al. (1985) showed that this
initial acceleration of smooth pursuit is under adaptive gain
control. Perhaps the mechanism that mediates this adaptive control
is immature in young infants. A similar form of adaptive control
has been documented by Miles (1985) for both accommodative vergence
and vergence accommodation. Interestingly, both systems are limited
to adaptations that compensate for increases in interocular
separation. Of course, increases in interocular separation require
higher vergence gain, and low gain seems to characterize both the
vergence and pursuit systems in young infants. Whether this low
gain is the result of sensory deficiencies alone is unclear at
present. Moreover, it is not known whether adaptive gain control
applies to fusional vergence. However, exploration of these
adaptive mechanisms may be a fruitful area for explaining the
origins of certain types of strabismus.
SUMMARY
Sensory deficits constrain accommodative accuracy, and

therefore accommodative vergence. Sensory deficits also constrain
disparity processing, and therefore fusional vergence. These
sensory deficits account for much of the steady-state error in
binocular alignment observed in young infants. However, by the
sixth postnatal month, normal infants have developed sensory
capacities that should not seriously constrain oculomotor control.
Dynamic aspects of vergence control have not been studied in young
infants, although analogous data from the pursuit system suggest
that gain will initially be low. A potentially important topic for
future oculomotor research on infants is the study of adaptive gain
control mechanisms for fusional vergence and for interocular
differences in the saccadic system.
REFERENCES
Aslin, R. N. (1977). Development of binocular fixation in human

infants. J. Exper. Child Psychol., ~. 133-150.
Aslin, R. N. (1981). Development of smooth pursuit in human

infants. In Eye Movements: Cognition and Visual Perception (eds. D.
F. Fisher, R. A. Monty, and J. W. Senders). Erlbaum, Hillsdale, NJ.
Aslin, R. N. (1985). Oculomotor measures of visual development. In

Measurement of Audition and Vision in the First Year of Postnatal
Life: A Methodological Overview. (eds. G. Gottlieb and N.
Krasnegor). Ablex, Norwood, NJ,
Aslin, R. N. (1987a). Anatomical constraints on oculomotor

development: Implications for infant perception. In Infant
Perception: The Minnesota Symposium on Child Psycholo~ed. A.
Yonas). Erlbaum, Hillsdale, NJ,
Aslin, R. N. (1987b). Motor aspects of visual development in

infancy. In Handbook of Infant Perception, Vol. 1. (eds. P.
Salapatek and L. B. Cohen). Academic Press, Orlando, FL.
Aslin, R. N. and Jackson, R. W. (1979). Accommodative-convergence

in young infants: Development of a synergistic sensory-motor system.
Canadian J, Psycho!., l,l, 222-231.
Aslin, R. N. and Salapatek, p, (1975). Saccadic localization of

visual targets by the very young human infants. Percept.
Psychophys., li• 293-302.
Banks, M. s. (1980). The development of visual accommodation during

early infancy. Child Develop.,~. 646-666.
Banks, M. s. and Dannemiller, J, L. (1987), Infant visual

psychophysics. In Handbook of Infant Perception, Vol. 1. (eds. p,
Salapatek and L. B. Cohen). Academic Press, Orlando, FL.
Birch, E. E., Shimojo, s., and Held, R. (1985). Preferential-

looking assessment of fusion and stereopsis in infants aged 1-6
months. Invest. Ophthal. Visual Sci.,~. 366-370.
Bronson, G. W. (1983). Potential sources of error when applying a

corneal reflex eye-monitoring technique to infant subjects. Behav.
Res. Meth. Instrum., ~. 22-28.
Duwaer, A. L. and van den Brink, G. (1981). Diplopia thresholds and

the initiation of vergence eye movements. Vision Res., l!_, 1727-
1737.
Enright, J, T. (1986). Facilitation of vergence changes by

saccades: Influences of misfocused images and of disparity stimuli
in man. J, Physiol. (London), 371, 69-87.
Held, R., Birch, E., and Gwiazda, J, (1980). Stereoacuity of human

infants. Proc. National Acad. Sci. u.s.A., II· 5572-5574.
Maurer, D. (1975). The development of binocular convergence in

infants. (Doctoral dissertation, University of Minnesota, 1974).
Dissertation Abstracts, ~. 6136-B.
142 R.N. ASLIN
Miles, F. A. (1985). Adaptive regulation in the vergence and

accommodation control systems. In Adaptive Mechanisms in Gaze
Control: Facts and Theories. (eds. A. Berthoz and G. Melvill
Jones). Elsevier, Amsterdam.
Optican, L. M. and Robinson, D. A. (1980). Cerebellar-dependent

adaptive control of primate saccadic system. J. Neurophysiol., ~.
1058-1076.
Optican, L. M., Zee, D. s., and Chu, F. C. (1985). Adaptive

response to muscle weakness in human pursuit and saccadic eye
movements. J. Neurophysiol., ~. 110-122.
Rashbass, C. and Westheimer, G. (1961). Disjunctive eye movement.

J. Physiol. (London),~. 339-360.
Rethy, I. (1969). Development of the simultaneous fixation from the

divergent anatomic eye-position of the neonate. J. Pediatric
Ophthal., ~. 92-96.
Schor, C. M., Wood, I., and Ogawa, J. (1984). Binocular sensory

fusion is limited by spatial resolution. Vision Res., 24, 661-666.
Shimojo, s., Bauer, J., O'Connell, K. M., and Held, R. (1986). Pre-
stereoptic binocular vision in infants. Vision Res.,~. 501-510.
Slater, A. M. and Findlay, J. M. (1975). Binocular fixation in the

newborn baby. J. Exper. Child Psychol., 1Q, 248-273.
Snow, R., Hore, J., and Vilis, T. (1985). Adaptation of saccadic

and vestibule-ocular systems after extraocular muscle tenectomy.
Invest. Ophthal. Visual. Sci.,~. 924-931.
Tychsen, L. and Lisberger, s. G. (1986). Maldevelopment of visual

motion processing in humans who had strabismus with onset in
infancy. J, Neurosci., ~. 2495-2508.
Vilis, T., Yates, s., and Hore, J. (1985). Visual patching of one
eye produces changes in saccadic properties in the unseeing eye.
Dev. Brain Res., .!2_, 290-292.
Westheimer, G. and Mitchell, D. E. (1969). The sensory stimulus for

disjunctive eye movements. Vision Res.,~. 749-755.
Wickelgren, L. (1967). Convergence in the human newborn. J, Exper.

Child Psychol., 2• 74-85.
Yuodelis, C. and Hendrickson, A. (1986). ~ qualitative and

quantitative analysis of the human fovea during development. Vision
Res., ~. 847-855.
12
DISCUSSION: OCULOMOTOR CONTROL
AND STRABISMUS
Moderator: GUNNAR LENNERSTRAND
It was agreed upon that the papers on the general

aspects of strabismus delivered by G.Lennerstrand
and B.Harcourt should be discussed at the end of
the session. The discussion started therefore with
the paper by R.Spencer on "Morphology of extra-
ocular muscles in relation to the clinical mani-
festations of strabismus".
Lennerstrand: A strong case was made by Dr Spencer

that the orbital singly-innervated fibres may have
a prominent role in ocular motility and strabismus.
Is it possible that some of the other fibre types
would be of importance in this respect? In our
studies on strabismic and binocularly deprived cats
we saw changes also in other fibres, particularly
with regard to size.
Spencer: It is quite possible that other fibre-
types are affected in strabismus, though perhaps to
different extents depending upon the underlying
factors and whether the deficit is manifested in
the primary position and/or during movements of the
eye. What we have seen with the botulinum toxin is
that a normal muscle has been made strabismic and
one fibre type shows the most dramatic adaptive
change. Certainly in the long term when the length-
tension relationship of the entire muscle has been
changed, undoubtedly other fibre types also may be
143
144 G.LENNERSTRAND
affected, though the changes, if any, were subtle
in comparison. In other instances, such as the
clinically overacting inferior oblique muscle, we
observed changes associated specifically with
another fibre type in the global layer of the
muscle. Recent findings by Porter indicate that yet
another fibre type in the global layer is affected
by bupivacaine. The most significant concept to
emerge from these various findings is that
different fibre types may be selectively involved
in ocular motility deficits largely as a conse-
quence of not only their distribution in the
orbital or global layers of the muscle, but also
their distinct histochemical and ultrastructral
features that relate to oxidative/glycolytic
function and contractile properties. While all
motoneurones and the different muscle fibre types
that they innervate participate in all types of eye
movements, the orbital singly-innervated muscle
fibres may be one of the first recruited during
movement and, being the most fatigue resistant, are
ideally suited to maintaining fixation during
vergence or gaze.
Bagolini: Dr Spencer showed a photograph of a
section of muscle in which it was easy to
distinguish between the orbital and global layers.
If I understand correctly, one part is mainly
related to tonic or probably vergence movements,
while the other part is possibly related to
saccadic and therefore duction movements. From a
practical point of view, would it be possible to
selectively destroy the global part or the orbital
part by surgery or chemicals thus influencing more
the saccadic system or the vergence system?
Spencer: Although the division of labor concept, as
proposed by Scott and Collins, might imply that
different functions are associated with the orbital
versus global layers, it is unlikely that by
lesioning one or the other part you would interfere
with one type of eye movements as opposed to
another. In the case of the orbital layer, the
result might simply be difficulty in maintaining
the position of gaze without necessarily inter-
fering with the ability to perform a saccadic or
vergence eye movement.
Helveston: Clinically we use Oculinum for the
treatment of strabismus as well as blepharospasm.
An acute paresis is created, but after a while
function of the muscle returns. Do you have any
comments on the long term morphological changes of
DISCUSSION: OCULOMOTOR CONTROL AND STRABISMUS 145
the muscles after Oculinum treatment?

Spencer: The findings in adult monkeys clearly
showed that 6-8 weeks after an injection of 10
units toxin into the medial rectus muscle, the
fibres are qualitatively normal in appearance,
although there are quantitative differences
compared to normal. The orbital singly-innervated
fibres are about 24 % smaller in cross-sectional
area and the vasculature has adapted propor-
tionately to the size of the fibres. It also is
about 24-25 % less in cross-sectional area than
normal. The potential efficacy of botulinum toxin
treatment of motility deficits, therefore, might
depend upon the presence and proportion of oxida-
tive muscle fibres. The duration of the paralysis
might be attributable to not only the dose-depend-
ent effect which is related to the recycling of
acceptors on the presynaptic nerve terminal mem-
brane, but also the postsynaptic restructuring of
the oxidative fibres and the proliferation of the
capillary network upon which they are dependent
for normal function. The ptosis that frequently is
secondary to toxin injections of the rectus muscles
thus might easily be explained by the vulnerability
of the levator palpebrae muscle with its high pro-
portion of oxidative fibres. On the other hand,
with the treatment of blepharospasm, the shorter
duration of paresis in comparison to the extra-
ocular muscles might be explained by the paucity of
such fibres. In this case, the duration would be
dependent only upon the presynaptic effect of the
toxin and is not potentiated by structural modi-
fications in the muscle fibres.
Steinbach: Where are the spindles in extra ocular

muscle located?
Spencer: In those species that have true muscle
spindles in the extraocular muscles, they appear to
be located at the junction between the orbital and
global layers, predominantly in the proximal part
of the muscle.
Eggers: Pachter and coworkers* found in serial
sections of mouse extraocular muscle that the
*Pachter,B,J. and Breinin,G.M. (1976) Light and

electron microscopic serial analysis of mouse
extraocular muscle: Morphology, innervation and
typographical organization of component fibre popu-
lations. Tissue & Cell, ~. 547-560.
146 G.LENNERSTRAND
morphology of single fibres changes along the

length of the muscle, implying that it is difficult
to classify the muscle fibres by cross-sectional
appearance at one point. Is there a corresponding
situation in monkey muscle?
Spencer: The orbital multiply-innervated muscle

fibre displays histochemical, ultrastructural, and
innervational variability at different locations
along its length. We have had no difficulty in
identifying the different muscle fibre types in
transverse sections of monkey extraocular muscle.
Lennerstrand: We should also discuss the develop-

ment of extraocular muscle fibres, particularly in
relation to strabismus, and whether the development
or maturation is influenced by the denervation by
botulinum toxin or other procedures of clinical
importance. Dr Spencer, do you have any comments on
this?
Spencer: It is quite clear that the extraocular
muscles undergo considerable development post-
natally in the cat and possibly also in primates.
Dr Lennerstrand has elegantly demonstrated the
extent to which the maturation of muscle fibres in
the extraocular muscles can be manipulated by
alterations in the early visual experience of the
animal, either by visual deprivation or in con-
genital strabismus. He has shown in the cat that
the muscle fibres do not develop to the same extent
as normal. In particular, the oxidative activity is
reduced as is the vasculature. These studies have
demonstrated the value of histochemical analyses of
extraocular muscle as it relates to metabolic and
contractile properties and should be performed in
humans as a means of assessing functional changes
in the muscle fibres in strabismus. We have per-
formed botulinum toxin injections in infant monkeys
at 4 months of age, which would be equivalent to
approximately 16 months of age in humans. At this
time, the orbital singly-innervated fibres are
approximately 40 % smaller than the normal adult.
While the adult muscle fibres had recovered to a
normal morphology by 6-8 weeks after toxin injec-
tion, by 10 weeks in the infant monkey they are far
from complete recovery. The infant muscle thus
appears to be more vulnerable to the effects of the
toxin than the adult. It remains to be determined
whether there is a critical period in the postnatal
development of extraocular muscle during which time
the differentiation and maturation of different
muscle fibre types is intrinsically determined
and/or dependent upon neural or hormonal factors.
Discussion of the paper by G. Lennerstrand: "Motor

and sensory function of normal and strabismic
extraocular muscle".
Lennerstrand: In my presentation I briefly mentio-
ned the studies by Mitsui and coworkers*. Dr von
Noorden will show us a video-tape recording of the
so called magician's forceps phenomenon and give us
his interpretation of the motility patterns.
von Noorden: (Commenting the video-tape). This lady
has an exotropia with strong dominance of the left
eye. The right eye is deviated about 30 prism
diopters while she is fixating with the left eye
and is suppressed. The left eye is now adducted
with a forceps. You see that the right makes an
adduction movement in response to displacement of
the left eye. Now we hold a semi-transparent filter
before the left eye which makes it impossible for
this lady to see anything but diffuse light, but we
can see her eye through the filter. In the third
sequence we have anesthetized her right eye which
is the deviated, suppressed eye. We adduct this eye
now with the forceps and see no movement of the
left eye. My interpretation of this phenomenon is
that it is purely visually elicited. By moving the
left eye with the forceps we move the image on the
retina. The patient attempts to refixate but this
movement is frustrated because the eye is held with
the forceps. On the basis of Hering's law the
impulse to abduct the left eye is transmitted to
the right eye which will adduct. No such movement
in the left eye occurs when we adduct the right eye
with forceps because of suppression of that eye. I
see no reason to explain this phenomenon on the
basis of inflow.
Steinbach: Tamura and Mitsui** say that they got
the same effect under general anesthetic. Can you
comment on how that could happen?
*See Mitsui,Y.(1986) Strabismus and the Sensory

Motor reflex. Excerpta Medica, Current Clinical
Practice Series No. 30, Amsterdam.
**Tamura,O. and Mitsui,Y.(1986). The Magician's
forceps phenomenon in exotropia under general
anesthesia. Brit. J. Ophthalmol., ZQ, 549-552.
148 G.LENNERSTRAND
von Noorden: I have looked for it but never
observed it under anesthesia but would like to ask
Dr. Kommerell who attended Mitsui's experiment to
comment on this.
Kommerell: Regrettably I wasn't able to attend
experiments in general anesthesia, but Prof. Tamura
was kind enough to show me an experiment very
similar to the one Dr von Noorden has just demon-
strated in the video. The only addition was that an
EMG-electrode was inserted in the muscle pulled.
The right eye was adducted with a forceps, and the
electrode was in the right lateral rectus. This
resulted in a series of saccades of the left eye.
In synchrony with them, burst activity could
clearly be heard in the loudspeaker from the EMG of
the right lateral rectus suggesting that the output
of the CNS corresponded to versional saccades. This
finding was compatible with Dr von Noorden's inter-
pretation that the movements of the left eye were
visually elicited.
Discussion of the paper by H.Bicas: "Kinetics of

the eye".
Eggers: Dr Bicas showed a three-dimensional diagram
on which he plotted eye position and innervation
and muscle force. I would like to know what he
takes as a measure of innervation. Is it eye
position or does he have a way of summating the EMG
or some analogous procedure?
Bicas: Usually innervation is related to the
increment of forces to initiate a rotation. How-
ever, in practice, direct correlations of inner-
vation of a muscle, generated tension and position
to where the eye should be moved, are not possible.
In the intact eye, for instance, developed forces
result not only from the contraction of the agonist
muscle (of which measurements of innervation are
presumably being done), but also from the contrac-
tion of one or more synergists and from the
relaxation of antagonists. In turn, such a decay of
forces of the relaxing muscles will be indirectly
expressed by a "liberation" of passive forces
(shortening of previously stretched structures
acting against opponent elastic elements and static
sliding friction).
Kommerell: I certainly agree with Dr Bicas that one
can distinguish between anatomical and inner-
vational types of strabismus by comparing the angle

of squint in the awake and in the anesthetized
patient, and this is of great theoretical interest.
However, as to the practical consequences, it might
not make a difference. What we need to correct is
the angle in the awake patient, and I don't think
the dosage of surgery depends on whether part of
the strabismus is innervational or not. The reason
is that the stiffness of the muscles is independent
of the innervational level, as long as the muscle
has not been made slack by excessive recession.
Bicas: Since the innervational and the mechanic

components of a strabismus may vary from one
patient to the other, different surgical plans have
to be done, accordingly to what those components
amount in each case. That is the reason why the
same angle of strabismus is not always corrected by
the same amount of a surgery.
von Noorden: We should be careful to draw con-

clusions from the position of the eyes under
anesthesia since the eye position changes depending
on the level of anesthesia. It is difficult to
ascertain that the level of anesthesia remains
constant when it has reached the surgical plane.
Discussion of the paper by R.Baker on "Phyletic

organization of brainstem neuronal circuits and the
etiology of strabismus".
Eggers: Dr Baker, why is a second oculomotor
integrator required for the vergence system in
addition to the existing versional motor inte-
grator.
Baker: The major reason is evolution itself as
evidenced by the absence of an obvious morpho-
logical correlate. There is little structural
support for direct or indirect synaptic pathways
between the midbrain and the prepositus.
Helveston: Dr Baker suggested that the human brain-
stem may be insufficiently wired and unable to
withstand insults with the result being strabismus.
Could you elaborate a bit further on this?
Baker: The humans brainstem contains neurons and
circuits largely designed to accomplish versional
eye movement without concern for alignment. Years
of evolution wasn't prepared for a fovea and the
ISO G.LENNERSTRAND
necessity to align the eyes preciSely. Binocu-
larity, retinal correspondance and vergence
required incorporating new circuits and neurons
with existing machinery constructed in different
reference frame works. As one example of such
organization, it appears that early in vertebrate
phylogeny visual and vestibular pathways were
separate onto the motor nuclei. However, as their
reference frame work became more closely aligned,
so did the central pathways to the extent that in
mammals both signals are combined on afferent
neurons to motoneurons.
Haase: I would like to ask Dr Baker if he can
explain the very frequent alphabetic phenomena in
infantile exotropia on the basis of the brainstem
because I can't believe the explanation given by
Gobin, that these syndroms are due to anatomical
aberrations in the orbit, or false insertion on the
globe. We made analysis on this topic for 20 years
and did not find different insertions of the
superior or inferior obliques muscles in V- or A-
phenomenon. So I can't explain these phenomena on
an anatomical basis.
Baker: I doubt very much whether you can explain
them in the brainstem either, given our current
knowledge. My paper argues that an innervational
mechanism is likely, based on the involvement of
the prepositus nucleus.
Bagolini: I think that there may be a mechanical
component influencing the A and V pattern. For
example has Dr Lennerstrand presented 21 cases of
Arnold-Chiari syndrome. The picture of a CT-scan of
one patient shows a case with hydrocephalus with a
bulging frontal bone. Now you stated that in five
cases out of 21, you have found an A pattern. This
can be explained by the fact that the trochlea may
be displaced frontward in hydrocephalus. The angle
that the reflected part of the superior oblique
forms with the antero-posterior axis of the globe
is narrower than usual. You get therefore an A
pattern according to the theory of Gobin. We
certainly can not exclude that the obliques are
sometimes overactive or underactive because of a
tonus variation, but there are certainly anatomical
cases. Very probably the 5 Arnold-Chiari cases you
presented have an anatomical basis if they are
hydrocephalic.
Lennerstrand: This mechanism has been postulated.
We have not yet examined the material I presented
DISCUSSION: OCULOMOTOR CONTROL AND STRABISMUS lSI
in this respect. I would be more inclined to favour

Dr Baker's view and support an innervational
mechanism.
Blakemore: I should like to raise the question of
the possible role of the cerebral cortex in the
regulation of disjunctive eye movements and
accommodation. Certainly the cortex is likely to be
involved in the detection of the sensory cues for
controlling such responses - retinal disparity in
the case of vergence movements, and blur in the
case of accommodation. Indeed, Jampel* reported
that vergence movements and accommodative changes
can be produced by faradic stimulation of extra-
striate cortical areas in the anesthetized monkey.
Stuart Judge in Oxford (unpublished observations)
has tried, so far without success, elicit such
responses with intracortical microstimulation in
the alert monkey, but his search for a cortical
centre for the near response has not been exhaust-
ive.
Singer: Does this imply that the vestibulo-ocular
reflex machinery is more adaptive than the machin-
ery that controls vergence? Can this be tested
experimentally?
Baker: The neuronal site(s) of plasticity in the
vestibular ocular reflex have not been established
yet. The cerebellum, either cortex and/or nuclei
are intimately involved and the extent might be
different for versional versus vergent movement.
The flocculus has been shown to be directly related
to horizontal conjugate gaze adaption. However, its
removal does not interfere with vergence adaption.
If, as likely, integrators are involved in the near
response, then it would be interesting to focus on
midline cerebellar structures like the nodulus and
uvula. According to Miles (1985) and Schor (this
volume) adaptive plasticity of vergence and accom-
modation is equally as well developed and testable,
as the VOR and saccadic system.
Aslin: To follow up on what Dr Blakemore was
asking, are you suggesting that visual input is of
secondary importance to the adaptive mechanisms
that presumably are mediated by the cerebellum?
*Jampel,R.S. (1960). Convergence, divergence,

pupillary reactions and accommodation of the eyes
from faradic stimulation of the macaque brain.
J.Comp.Neurol. 115, 371-400.
!52 G.LENNERSTRAND
Raker: I would say that visual input is necessary
but not sufficient for explaining either adaption
or the etiology of strabismus.
Hoffmann: First a comment to the question about
cortical influences on subcortical structures. I
would say that there is no subcortical visuo-motor
structure which is not getting cortical input, even
the pupillar response is under cortical influence
and I am certain that near response movements are
influenced profoundly cortically. Secondly to the
notion of cortical plasticity and that the cause of
strabismus will be at the subcortical level. What
happens if the cortical output is faulty? The
brainstem mechanism, I think, have to assume to a
certain degree that the cortical output provides
true information. I don't see a possibility how the
brainstem could get around this problem.
Schor: Vergence adaption is not directly dependent

on the stimulus. The adaption is a secondary
response to another primary motor response. For
example vergence adapts in response to vergence
stimulated by disparity, accommodation and perhaps
proximity. The adapter seems to have no knowledge
of the source of the original vergence stimulus, so
it could reside in the midbrain or brainstem and
have virtually no cortical contribution whatsoever.
Baker: Dr Schor's deduction is at the heart of the
issue, namely separation of the sensory organi-
zation from the motor.
Lennerstrand: Dr Baker, what would you suggest to
look for experimentally in order to test the
hypothesis that infantile strabismus is an oculo-
motor disorder?
Baker: Experimentally I would first examine the
structure-function organization of the vergence
system, utilizing existent intracellular electro-
physiological and HRP techniques in the alert
animal. Then I would proceed to the more difficult
task of studying the development of accommodation
and vergence related neurons in the mammal. Time
consuming but absolutely necessary.
Discussion of the papers by K.P.Hoffmann on "Neural

basis for changes in the optokinetic reflex with
strabismus and amblyopia" and G.Kommerell on
"Ocular motor phenomena in infantile strabismus".
von Noorden: It seems to me premature to equate a
loss of binocularity with optokinetic asymmetry. It
is misleading to base any judgement regarding
symmetry of the OKN on clinical observation. Eye
movement recordings are essential to come to such a
conclusion. Obviously we cannot do this in young
children. One must study a population old enough to
obtain reliable ENG recordings. Studies that I have
seen so far do not describe the type of strabismus
very well and I have recorded OKN asymmetry in
patients with normal binocular vision and symmetric
OKN in patients with documented infantile eso-
tropia. Clearly, there is no consistancy. We need
better clinical studies that also consider such
elements as the location of suppression scotomas on
the retina which could presumably influence pursuit
movements, and the increased tonus of the medial
rectus muscle in esotropia that could influence the
symmetry of OKN.
Hoffmann: One suggestion to study OKN more care-
fully would be to look closer at the open loop
performance of the system. That could be done quite
easily by not using an optokinetic drum moving for
5 min in the same direction but having alternating
directions, so that the eye can't catch up with the
stimulus. This procedure probably will show
asymmetries where you haven't seen them so far.
With respect to the idea of a decrease of binocu-
larity being the main cause for asymmetry I would
like to repeat that the NOT is the critical link
between the eyes and the oculomotor system for OKN.
This has also been shown by lesion studies*.Each
eye has to be connected to the contralateral as
well as to the ipsilateral NOT for symmetric OKN.
Our data show that the only way the ipsilateral
projection can reach the NOT is through binocular
cortical cells. With a break down of binocularity
in the movement processing cortical system OKN has
to become asymmetric.
*Kato,L., Harada,K., Hasegawa,T., Igarashi,T.,

Kojke,Y. and Kawasaki,T.(1986). Role of the nucleus
of the optic tract in monkeys in relation to
optokinetic nystagmus. Brain Res., 364, 12-22.
154 G.LENNERSTRAND
Kommerell: As mentioned in my talk, I quite agree
with Dr von Noorden that there are exceptions to
the rule. But the majority of cases suggests that
it is very difficult for the brain to establish a
symmetrical OKN system if binocularity fails to
develop in the first months of life.
Harcourt: Both Joyce Mein* and John Flynn** have
indicated that there is clinically no direct corre-
lation between asymmetrical OKN responses and
absence of binocular vision. Many of Mein's
patients who had infantile esotropia without
nystagmus or DVD had no clinical evidence of
binocular single vision and yet had symmetrical OKN
responses.
Singer: We need to define what we mean when talking
about binocularity. The visual system is a dis-
tributed system. Different aspects of the visual
input are processed in parallel in numerous
prestriatal cortical areas and several of these
functions require evaluation of binocular signals.
The analysis of the trajectories of moving objects
is one example. Clinical tests, however, tend to
assess cooperativity between the two eyes only with
respect to stereopsis and fusion. It is commonly
ignored, therefore, that most of the strabismic
amblyopes are still capable of integrating binocu-
lar signals for the computation of movement
trajectories***· Likewise, what physiologists
usually address as loss of binocularity is a
reduction of binocular cells in primary visual
cortex. Such loss does, of course, not exclude
persistence of binocular neurons in other cortical
areas such as e.g. the superior temporal sulcus in
monkey or the Claire-Bishop area in cat. These
areas do influence
*Mein,J.(1983). The OKN response in early onset

strabismus. Australian Orthopt. J., £Q, 13-17.
**Flynn,J.T., Pritchard,C. and Lasley,D.(1984).
Binocular vision and OKN asymmetry in strabismic
patients. In Strabismus II. (Ed. R.D.Reinecke),pp.-
35-44, Grune & Stratton, Orlando.
***Sireteanu,R., Fronius,M. and Singer,W.(1981).
Residual binocular interaction in the peripheral
visual field of squint and anisometropic amblyopes.
Vision Res., ~. 1065-1074.
subcortical premotor centers such as the NOT and

can convey binocular signals even if binocularity
s lost in area 17.
Campos: DVD usually does not occur at birth. It is
usually detected later in life and often after
horizontal strabismus surgery. Therefore, how does
DVD fit with the relationship between nystagmus and
loss of binocularity as suggested by Dr Kommerell?
I would also add that a relationship between
strabismus and latent nystagmus cannot hold always,
as we do see patients with latent nystagmus and
healthy binocular vision.
Goller: Dr Hoffmann, could you elaborate a bit on
the asymmetry you find in the motor response when
you use different velocities of stimulation.
Hoffmann: The asymmetry was found to be largest for
medium velocities. 20-40 degrees per second on the
retina (i.e. retinal slip velocity) leads to the
highest gain of the optokinetic reflex in normal
human subjects or macaque monkeys. At these
velocities you will see the largest asymmetry. I
also would like to make a very brief comment on the
velocity response properties of different inputs to
the system. I think it should be very clear that
the asymmetric response at birth is due to the
retinal input. The retinal ganglion cells prefer
very low velocities only up to 10-20 degrees per
second and their best performance is at about 1 or
even below 1 degree per second whereas the high
velocity input to the system which is the more
important one in adults comes via cortical projec-
tions.
Aslin: I must admit I still have great difficulty
understanding the relationship between the develop-
ment of asymmetrical OKN and the emergence of
esotropia. Thus, I would like to direct a question
to Drs Hoffmann and Kommerell: Dr Harcourt raised a
question in his presentation as to why, under
binocular viewing conditions, OKN is not asym-
metrical? Let us say we have stripes moving from
right to left and the left eye is patched. This
nasalward motion will be a very effective signal
for driving OKN. However, with both eyes viewing,
the same right to left movement will be a good
signal in the right eye and a poor signal in the
left eye. But since we don't see asymmetrical OKN
under binocular viewing conditions, the degraded
signal presented to the left eye must somehow be
suppressed. That is, the OKN system must use the
156 G.LENNERSTRAND
greater magnitude signal from the two eyes to drive

OKN. How would this situation lead to manifest
esotropia?
Hoffmann: You have symmetrical binocular OKN
responses in rats or rabbits which have totally
asymmetric systems all their life. The system being
supplied with direction specific error signal just
from one eye works perfectly. The two eyes work
together by supplying an error signal for one
direction each, for the leftward movement it would
be the right eye and for the rightward movement it
would be the left eye. There seems to be no problem
that the other eye isn't seeing that particular
direction.
Kommerell: The squinting eye is not completely sup-
pressed, we know this from psychophysical investi-
gations. In particular, the periphery of the
squinting eye contributes to binocular vision. So,
I don't see a difficulty in explaining symmetrical
OKN with both eyes open. As to the other part of
the question, I don't see how asymmetrical OKN
should lead to esotropia, but strabismus may well
lead to an asymmetry of OKN.
Bagolini: This is an amateur question. The exci-
tation of the retina goes directly to the opposite
nucleus of the optic tract in about 0.7 msec. The
excitation from the retina also reaches the nucleus
through a long loop passing through the cortex. The
signals that reach the NOT via the cortex should be
out of phase with the excitation coming directly
from the eye. Since the eye moves quickly, the
excitation from the two retinae should frequently
not match. I am curious to know more about this
problem.
Hoffmann: It turns out that in the cat different
conduction velocities in the subcortical projection
and the cortical loop more or less lead to the same
overall latencies at the subcortical target
structure. The projection through the cortical
relay is much faster than the direct projection
from the retina to the NOT and I would assume that
also in the monkey the fastest conducting retinal
axons supply the information to the cortex from
which direction specific information is then
elaborated and send to the NOT. In fact the two
informations may arrive simultaneously. In addition
these conduction velocities are only a fraction of
the overall visual latency with about 25-30 msec in
the retina which has to be added to the conduction
time along the optic tract. The third point, OKN is

a global phenomenon and the most important para-
meter for this response is that movement on
different parts of the retina are in the same
direction. The temporal sequence doesn't seem to be
so important.
Discussion of the paper by C.Schor on "The phasic-

tonic organization of accommodation and vergence".
Bagolini: You seem to consider both the fast and

the slow integrator to be central and at the
midbrain level. Could it instead be that the slow
integrator is a slow relaxation process of the
muscles and thus peripheral?
Schor: The main evidence has to do with the
duration of the time constant for the decay of
these after effects. The longest time recorded for
the durations of a prism vergence after effect is
about 8 hours. It is hard to imagine any cell in
the extraocular muscle with such a long decay time
constant. I assume then that adaption would have
central neurological origin.
Lennerstrand: Relaxation of even the slowest of the
muscle fibres in the extraocular muscles, the
amphibia-like multiply innervated fibres, would be
completed within a few seconds, which excludes a
peripheral cause for the slow adaptation.
Sjbstrand: I have a question both to Dr Schor and

Dr Aslin. As a clinician I am intrigued by the fact
that we haven't discussed the effect of hyper-
metropia. I think it is important to consider the
epidemiological studies of Ingram and coworkers*
showing that one of the risk factors in recruiting
squinters is a high hypermetropia at an age of 1
year. It would be very important to know the
characteristics of interaction between vergence and
accommodation in adults or during development as Dr
Aslin has studied. Does accommodation effort to
overcome hyperopia create the convergence squint or
is an excessive convergence induced voluntarily to
further stimulate accommodation?
*Ingram et al (1986). Prediction of amblyopia

and squint by means of refraction at age 1 year.
Br.J.Ophthalmol., 170, 12-15.
!58 G.LENNERSTRAND
Schor: There may be cases where children converge
excessively to utilize convergence accommodation to
overcome hyperopia. This was Fincham's* coarse
adjustment mechanism of accommodation. I am
suggesting that when the AC/A ratio is high,
accommodation is usually unadaptible. But it is
also clear that in high hypermetropia with a high
AC/A ratio there is excessive convergence at both
far viewing distances and at near ones. If there is
no hypermetropia but there is a high AC/A ratio you
might expect an esotropia if the person favours
clear vision instead of single vision. However,
instead of favouring clear vision many subjects
favour single vision and they simply do not
accommodate accurately at near distances. When they
do accommodate it is a very slow response. In this
way they avoid percipitation of esotropia.
Aslin: It seems to me that it would be useful for
both clinicians and basic researchers to consider
the amplitude of accommodation in patients who are
hypermetropic. One could imagine high hypermetropes
who have a large accommodative amplitude and
therefore do not lead themselves into accommodative
esotropia. I wonder if perhaps the amplitude of
accommodation could be potentially useful in the
clinic?
Schor: The real question is whether accommodative
convergence increases linearly with accommodative
response amplitude. Near the limit of accommoda-
tion, accommodative vergence increases non-linear-
ly. This is usually seen in pre-presbyopia when the
AC/A ratio suddenly increases. If you had a very
limited amplitude of accommodation you would expect
to get much more accommodative vergence than if you
could accommodate in the initial portion of the
accommodated range. The non-linearity results from
variation in effort required to accommodate
(myodiopter).
Lennerstrand: Do you have any suggestions for the
orthoptists with respect to the type of training
procedures to strengthen the weak adaptor and
reduce the functions in the excessive adaptor.
*Fincham,E.F. and Walton,J.(1957). The reciprocal

actions of accommodation and convergence. J.
Physiol. (Lond) 137, 488-508.
Schor: I doubt very much if you can permanently

weaken the stronger adaptor with ramp tracking
exercise but Vaegan* has shown that you can indeed
strengthen adaptors with the tonic exercise. This
approach is to stimulate sustained convergence and
sustained accommodation which results in marked
increases in the amount of adaptability. David
Hanson who is at Cardiff, also has shown similar
results. He has found that some disorders of con-
vergence and accommodation are associated with the
lack of adaptation. Using these tonic exercises he
has found an improvement of adaptation of the
vergence system.
Campos: I believe that Dr Schor's findings may have
a clinical application only in patients orthophoric
for distance and esotropic for near; i.e. in the so
called accommodative strabismus of non-refractive
type.
Goller: I would like to ask once more perhaps to
clearify Dr Sjostrand's question. Do you think that
hyperopia is just a question of an inherited short
axial length or is it an acquired desease?
Schor: Hypermetropia is usually considered a

genetic disorder. However, visual deprivation can
also produce large refractive disorders.
Sjostrand: We know that we have a higher frequency

of bilateral amblyopia in highly hypermetropic
children as shown in Ingram's studies. This
indicates that they probably have a blurred retinal
image in both eyes. Perhaps, Dr Haase can present
his data concerning the risk of amblyopia in
hyperopic children.
Haase: The group of squinters who do not suffer
from middle or a high degree of ametropia (<2.5 D)
amounts to 20 % of the strabismic patients. 80 %
have either myopia, hypermetropia or astigmatism.
In 2 % of the population of six years old children
we found high hypermetropia and/or astigmatism as
an amblyopogenic factor in both eyes. They all
suffered from bilateral amblyopia. They needed
months to years in order to improve their visual
acuity in both eyes after the correction of the
ametropia.
*Vaegan,D.R.(1979). Con- and divergence show large

and sustained improvement after short isometric
exercise. Am.J.Optom.Physiol.Optics, ~. 23-33.
160 G.LENNERSTRAND
Schor: Actually there may be a connection in the
emmetropization process and the AC/A ratio in which
there is some active tuning of refractive error.
Adaptability of the tonic process of accommodation
may be inheritly involved in the process of
emmetropization. In cases where you have extremely
high refractive errors, a high AC/A ratio could be
due to a failure of the adapted process of accommo-
dation. You would expect to find a higher accommo-
dation convergence ratio without the adaptive
process because of the great effort of the optical
reflex accommodated process, which stimulates
accommodative vergence. Accordingly there is a
possible connection. However, you might also expect
to find a lot of high myopes with high AC/A-ratios
for the same reasons.
Lennerstrand: Do you think there is a role for
visual feedback in adaption?
Schor: Fred Miles has suggested that a stimulus
conflict can result in adaptation. Initially
conflict was used to study adaptation of the
vestibular ocular reflex in which head rotation was
unequal to retinal image rotation when a magnifier
was worn before the eyes. He and Judge also
attempted to change accommodative convergence
ratios in response to conflicts between accom-
modation from one distance and convergence for a
different distance. In our procedures we have no
feedback to manifest this conflict. Accommodation
was stimulated monocularly so that the vergence
system had no knowledge of its accuracy or we
stimulated vergence with pinholes before the eyes
so that accommodation had no knowledge of its
accuracy. However, we are still changing AC/A and
CA/C ratios in a way that would bring them towards
more normal values. Consequently the feedback is
really not necessary to change the cross coupling
between accommodation and vergence although it may
still improve the modifications or accelerate them.
Discussion of the paper by B.Bagolini on "Clinical

aspects of vergent mechanisms".
von Noorden: The existence of anomalous fusional
movements cannot be doubted but I am not sure
whether I can wholeheartedly agree with the
etiologic mechanism that Dr Bagolini suggested
today, namely that they occur on the basis of
anomalous retinal correspondence. First these
movements can also be elicited by prism in patients

with strabismus and normal retinal correspondence.
Second, the mechanism suggested by Dr Bagolini is
incompatible, in my opinion, with the point-to-area
character of anomalous correspondence. If there are
any number of retinal elements in the deviated eye
that are capable of forming a common visual
direction with the fovea of the fixating eye, why
should the eyes return to a particular position if
the deviation is neutralized with prisms? And
third, if these anomalous fusional movements occur
on the basis of anomalous correspondence why don't
most of our patients with infantile esotropia have
a recurrence of their original deviation after
surgery?
Bagolini: Anomalous movements when sufficiently
strong bring the retinal images over areas of
acquired anomalous correspondence. I have seen it
hundreds of times in cases carefully studied. I
repeat that for anomalous movements I mean those
movements that patients with convergent squint
perform when we correct the angle of strabismus
with base-out prisms. It is a movement that is not
always performed by all patients. It is usually
evident in those who have been squinting for a
rather long time. They tend to bring the retinal
images over anomalous corresponding areas and not
to a specific point inside these areas. Then I
would like to point out that anomalous retinal
correspondence is a modification of the retinal
directional localization in binocular vision.
Anomalous movements are instead a modification of
the motorial value of the retinal elements in
binocular vision. You do not detect them mono-
cularly. Putting together these elements it seems
logical to think that both are adaptions to the
deviation and have the same aim. The aim seems to
be that of achieving an anomalous type of binocular
vision in spite of a deviated eye. If the aim is
this, these anomalous movements should be con-
sidered fusional. What triggers them? They are
triggered by displacing the retinal images just as
for normal fusional movements. I don't know if I
have answered all of your questions. I believe you
further mentioned that you have observed them in
patients with normal correspondence without an
angle of strabismus.
von Noorden: I am talking about heterotropic
subjects with normal retinal correspondence that
can eat up prisms just like the patients that you
describe.
162 G.LENNERSTRAND
Bagolini: If there was normal correspondence they
were probably normal fusional movements, I have to
mention another important aspect. Anomalous retinal
correspondence and anomalous movements are probably
two different adaptational phenomena to the
strabismic deviation. They may develop in different
times. I have seen many cases with normal retinal
correspondence after treatment which still had
anomalous movement and vice versa. Anomalous
movements and anomalous retinal correspondence are
two adaptational phenomena which may develop in
different times and they are not necessarily
interwoven like the normal fusional movements and
the normal retinal correspondence. In strabismus
these two aspects of binocular vison may be
dissociated.
Harcourt: Dr Bagolini mentioned that the explana-
tion for the blockage of nystagmus by base-out
prisms might be the inhibition of lateral rectus
tone induced by convergence. Would one not then
expect that base-in prisms would have a similar
effect by inhibiting medial rectus tone. Is it just
that sufficient divergence cannot be induced in
this way?
Bagolini: I think that only base-out prisms can be
used because divergent fusional movements are very
limited, so you can not rely on them for blocking
nystagmus.
Westheimer: Could the aim of the anomalous fusional
movements be merely to move the fovea of one eye on
to the blind spot of the other?
Bagolini: No. For example, in microstrabismus these
movements are very strong. They may slowly compen-
sate even 40 or 60 base-out prisms by steps of 10-
15 prism diopters. The retinal image is in the
deviated eye quite far from the blind spot in these
patients.
Schor: Dr Bagolini, what makes you think there are
no phasic disparity vergence responses in anomalous
retinal correspondence? It is clear that they have
no fast disparity vergence responses but they could
have very slow ones that are controlled by the
phasic and not the tonic process. There could still
be small responses to disparity which could then be
integrated by the adaptable tonic process to
produce anomalous slow movements.
Bagolini: We are tempted to compare the slow

movements I have discribed with normal fusional
movements. I don't know, referring to your model,
what substitutes the "fast integrator" to trigger
this type of "slow integrator". I am looking to see
if accommodation may trigger this phenomenon.
Accommodation however works only for horizontal
movements and Campos has seen in esotropic patients
that if you use vertical prisms you have a small
slow vertical movement which can not be triggered
by accommodation. So perhaps the hypothesis of
accommodation as a triggering factor is not
tenable.
Schor: Yes, I agree. However, if you compare the
velocity of disparity vergence stimulated with
foveal targets and parafoveal targets, as we have
done*, you find that with foveal targets the eyes
will converge as much as 4 or 5 degrees per second
per degree of disparity. But as soon as you move
targets on to the peripheral retina the velocity of
vergence drops. Perhaps in strabismus with central
suppression there is more periferal fusion, which
would produce a much slower velocity of fusional
vergence but yet disparity vergence could still be
evoking these adaptive responses.
Discussion of the paper by R.Aslin on "Normative

oculomotor development in human infants".
Levi: I was interested in Dr Aslin's suggestion
that the eye movements of infants may not be
perfectly conjugate. Have you done any calculations
as to how dysconjugate they might be? Do you know
the magnitude of Panum's fusional area in infants,
as that would impact on how dysconjunctive eye
movements might effect sensory development.
Aslin: With regard to the first question, it is of
course difficult to say because we don't have
simultaneous eye movement recordings from the two
eyes. However, one thing we have noted when
recording from a single eye is that for horizontal
saccades there is very little vertical cross-talk
or scatter whereas for vertical saccades there is
considerable horizontal cross-talk. For vertical
*Schor,C.M., Robertson,K.M. and Wesson,M.(1986).

Disparity vergence dynamics and fixation disparity.
Am.J.Ophtom. Physiol.Optics, 63, 611-618.
164 G.LENNERSTRAND
saccades, either the infant is not able to execute
reliably a vertical rotation or there is inter-
ference from the other eye that leads to slight
dysconjugacy. The magnitude of horizontal cross-
talk on vertical saccades is up to 4 degrees,
suggesting the presence of this degree of dys-
conjugacy. With regard to the size of Panum's
fusion area, if you believe the prism data that I
presented, it suggests that Panum's area may be as
large as 4 to 5 degrees until the fourth month
after birth.
Eggers: A comment in regard to Dr Aslin's remarks
on the need for recalibration of the saccadic
system with growth. One thing that remains constant
is the percent contraction of the muscle in length
for a given degree of gaze shift. This provides
simplification of the muscle innervation required
to move the eye. You also implied that as the cones
migrate into the fovea in postnatal development
they drag the bipolar cells and the ganglion cells
along with them over large distances. I would
suggest perhaps that the cones migrate and change
the bipolar cell they talk to, thus not having to
drag the neural tissue so far along with them.
Aslin: With regard to your second point, the
evidence presented by Yuodelis and Hendrickson
(1986*) strongly suggests that the latter situation
does not hold. That is, the connectivity of cones
to bipolar cells does not change postnatally. More-
over, the 300 micron length of the fibres of Henle
can be taken as an estimate of the magnitude of
cone migration. If you use that estimate rather
than the one I presented in my talk, you still get
significant postnatal shifts in the local sign of
photoreceptors stimulated at a given retinal
eccentricity. These shifts are on the order of 1 to
2 degrees, but the point is that the combination of
optical and photoreceptor changes requires some
degree of oculomotor control recalibration post-
natally.
Blakemore: My understanding of the results of
Hendrickson and her colleagues is that the in-
creased packing of photoreceptors is mainly
restricted to the fovea. Indeed, the lengthening of
*Yuodelis,C. and Hendrickson,A.(1986). A quali-

tative and quantitative analysis of the human fovea
during development. Vision Res., ~. 847-855.
the fibres of Henle is presumably due to the

relative migration of foveal receptors and their
associated bipolar and ganglion cells. The major
changes in the relationship between retinal
position and directional sign might apply to the
central retina.
Aslin: It must also apply to the retinal periphery
if the migration is monotic across the retina,
although the absolute magnitude would be inversely
proportional to retinal eccentricity. Given a
constant number of photoreceptors in the postnatal
retina, the migration towards the fovea must come
from the periphery. However, one could assume that
there is initially a maximum density of cones in an
annular arrangement around the eventual fovea. This
would not require any migration further than
approximately 5 degrees from the fovea. Unfortu-
nately, we simply do not have the anatomical data
to know whether this is the case. But you are
correct in pointing out that photoreceptor migra-
tion would be greatest within the central 5 degrees
of the retina.
Discussion of the papers by B. Harcourt on "Aetio-

logy, classification and clinical characteristics
of esotropia in infancy" and by G.Lennerstrand on
"Motor dysfunction in strabismus".
Harcourt: If some types of infantile esotropia are
induced by an inherent defect in potential bin-
ocular function, it surprises me that there is not
a stronger hereditary tendency.
von Noorden: Hydrocephalus is frequently associated
with manifest nystagmus. The question arises
whether the esotropia in such cases is secondary to
the nystagmus (dampening by convergence) or whether
it is an independent entity. With regards to your
last comment there is considerable evidence of a
multifactorial genetic transmission of strabismus.
Harcourt: Accommodative esotropia certainly does
have a strong familial tendency, but my experience
is that this is not nearly so common in essential
infantile esotropia with nystagmus. This was
emphasised by Lang*.
*Lang,J.(1968)."Squint dating from birth or with
early onset", Proceedings of the First Inter-
national Congress of Orthoptics,p. 231-7. Kimpton,
London.
166 G.LENNERSTRAND
Held: What do the monkey data say about this?
Lennerstrand: I would like Dr Eggers to make some
comments on the incidence of strabismus in monkeys
and perhaps modify the information that was given
during my presentation.
Eggers: I have been working with Ron Boothe who is
now at Emory University, Atlanta. He brought his
strabismic nemestrina macaque monkeys from Seattle
and he has continued to make trips to Seattle to
find new strabismic animals. When the animals are
brought in for yearly health check it is possible
to get a quick look at the animal. Those that look
esotropic he reexamines in more detail later. I
think it is very difficult to say what the true
incidence of naturally occuring strabismus is in
monkeys. We presented Boothe's animals at the
Association for Research in Vision and Ophthalmo-
logy in May 1987*. We have about 8 animals now that
form a homogeneous group. They are esotropic and
show hyperopia greater than the population norm.
They also show anisometropia, which is unusual in
this animal.
*Eggers,H.M. and Boothe,R.G.(1987). Naturally

occuring accommodative exotropic in macaques.
Invest. Ophthalmol. Suppl., 28 (3):103.
Session II
NORMAL AND ABNORMAL
VISUAL DEVELOPMENT
13
AMBLYOPIA IN HUMANS AND CLINICAL RELEVANCE
OF ANIMAL MODELS
GUNTER K. von NOORDEN
Amblyopia means literally "dullness of vision" and is defined

as subnormal visual acuity in one or both eyes which on physical
examination appear normal. This acuity deficit cannot be improved
with corrective spectacles but, if treated early in life by
occlusion of the fellow eye, is partially or completely reversible.
Amblyopia occurs in 2-2.5% of the population and develops only
during childhood; patients older than eight years are resistant to
amblyopiogenic conditions. There are three clinical conditions
that cause amblyopia: strabismus, anisometropia and form vision
deprivation.
This afternoon I shall briefly review the amblyopiogenic
mechanisms for each of these conditions. Suffice to say here that
there is strong clinical evidence, substantiated by experiments in
animal models that both the conflicting visual input from the two
eyes and form vision deprivation, acting in unison or individually,
can be-5ingled out as amblyopiogenic mechanisms (von Noorden,
1985).
Animal research has added an exciting new dimension to our
quest to solve the riddle of amblyopia. Strabismic, anisometropic
and form vision deprivation amblyopia have been successfully
produced in monkeys by experimentally altering the visual input to
the eyes during infancy in macaque monkeys, a species whose visual
system parallels that of the humans in terms of development,
function and anatomy (Booth et al, 1985) (Garey, 1987). As in our
human patients the amblyopia in monkeys occurs only during infancy
and early childhood and is reversible by occluding the fixating eye
and enforcing the usage of the amblyopic eye.
Supported by grants EY 01120, EY 07001 and EY 02520 from the

National Institutes of Health
169
170 G.K. von NOORDEN
Severe histological changes have been described in the lateral

geniculate nucleus (LGN) of monkeys, with amblyopia caused by
strabismus, anisometropia or form vision deprivation (for a review
see Booth et al, 1985). Does the human visual system react in a
similar manner? We can answer this question affirmatively, at least
as far as anisometropic amblyopia is concerned after having shown
that cells in the parvocellular layers of a human LGN connected
with the amblyopic eye were on the average 18% smaller than those
receiving input from the normal eye (von Noorden et al, 1983).
There is no reason to doubt that similar anomalies exist in LGN's
of patients with strabismic or form vision deprivation amblyopia
even though histologic proof is still lacking.
Thus, the relevancy of the monkey model for further study of
amblyopia is firmly established. While animal research never
should replace astute clinical observations, psychophysical,
electrophysiological and histological exploration of the amblyopic
phenomenon in animal models offers several advantages.
Extracellular recordings from and histological analyses of various
parts of the afferent visual system have provided data that could
never have been assembled from clinical research. Moreover, unlike
in humans the onset and duration of the amblyopiogenic events can
be precisely defined in animal models and their effects on the
immature visual system be studied in a prospective and controlled
fashion. In humans, the onset of strabismus, of anisometropia or
of a visually depriving cataract is often uncertain and our
knowledge about the sensitivity of the human infantile visual
system to abnormal visual stimulation is based on retrospective
data with all their intrinsic limitations. As a result we know less
about sensitive periods in humans than what we have learned from
animal models.
While animal research has contributed to our understanding of
various forms of amblyopia our search for a useful animal to study
the neurophysiology and neuroanatomy of essential infantile
esotropia and of exotropia continues. Unlike in accommodative
esotropia or paralytic strabismus or strabismus caused by
structural anomalies of the extraocular muscles or the orbit, the
etiology and pathophysiology of these frequently occurring forms of
strabismus have remained essentially unknown.
A glimmer of hope appeared on the horizon with the report of
naturally occurring esotropia in a colony of Macaca nemestrina
(Kiorpes et al 1985). However, a communication at the recent ARVO
meeting established that significant degrees of hypermetropia were
present in most of these animals (Eggers and Booth, 1987), raising
the distinct possibility that they were afflicted with
accommodative rather than essential infantile esotropia. A
AMBLYOPIA IN HUMANS 171
prom1s1ng beginning to link neuroanatomic anomalies with strabismus

was made by the contributions of Guillery and his co-workers (Kaas
&Guillery, 1974) (Guillery et al, 1984). These authors showed in
Siamese cats and albino monkeys that some of the retino-fugal
fibers are routed wrongly through the optic chiasm. This causes a
segment of temporal retina to be abnormally connected with the
contralateral (rather than ipsilateral) hemisphere. Thus, the
visual cortex in albino cats, monkeys and also in the human albino
(Witkop et al, 1982) receives contradictory messages from the two
eyes. The behavioral consequence of this abnormal arrangement in
terms of the etiology of strabismus so frequently found in albinos
is not yet fully understood. However, Guillery's work has raised
the distinct possibility that at least some forms of strabismus may
be caused by the sensory consequences of abnormal neuroanatomic
connections in the afferent visual pathways.
Further progress in our efforts to clarify amblyopia and
strabismus will clearly depend on an interdisciplinary approach.
Neither the clinician nor the basic scientist can afford any longer
to work in isolation. It is the purpose of this symposium to bridge
the gap between basic and clinical research and it is my hope that
all of us here will return to our hospitals and laboratories with a
better understanding of each other's work and a bagful of new
ideas.
References
1. Boothe, R.G., Dobson, V., and Teller, D.Y. (1985). Postnatal
development of vision in human and nonhuman primates. Annu.
Rev. Neurosci., 8, 495-545.
2. Eggers, H.M., Boothe, R.G. ( 1987). Naturally occurring
esotropia in macaques. Invest. Ophthalmol. Vis. Sci. (suppl)
28, 103.
3. Garey, L. (1987). Normal anatomic development of the primate
primary visual pathway. Wenner Gren Center Symposium
Strabismus and Amblyopia - Experimental Basis for Advances
in Clinical Management, Stockholm.
4. Guillery, R.W., Hickey, T.L., Kaas, J.H., Fellman, D.J.,
de Bruyn, E.J., Sparks, D.L. (1984). Abnormal visual pathways
in the brain of an albino green monkey (Cercopithecus
aethiops). J. Comp. Neurol., 226, 165-183.
5. Kaas, J.H., Guillery, R.W. (1974). The transfer of abnormal
visual field representatives from the dorsal lateral geniculate
nucleus to the visual cortex in Siamese cats. Brain Res., 59,
61-95.
6. Kiorpes, L., Boothe, R.G., Carlson, M.R., Alfi, D. (1985).
Frequency of naturally occurring strabismus in monkeys. J.
Pediatr. Ophthalmol. Strabismus, 22, 60-64.
7. von Noorden, G.K. (1985). Amblyopia: a multidisciplinary
approach. (Proctor Lecture) Invest. Ophthalmol. Vis. Sci., 26,
52-64.
~. von Noorden, G.K., Crawford, M.L.J. and Levacy, R.A. (1983).

The lateral geniculate nucleus in human anisometropic
amblyopia. Invest. Ophthalmol. Vis. Sci. 24, 788-790.
9. Witkop, C.J., Jay, B., Creel, D., GuillerY: R.W. (1982).
Optic and otic neurologic abnormalities in oculocutaneous and
ocular albinism In Genetics in Ophthalmology (eds. E. Cotlier
and E. Berman). Alan R. Liss, 299-318.
14
NORMAL ANATOMICAL DEVELOPMENT OF THE
PRIMATE PRIMARY VISUAL PATHWAY
L. GAREY
For many years a considerable effort has been made

to understand the possible underlying physiopathologi-
cal mechanisms of amblyopia by the study of animal
models. Initially the cat and, more recently, the
monkey have been extensively used in such studies
(Wiesel, 1982) • This experimental approach has shown
that during a "critical" period in the first few weeks
of the animal's life a normal visual environment is
necessary to prevent major structural and functional
abnormalities occuring in the developing visual system,
and especially the lateral geniculate nucleus (LGN) and
visual cortex. Attention has also been turned to the
normal development of primate visual pathways, and the
comparison of such development in human and non-human
primates. Such comparisons will certainly be of value
in possible extrapolations from monkey to human patho-
logical situations.
RETINA
In the foetal monkey there is a migration of neu-
rons away from the fovea, leading to a thinning out of
the bipolar and ganglion cells. At birth the retina
has a fairly mature appearance, even to the extent of
there being a true foveal pit (Samorajski et al., 1965;
Hendrickson and Kupfer, 1976). In spite of this, there
are still subtle changes that occur only postnatally,
especially concerning synapse formation (Smelser et
al., 1974; Nishimura and Rakic, 1985) and an increase
in length and numerical density of cone photoreceptors
(Hendrickson and Kupfer, 1976).
The peripheral portions of the human retina are
already fairly mature at birth. However, the central
173
174 L. GAREY
parts, especially in the foveal region, remain relati-

vely immature to histological inspection (Bach and
Seefelder, 1911; Mann, 1964) and this is also true of
the development of synaptic contacts (Hollenberg and
Spira, 1973; Spira and Hollenberg, 1973). More recent
work has demonstrated that in man the rudimentary fovea
is already distinguishable in the mid-term foetus as a
thick layer of ganglion cells and a zone marked by the
absence of rods (Figure 1). The foveal pit is visible,
but not yet fully formed, just after birth, for bipolar
and ganglion cell migration is not complete. Something
approaching maturity is only reached in the second year
postnatally, but even then full photoreceptor matura-
Figure 1. The maturation of the human fovea. A: 22-

week-old foetus which lacks a foveal depression. B: 24
to 26-week-old foetus with beginning of foveal depres-
sion (arrow). C: 5 days postnatal, with deeper depres-
sion due to thinning of ganglion cell layer. D: 15
months old showing almost complete lack of inner
nuclear and ganglion cell layers at arrow. E: 45 months
old with only cones and glia at the fovea. F: Adult
fovea. (Cr =choroid; P = photoreceptors). From Hend-
rickson and Yuodelis (1984), with permission.
NORMAL DEVELOPMENT OF THE PRIMARY VISUAL PATHWAY 175
tion continues until nearly four years of age (Abramov

et al., 1982; Hendrickson and Yuode1is, 1984).
LATERAL GENICULATE NUCLEUS

The human LGN doubles in volume during the first
six months of postnatal life, from about 70 to 140 mm3,
after which it remains stable throughout infancy and
adult life (de Court en and Garey, 19 8 3 ; Garey and de
Courten, 1983). A similar increase in LGN volume be-
tween birth and adulthood has been described in non-
human primates. For example, Fritschy and Garey
(l986b) showed that in the New World marmoset monkey
the volume almost doubles in the first week, reaching
about three times the birth volume of approximately 4
mm3 in the second postnatal month, and then decreases
again by about 25% to the adult figure of about 11 mm3
(Figure 2). Smaller changes have been described in the
Old World macaque (Gottlieb et al., 1985).
12 I
VOLUME
.r-·-·, .,.,
>-
I ., (J)
z
10
I-
./ 0
CY
(J)
z ./ ::::J
w ./ w
8 z
D
./ LJ._
0
w
:E 6 ./· 600 CY
w
::::J
_J
0
/
---------------~--~
-- --~f.l..M_I!~.Ef
------------
NEURONS CD
:E
400 ::::J
> 4 z
DENSITY
2 200
Birth 7 28 42 180 Adults

AGE
Figure 2. Postnatal changes with age (in days) in LGN
volume (in mm3, left scale), neuronal numerical density
(xlo4 neurons;mm3, also on left scale) and total number
of neurons (xlo3, right scale). From Fritschy and Garey
(1986b).
176 L.GAREY
In man, the lamination of the LGN is laid down

between 22 and 25 weeks of gestation (Hitchcock and
Hickey, 1980b). As to the size of somata of human LGN
neurons, Hickey (1977, 1981) described a rapid growth
of parvocellular cells during the first six months
after birth, followed by a slower growth period until
the end of the first year. Magnocellular cells also
grow during this time, but continue rather longer than
the parvocellular cells, until at least two years of
age. This period corresponds to that when the develop-
ing human visual system is particularly sensitive to
the effects of visual deprivation (Awaya et al., 1973;
Von Noorden, 1977). In monkeys, the parvocellular
neurons reach adult size within a few days after birth,
but magnocellular growth continues up to adulthood
(Headen et al., 1981).
Fairly recently, more attention has been turned to
the question of the development of the dendritic trees
of LGN neurons. De Courten and Garey (1982) demonstra-
ted that the various neuronal types described in the
adult LGN are already identifiable at birth, and even
in the late foetus. However most neurons, especially
the multipolar variety, undergo considerable changes
during their maturation. The most striking difference
between immature and mature LGN dendrites is that the
former bear large numbers of spines (de Courten and
Garey, 1982) . In the late foetus and at birth den-
drites already have moderate numbers of spines and bear
filopodia, growth cones and "hair-like" processes. The
maximum number of these various profiles is reached at
about four months postnatally, after which their number
decreases. The filopodia, growth cones and hair-like
processes disappear first, and by nine months very few
spines remain, so that the dendrites take on a mature
appearance. The diameter of the dendritic arbor in-
creases during this period. At birth the average
multipolar dendritic arbor is between 100 and 200 um in
diameter, reaching 300 um by two months and 400 um by
four months. By nine months the "adult" diameter is
reached, many neurons having a dendritic tree diameter
of 600 um. Similar overproduction of spines, together
with early proliferation of growth cones and filopodia,
also occurs in the sub-human primate LGN where the
morphology of LGN neurons is similar to that in man
(Garey and Saini, 1981; Saini and Garey, 1981).
The actual mean lengths of individual dendrites
also increase after birth, as does the total length of
the dendritic tree of a given neuron (Leuba et al.,
1985), although this increase is restricted to the
terminal branches of the dendrites. The same is true

for monkeys, both Old and New World (Leuba and Garey,
1984; Fritschy and Garey, 1987), but in these non-human
forms there is also a decrease in the total and mean
dendritic lengths after birth, once again involving the
terminal dendritic segments. Further analysis is
necessary to ascertain whether human LGN dendrites are
also subject to this late shrinkage.
It is now known that a common feature in the deve-

loping monkey visual system is loss of neuronal ele-
ments. The number of optic nerve fibres decreases to
reach adult levels by the second month postnatally
(Rakic and Riley, 1983), and after an early postnatal
increase retinogeniculate synapses are eliminated over
a period extending to some four months after birth
(Holstein et al., 1985).
VISUAL CORTEX
As for the LGN, the human area 17 increases in
volume rapidly during the perinatal period. It quad-
ruples between 28 weeks of gestation and birthj when
the volume of a single area 17 is some 1. 5 em . It
quadruples again by about four months of age (Rutten-
locher et al., 1982), reaching a volume of about 6 cm3,
that is maintained into adulthood. The basic laminar
pattern is already visible at birth in area 17 of Old
and New World monkeys (Wiesel and Hubel, 1974; Blake-
more et al., 1979; Fritschy and Garey, 1986a) and man
(Leuba and Garey, 1987). In man, cortical lamination
is already recognizable before mid-gestation, but the
various sublaminae are only really distinguishable from
about 26 gestational weeks (Takashima et al., 1980;
Sauer et al., 1983). The stria of Gennari, in layer
IVb, and the cell dense granular layer IVc are visible
by mid-gestation (Leuba and Garey, 1987) thus making
area 17 easily distinguishable from area 18 in the
foetus.
In man there is no evidence for any loss of neu-
rons between mid-gestation and old age, for there are
compensatory decreases in cell density and increases in
cortical volume (Leuba and Garey, 1987). The neuronal
numerical density at mid-gestation is over one million
per mm3, decreasing to 90,000 by birth. There is a
further decrease to about 40, 000 by some four months
postnatally, a value that is approximately maintained
in the adult. However, 0 1 Kusky and Colonnier ( 1982)
did describe a moderate neuronal loss in macaque area
17 between birth and adulthood. As there is a more
178 L. GAREY
definite neuron loss with aging in rodents (Heumann and

Leuba, 1983), it is possible that there is in fact
cortical neuronal death with age in lower mammals and
some primates, but not in man, although Fritschy and
Garey (1986a) found no such cell loss in the New World
marmoset area 17.
There are various changes in dendritic morphology
in developing human visual cortex (Takashima et al. ,
1980). Among other changes, dendritic spines, at least
on layer III pyramidal cells, double in number between
about 33 foetal weeks and birth and continue to in-
crease until about five months postnatally. The spine
number then decreases again until two years of age when
approximately the same number is found as at birth
(Michel and Garey, 1984). This phenomenon is similar,
although less marked, to that described above for spine
loss in the LGN (de Courten and Garey, 1982). In
monkeys, there is also visual cortical dendritic spine
loss after a peak reached at about two months postna-
tally (Lund et al., 1977~ Boothe et al., 1979).
At the ultrastructural level there are major chan-

ges in the synaptic organization of area 17 detectable
during the perinatal and adolescent periods in both man
and monkey. In man, synaptic density increases rapidly
between late gestation and about eight months postna-
tally, when a maximum is reached (Huttenlocher et al.,
1982). It has been calculated that this maximum repre-
sents about 3. 5 x 1012 synapses in a single area 17.
After this age there is loss of about 40% of these
synapses, so that by apP.roximately 11 years of age
there are only 2.1 x lol2 contacts recognizable, a
figure that is only slightly higher than that recorded
at birth. A similar phenomenon had been reported in
monkey visual cortex (O'Kusky and Colonnier, 1982~
Rakic et al., 1986) where the maximum synaptic density
is reached around four months postnatally, followed by
a decline to adult levels by about three years of age.
This decline is most noticeable for synapses onto
dendritic spines.
Since the presence of ocular dominance columns was
shown in area 17 of the Old World monkey (Wiesel et
al., 1974), various attempts have been made to demon-
strate then in the human. As it is not possible to use
invasive experimental techniques in man, only those
methods that can be applied to post mortem material can
be utilised, but good evidence has now been obtained
that ocular dominance columns do exist in man (Hitch-
cock and Hickey, 1980a) • These columns may provide
basic structural elements for the primate visual cor-
tex. More recently, new modular units have been de-

scribed in sub-human primate and in human visual cor-
tex. They include neuropil patches that stain for the
mitochondrial enzyme cytochrome oxidase, and that are
concentrated in layers II and III at the centres of
ocular dominance columns (Hendrickson et al., 1981;
Horton, 1984; Horton and Hedley-Whyte, 1984) . There
would, thus, appear to be a basic structural similarity
between the human and monkey visual cortex, although
ocular dominance columns do not seem to be ubiquitously
present in New World primates (Hendrickson et al. ,
1978; Spatz, 1979).
CONCLUSION
In both man and monkey the periods during which
visual function is developing most rapidly are also the
times when the visual system is most susceptible to
permanent damage if the visual environment is abnormal.
In the human, visual acuity improves during the first
half of the first postnatal year (Dobson and Teller,
1978; Pirchio et al., 1978; Held, 1979), a time when
interference with the normal visual environment is most
likely to cause amblyopia (Awaya et al. , 1973; Von
Noorden, 1977, 1981). This is just the period when the
changes described above are at their most dramatic.
Similar critical periods exist in the monkey, when
acuity is increasing and experimental visual depriva-
tion is most devastating (Vital-Durand et al., 1978;
Blakemore and Vital-Durand, 1979; Wiesel, 1982) but
these periods occur earlier than in man. It has been
suggested that the processes in man and monkey are
parallel but about four times faster in monkey (Teller,
1981).
REFERENCES
Abramov, I., Gordon, J., Hendrickson, A., Hainline, L.,
Dobson, V. and LaBossiere E. (1982). The retina of the
newborn human infant. Science, 217, 265-267.
Awaya, s., Miyake, Y., Imaizumi, Y., Shiose, Y., Kanda,
T. and Komuro, K. (1973). Amblyopia in man, suggestive
of stimulus deprivation amblyopia. Jap. J. Ophthalmol.,
17, 69-82.
Bach, L. and Seefelder, R. (1911). Atlas zur Entwick-

lungsgeschichte des menschlichen Auges. Engelmann,
Leipzig.
180 L. GAREY
Blakemore, c., Garey, L.J. and Vital-Durand, F. (1979).

Developmental plasticity in the monkey visual system.
In Neural Growth and Differentiation (eds. E. Meisami
and M.A.B. Brazier), Raven Press, New York.
Blakemore, c. and Vital-Durand, F. (1979). Development
of the neural basis of visual acuity in monkeys. Specu-
lation on the origin of deprivation amblyopia. Trans.
Ophthalmol. Soc. U.K.
99, 363-368.
Boothe, R.G., Greenough, W.T., Lund, J.S. and Wrege, K.
(1979). A quantitative investigation of spine and den-
drite development of neurons in visual cortex (area 17)
of Macaca nemestrina monkeys. J. Comp. Neurol., 186,
473-490.
Courten, c. de and Garey, L.J. (1982). Morphology of

the neurons in the human lateral geniculate nucleus and
their normal development. A Golgi study. Exp. Brain
Res., 47, 159-171.
Courten, c. de and Garey, L.J. (1983). Developpement
morphologique de la voie visuelle primaire chez 1 'en-
fant. J. Fr. Ophtalmol., ~. 187-202.
Dobson, v. and Teller, D.Y. (1978). Visual acuity in
human infants: a review and comparison of behavioral
and electrophysiological studies. Vision Res., 18,
1469-1483.
Fritschy, J.M. and Garey, L.J. (l.986a). Quantitative
changes in morphological parameters in the developing
visual cortex of the marmoset monkey. Devel. Brain
Res., 29, 173-188.
Fritschy, J.M. and Garey, L.J. (1986b). Postnatal deve-
lopment of quantitative morphological parameters in the
lateral geniculate nucleus of the marmoset monkey.
Devel. Brain Res., 30, 157-168.
Fritschy, J.M. and Garey, L.J. (1987). Postnatal deve-
lopment of dendrites in the lateral geniculate nucleus
of the marmoset monkey (Callithrix jacchus): a quan-
titative Golgi study. Submitted.
Garey, L.J. and de Courten, c. (1983). Structural deve-
lopment of the lateral geniculate nucleus and visual
cortex in monkey and man. Behav. Brain Res., 10, 3-13.
Garey, L.J. and Saini, K.D. (1981). Golgi studies of
the normal development of neurons in the lateral geni-
culate nucleus of the monkey. Exp. Brain Res., 44, 117-

128.
Gottlieb, M.D., Pasik, P. and Pasik, T. (1985). Early
postnatal development of the monkey visual system. I.
Growth of the lateral geniculate nucleus and striate
cortex. Devel. Brain Res., 17, 53-62.
Headen, M.P., Sloper, J.J., Hiorns, R.W. and Powell,

T.P.S. (1981). Cell sizes in the lateral geniculate nu-
cleus of normal infant and adult rhesus monkeys. Brain
Res., 229, 183-186.
Held, R. (1979). Development of visual resolution. Can.

J. Psychol., 22, 213-221.
Hendrickson, A.E., Hunt, S.P. and Wu J.Y. (1981). Im-

munocytochemical localization of glutamic acid decar-
boxylase in monkey striate cortex. Nature, 292, 605-
607.
Hendrickson, A. and Kupfer, c. (1976). The histogene-

sis of the fovea in the macaque monkey. Invest.
Ophthalmol., 15, 746-756.
Hendrickson, A.E., Wilson, J.R. and Ogren, M.P. (1978).

The neuroanatomical organization of pathways between
the dorsal lateral geniculate nucleus and visual cortex
in Old World and New World primates. J. Comp. Neural.,
182, 123-136.
Hendrickson, A.E. and Yuodelis, c. (1984). The morpho-

logical development of the human fovea. Ophthalmol. ,
91, 603-612.
Heumann, D. and Leuba, G. (1983). Neuronal death in the
development and aging of the cerebral cortex of the
mouse. Neuropathol. Appl. Neurobiol., ~, 297-311.
Hickey, T.L. (1977). Postnatal development of the human
lateral geniculate nucleus: relationship to a critical
period for the visual system. Science, 198, 836-838.
Hickey, T.L. (1981). The developing visual system.

Trends in Neurosci., ~, 41-44.
Hitchcock, P.F. and Hickey, T.L. (1980a). Ocular domi-

nance columns: evidence for their presence in humans.
Brain Res., 182, 176-179.
Hitchcock, P.F. and Hickey, T.L. (1980b). Prenatal

development of the human lateral geniculate nucleus. J.
182 L. GAREY
Comp. Neurol., 194, 395-411.

Hollenberg, M.J. and Spira, A.W. (1973). Human retinal
development: ultrastructure of the outer retina. Amer.
J. Anat., 137, 357-385.
Holstein, G.R., Pasik, T., Pasik, P. and Hamori, J.
(1985). Early postnatal development of the monkey
visual system. II. Elimination of retinogeniculate
synapses. Devel. Brain Res., 20, 15-31.
Horton, J.C. (1984). Cytochrome oxidase patches: a new

cytoarchitectonic feature of monkey visual cortex.
Phil. Trans. Roy. Soc. B, 304, 199-253.
Horton, J.C. and Hedley-Whyte, E.T. (1984). Mapping of

cytochrome oxidase patches and ocular dominance columns
in human visual cortex. Phil. Trans. Roy. Soc. B, 304,
255-272.
Huttenlocher, P.R., Courten, c. de, Garey, L.J. and Van

der Leos, H. (1982). Synaptogenesis in human visual
cortex - evidenc~ for synapse elimination during normal
development. Neurosci. Letters, 11, 247-252.
Leuba, G. and Garey, L.J. (1984). Development of den-

dritic patterns in the lateral geniculate nucleus of
monkey: a quantitative Golgi study. Devel. Brain Res.,
16, 285-299.
Leuba, G. and Garey, L.J. (1987). Evolution of neuronal

numerical density in the developing and aging human
visual cortex. Human Neurobiol., In Press.
Leuba, G., Kraftsik, R. and Garey, L.J. (1985). Deve-

lopment of dendrites in the human lateral geniculate
nucleus. Neurosci. Letters, Suppl. 22, 309.
Lund, J.S., Boothe, R.G. and Lund, R.D. (1977). Deve-
lopment of neurons in the visual cortex (area 17) of
the monkey (Macaca nemestrina): a Golgi study from
fetal day 127 to postnatal maturity. J. Camp. Neurol.,
176, 149-188.
Mann, I. (1964). The Development of the Human Eye.
British Medical Association, London.
Michel, A.E. and Garey, L.J. (1984). The development of
dendritic spines in the human visual cortex. Human
Neurobiol., ~' 223-227.
rhesus monkey retina. I. Emergence of the inner plexi-

form layer and its synapses. J. comp. Neurol. , 2 41,
420-434.
O'Kusky, J. and Colonnier, M. (1982). Postnatal changes

in the number of neurons and synapses in the visual
cortex (area 17) of the macaque monkey: a stereological
analysis in normal and monocularly deprived animals. J.
Comp. Neurol., 210, 291-306.
Pirchio, M., Spinelli, D., Fiorentini, A. and Maffei,

L. (1978). Infant contrast sensitivity evaluated by
evoked potentials. Brain Res., 141, 179-184.
Rakic, P., Bourgeois, J.P., Eckenhoff, M.F., Zecevic,

N. and Goldman-Rakic, P.S. (1986). Concurrent overpro-
duction of synapses in diverse regions of the primate
cerebral cortex. Science, 232, 232-234.
Rakic, P. and Riley, K.P. (1983). Overproduction and

elimination of retinal axons in the fetal rhesus mon-
key. Science, 219, 1441-1444.
Saini, K.D. and Garey, L.J. (1981). Morphology of neu-

rons in the lateral geniculate nucleus of the monkey. A
Golgi study. Exp. Brain Res., 42, 235-248.
samorajski, T., Keefe, J.R. and Ordy, J.M. (1965).

Morphogenesis of photoreceptor and retinal ultrastruc-
ture in a sub-human primate. Vision Res.,~, 639-648.
Sauer, B., Kammradt, G., Krauthausen, I., Kretschmann,

H.J., Lange, H.W. and Wingert, F. (1983). Qualitative
and quantitative development of the visual cortex in
man. J. Comp. Neurol., 214, 441-450.
Smelser, G.K., Ozanics, v., Rayborn, M. and Sagun, D.
(1974). Retinal synaptogenesis in the primate. Invest.
Ophthal1nol., 13, 340-361.
Spatz, W.B. (1979). The retino-geniculo-cortical path-

way in Callithrix. II. The geniculo-cortical projec-
tion. Exp. Brain Res., 2§, 401-410.
Spira, A.W. and Hollenberg, M.J. (1973). Human retinal
development: ultrastructure of the inner retinal
layers. Devel. Biol., 31, 1-21.
Takashima, s., Chan, F., Becker, L.E. and Armstrong

D.L. (1980). Morphology of the developing visual cortex
of the human infant. A quantitative and qualitative
Golgi study. J Neuropathol. Exp. Neurol. 39, 487-501.
184 L. GAREY
Teller D.Y. (1981). The development of visual acuity in

human and monkey infants. Trends in Neurosci., ~' 21-
24.
Vital-Durand, F., Garey, L.J. and Blakemore, C. (1978).
Monocular and binocular deprivation in the monkey:
morphological effects and reversibility. Brain Res. ,
158, 45-64.
Von Noorden, G.K. (1977). Mechanisms of amblyopia. Adv.
Von Noorden, G.K. (1981). New clinical aspects of

stimulus deprivation amblyopia. Amer. J. Ophthalmol.,
92, 416-421.
Wiesel, T.N. (1982). Postnatal development of the

visual cortex and the influence of environment. Nature,
299, 583-591.
Wiesel, T.N. and Hubel, D.H. (1974). Ordered arrange-

ment of orientation columns in monkeys lacking visual
experience. J. Comp. Neurol., 158, 307-318.
Wiesel, T.N., Hubel, D.H. and Lam, D.M.K. (1974). Au-

toradiographic demonstration of ocular-dominance co-
lumns in the monkey striate cortex by means of trans-
neuronal transport. Brain Res., 79, 273-279.
IS
EFFECTS OF ABNORMAL VISUAL EXPERIENCE ON THE
MORPHOLOGY OF LATERAL GENICULATE NEURONS IN
THE INFANT PRIMATE
M.P. HEADON, J.J. SLOPER and T.P.S. POWELL
INTRODUCriON
In infant monkeys early monocular lid suture causes shrinkage

of neurons in the deprived laminae of the lateral geniculate
nucleus (LGN) relative to undeprived cells in the same animal
(Headon and Powell 1973, Von Noorden 1973b, Hubel et al. 1977,
J.JaVay et al. 1980, Vital-Durand et al. 1978, Von Noorden and
Crawford 1978). Measurements have shown that major changes in size
of LGN cells occur if closure is started during the first two to
three months of life (Von Noorden 1973b, Vital-Durand et al. 1978,
Headon et al. 1985b), with a small degree of residual sensitivity
remaining for approximately the first year. In the cat it has been
shown that a competitive interaction between visual pathways
related to the two eyes is important in producing the abnormalities
seen after monocular closure (Wiesel and Hubel 1963 a,b, Guillery
and Stelzner 1970, Guillery 1972). A competitive interaction also
appears to be important in the early development of the primate
visual system (Von Noorden 1973a, Hubel et al. 1977), but a small
degree of cell shrinkage has been described in the monocular
segment of the LGN, where competition should not occur, and it has
therefore been suggested that disuse also plays a part in causing
the changes in LGN cell size (Von Noorden and Middleditch 1975,
Von Noorden et al. 1976).
The above conclusions regarding the morphological

sensitivity of LGN cells are based on comparisons made between
cells in deprived and undeprived LGN laminae within individual
animals, so that each animal serves as its own control. This method
has the advantage that variability in cell size measurements
between individual animals, whether for natural or technical
reasons, is controlled for because results are expressed as a
percentage difference in size between cells in corresponding
deprived and undeprived laminae within an animal. However
interpretation of these results has depended on the assumption that
cells in the undeprived laminae are unaffected by deprivation.
185
186 M.P. HEADON, J.J. SLOPER and T.P.S. POWELL
Although the possibility of changes in undeprived cells has been

considered by several authors (Wiesel and Hubel 1963a, Guillery
1972, Headon and Powell 1973, Hubel et al. 1977) and a degree of
hypertrophy of undeprived cells has been shown in the cat and dog
in addition to major shrinkage of deprived cells (Sherman and
Wilson 1975, Hickey et al 1977), previous results in the monkey
have always been interpreted as showing shrinkage or failure of
growth of deprived cells.
In order to interpret the effects of binocular closure in

infant monkeys (Headon and Powell 1978) it became necessary to
measure LGN cell sizes in a series of normal monkeys. From these
results it was apparent that there was much less variability in
normal cell size between different monkeys than had been previously
thought and much of the variability of undeprived cell size in
experimental animals is in fact due to changes in these cells
resulting from the visual deprivation. (Headon et al. 1985a). These
cl>anges in undeprived cells have now been systematically examined
by comparing sizes of both deprived and undeprived LGN cells in
visually deprived monkeys with measurements of normal LGN cells in
18 monkeys ranging in age from 8 days to fully adult. This has
shown that undeprived cells in experimental animals undergo
surprising and extensive changes in size following visual
deprivation and these may mask changes in the deprived cells when
only the size changes relative to undeprived cells are measured
(Headon et al. 1985b). Cells in the LGN are sensitive to visual
deprivation for much longer than previously thought, but the
changes seen in LGN cells depend critically on timing and may be
qualitatively different both according to the duration of the
deprivation and the age at which deprivation is started. There is a
second period of sensitivity extending from about two months of age
to one year during which it appears that cooperation between the
visual pathways related to the two eyes is necessary for normal
development to occur. It was also apparent from these studies that
there are important differences in the response to deprivation
between cat and monkey.
THE NORMAL .LATERAL GENICULATE NUCLEUS
The LGN of the rhesus monkey, like that of man, consists

essentially of six laminae of neurons. Of these the inner two
(Laminae I and II) normally consist of large cells and are termed
the magnocellular laminae and the outer four (Laminae III to VI)
contain smaller cells and are termed the parvocellular laminae.
These two sets of cells react differently to visual deprivation
under certain conditions. There were no systematic differences in
size between left and right sides or between male and female
animals. Growth of cells in the parvocellular LGN laminae is
complete by one week of age, there being no change in mean size
between then and adulthood (Fig. 1 ). Cells in the magnocellular
laminae grow by about 1 0% over this same perioo. but most of their
EFFECTS OF VISUAL DEPRIVATION ON LGN CELL SIZE 187
•
I
•
300 • •• •
• • ~~
• •
•
NE •
:J... • •
0
Cl!
....
0
200
•• • •
•
•
•
•
•
•
• .
~
•
Q;
u •
c
0
Cl! •- Magnocellular
~
100 •- Parvocellular :!::
:::J
"0
<{
os~~~----~~~~~~----~-L-L~~~----
10 50 100 500
Age(days)
Fig 1. Mean LGN cell area for 18 normal monkeys plottoo against
age. Cells in the parvocellular laminae do not grow between a week
of age and adulthood. From Headen et al. 1985a.
growth is also completoo early. In this the monkey differs from man
because growth in the human LGN continues until 6 months of age in
the parvocellular laminae and about two years of age in the
magnocellular laminae (Hickey 1977). There is also evidence from
the development of the detailoo structure of LGN cells (Garey and
Saini 1981, De Courten and Garey 1982) and the timing of the
appearance of cytochrome oxidase patches in the visual cortex
(Horton 1984, Horton and Hooley-White 1984), that the infant monkey
is considerably more mature at birth than the human, although the
basic pattern of cell growth appears to be similar (Headon et al.
1985a). In the cat the sensitive period corresponds to a period of
rapid LGN cell growth (Garey et al. 1973) and it has been suggestoo
that this may be a general principle (Hickey 1 977), but it is
clearly not so in the monkey (Headon et al 1985a).
EFFECI'S OF M::JN<Xm.AR CWSURE AT BIRTH
In the monkey the major initial change following monocular

closure at birth is not shrinkage of deprived LGN cells but
hypertrophy of cells in the undeprivoo parvocellular laminae
g •- Undeprived
o- Deprived
"- Approx age at dosurP E " Approx.oge at closure
'/,
"
-10
or---------------~~~
t
£
g
E
"
-30
0 ~--------------~----v~
l
~
-10
-20
g
t
E
"
-3()
0
£ -10
g
"t£
~ -20
E
~
-30
l
-10
-20
-30 t
l
0 £
-10
-20
-30
=;~L
0o 100 200~---soo 600
<'---//---
~ 1cX:x:t'1 aOO --i~

l
- 10
-20
0 o!;--;;:100~2'*'oo;--;;3()0~-;;400:;;-;5o'oooc:--:o~oo 900 1000 ~-,~oo
Age (days) Age , days)
Fig 2. Changes in LGN cell area in a series of monkeys monocularly

deprived from birth for varying periods. The upper graph shows the
shrinkage of deprived cells relative to undeprived cells in the
same animal. The lower graph shows mean areas for both deprived and
undeprived cells from the same animals plotted in relation to cell
areas in normal animals. (Headon et al 1985b).
(Fig. 2, bottom). A difference between deprived and undeprived

cells is detectable as early as two days after closure and by three
to four weeks undeprived cells are about 25 to 30% larger than
normal. This hypertrophy is responsible for the difference in size
seen between deprived and undeprived parvocellular cells because,
surprisingly, there is little immediate change in the deprived
cells. Less hypertrophy occurs in the magnocellular laminae. The
hypertrophy is maintained until about eight weeks of age, but then
the undeprived parvocellular cells shrink back to normal size and
the deprived parvocellular cells shrink in parallel (Fig. 2,
bottom). By three months of age the undeprived cells are of normal
size and the deprived cells are very shrunken. This later stage of
parallel shrinkage does not affect the magnocellular cells.
The changes in size of LGN cells following monocular closure

at birth thus occur in two distinct phases. Initially the
difference between deprived and undeprived cells is produced mainly
by hypertropy of the undeprived cells. This is followed by a second
phase during which deprived and undeprived cells shrink in parallel
but only cells in the parvocellular laminae are involved. This

later phase has not been described previously because it is not
apparent if deprived cells are only compared to undeprived cells
(Fig. 2, top) since the two sets of cells are changing in parallel.
EFFECI'S OF IDOOCULAR cr.DSURE STARTED AT LATER AGES.
A period of normal visual experience prior to deprivation does

not simply reduce the sensitivity of LGN cells to closure but
rather causes a change in the pattern of reaction of LGN cells as
closure is started at progressively later ages. If deprived cells
are only compared to undeprived cells there is close agreement
between the present studies (Headon et al 1985b) and previous work
(Headon and Powell 1973, Von Noorden 1973b, Von Noorden and
Crawford 1978, Vital-Durand et al. 1978, LeVay et al. 1980).
The difference between deprived and undeprived cells is markedly
less when closure is started at about two months of age or later,
although a small effect is still detectable even with closure
started at eighteen months of age (Fig 3, left). This marked
reduction in sensitivity at about two months of age has been
thought to mark the errl of the most sensitive period of development
and corresponds to the time when the width of cortical ocular
dominance is largely fixed.
However, comparison of parvocellular cell sizes in
experimental animals with those of normal animals shows that major
sensitivity to closure does not end at this time. In fact the
reaction of the deprived parvocellular cells to closure is
unaltered over this period (Fig. 3, right) but changes in them
become masked because there is a marked change in the reaction of
the undeprived cells. Whereas the undeprived cells underwent
hypertrophy following closure at birth they shrink in response to
later closure. This gives the appearance of a reduction in
sensitivity i f comparisons are only made between deprived and
undeprived cells when what is in fact changing is the type of
sensitivity of the undeprived parvocellular cells. Shrinkage of the
undeprived cells following monocular closure is marked for closures
started at up to nine months of age and sensitivity then tails off
at a year to eighteen months of age. It only affects cells in the
parvocellular layers, mean cell areas in the magnocellular laminae
being unchanged.
Normal visual development prior to closure thus causes a
marked change in the effect of closure on LGN cells as development
proceeds, there being a reversal in the reaction of the undeprived
parvovellular cells so that monocular closure started at between
two and twelve months of age causes shrinkage of both deprived and
undeprived parvocellular cells. At the peak of late sensitivity
these changes in cell size are as marked, and occur as rapidly, as
changes in the early sensitive period following closure at birth,
but the pattern of change is different and only parvocellular cells
and not magnocellular cells are affected.
2!
c Magnocellular
~ Parvocellular
~ .,,
i 0
''
£
~
-10 •
:e
~
-20 '~·
&" ...........,
- - #---
~c -30
~,.____._
~ -40
-50
•- - - - - • Undeprived magnocellular
'1.
Deprived magnocellular
- Undeprived parvocellular
2!
c
Deprived parvocellular
~
__ ..-----#'---- _ ..
~ - .#--
i
£
~
:e
~ -20
' 'c: '
.,.o-- - -- -o- - - - - #- - - - o, .... '
b--- -<>....
~c -30
~ -40
-soo~--~20~---4~0~--~67
o----~80~---1~oo~~--4-00L----4~2-o-*~18~4-0--~1860
Durat 1on (days)
Fig 3. Left: Changes in deprived LGN cell area relative to

undeprived cells from the same animal in a series of monkeys with
monocular closure started at progressively later ages. For each
series of animals the age at which closure was started is plotted
as a cross. Little change is seen with closure started at 60 days
or later.
Right: Changes in both deprived and undeprived cells in the same
animals are plotted in relation to normal cell sizes. At later ages
shrinkage of the undeprived cells can be seen to mask the changes
in the deprived cells. From Headon et al. 1985b.
During the first six to eight weeks of life it appears that

the balance between the input from the two eyes is being determined
by a competitive mechanism and that if the activity of one eye is
impaired then the other eye becomes dominant. This is reflected not
only in changes in LGN cell size but also in expansion and
contraction of cortical ocular dominance stripes and in the
physiological dominance of cortical neuronal responses by the open
eye (LeVay et al. 1980). A small degree of this effect is still
apparent until at least a year of age in the LGN but the major
effect of late deprivation started at this time, namely the
simultaneous shrinkage of both deprived and undeprived
parvocellular cells, cannot be explained in this way. Because
closure of one eye causes shrinkage of LGN cells related to both
eyes it has been suggested that the changes are a reflection of
impaired binocular function, probably initially at cortical level
and possibly involving connections from layer IV to supragranular
layers and other intracortical connections (Headon et al. 1985b).
Although this later period has received much less attention,
changes have been described in the physiological responses of
neurons in the supra and infragranular layers where binocularly
driven neurons are first found (LeVay et al.1980) and there is
evidence for histological changes in the cortex following late
closure from both the Liesgang silver stain (LeVay et al. 1 980) and
the cytochrome oxidase method (Horton 1984). In view of the changes
in both deprived and deprived LGN cells it is of interest that the
cytochrome oxidase changes involve a reduction in staining of the
border regions of both deprived and undeprived ocular dominance
stripes in layer IV. Although it is not clear how the late cortical
and geniculate changes are interrelated, it seems that the pathways
related to the two eyes are in some way dependant on each other for
normal development during this late period and this probably
involves the elaboration or maturation of cortical binocular
function. In the absence of normal binocular cooperation there is
shrinkage of LGN cells related to both eyes, regardless of whether
the corresponding eye is directly deprived of visual input or not.
In other words a normal monocular input is not sufficient for the
maintenance of normal LGN cell sizes. This concept has been tested
by making measurements in the monocular segment of normal monkeys.
CHANGES OF CELL SIZE IN THE MJNOCULliR SEG1ENT
The lack of change in cell size in the monocular segment of

the kitten LGN is an important part of the evidence for the
involvement of a competitive mechanism in producing the shrinkage
of cells in the binocular part of the nucleus (Guillery and
Stelzner 1970). In the monkey there is much less cell shrinkage in
the monocular segment than the binocular part of the nucleus, which
similarly indicates the involvement of a competitive mechanism in
mediating the effects of early monocular closure. However in the
monkey there is a small degree of change in this region (c.1 0% c.f.
c.30% in binocular regions) and it has therefore been suggested
that disuse may also play a part in causing the cell size changes
(Von Noorden and Middleditch 1975, Von Noorden et al. 1976).
Although confirming the smaller difference in size between deprived
and urrleprived monocular segment cells, comparisons with cells
sizes in the normal monocular segment have shown this difference to
be due to a small degree of hypertrophy of cells in the urrleprived
monocular segment (Sloper et al. 1987). Because there is no direct
influence from the deprived eye in this region this cannot be due
to disuse but probably indicates a weak indirect or diffuse
binocular effect influencing cells in the monocular segment.
The above changes occur following early closure. It has been

suggested that during the later, secorrl phase of development a
ccoperative interaction between inputs from the two eyes is
important in maintaining normal cell sizes in the parvocellular
laminae and that it is the absence of this which causes the
shrinkage of both deprived and urrleprived cells following late
closure (Headon et al. 1985b). Such ccoperation should be greatly
reduced or absent in the monocular segment of the nucleus during
normal development. It was therefore predicted that cells in this
region of the LGN would shrink during the later stage of normal
development. Measurements in a series of normal monkeys have shown
this to be true, cells in the parvocellular monocular segment
shrinking by about 1 5% during normal development between 3 months
of age and adulthood. This shrinkage occurs both in absolute terms
and in relation to cells in the binocular parts of the same
nucleus. Magnocellular cells grow slightly over this same period as
do cells in the binocular parts of the parvocellular laminae
(Sloper et al. 1987).
ffiiTICAL PERIODS
The studies of the development of LGN cells described above

have shown that the development of the primate visual system has
two distinct critical periods during which different aspects of
visual experience are important for normal development. Between
birth and about 8 weeks the balance between the two eyes is largely
determined by a competitive mechanism. If one eye is closed at
birth parvocellular LGN cells related to the open eye enlarge and
there is initially little shrinkage of deprived cells. After about
8 weeks of age binocular ccoperation between inputs from the two
eyes is most important for normal development. If an animal enters
this second phase with a preexisting a.lnormality in the balance
between the parts of the visual system related to the two eyes as a
result of early closure then all parvocellular cells shrink during
this secorrl phase, although the size difference between deprived
and urrleprived cells is unchanged. The end result is a return to
normal size for the undeprived cells, with the deprived cells
becoming markedly smaller than normal. However the normal size of
the urrleprived cells is probably the result of two different
a.lnormal processes whose effects happen to cancel each other rather
than a reversal of the initial abnormality. If closure is started

between about 8 weeks and one year of age there is similarly
shrinkage of both deprived and undeprived parvocellular cells but
little difference in size between them and all parvocellular cells
become smaller than normal. For reasons outlined above it is
probably the failure of binocular cooperation at cortical level
which is the cause of the parvocellular cell shrinkage in both
these instances.
CLINICAL IMPLICATIONS
Both the morphological and behavioral studies suggest that

binocular function is sensitive to disruption later that the other
aspects of visual development studied and this has important
implications for the understanding and treatment of visual problems
occuring during development. In particular, abnormalities occuring
early not only disrupt development at that time but also prevent
normal development during the subsequent phase. Although the
initial abnormality may simply be that the balance between the two
eyes has been upset, this is soon compounded by the failure of
development of binocular cooperation.
It is difficult to give an accurate comparison of ages

between monkey and man but a ratio of 3 or 4 to 1 is usually used.
In addition, evidence that the monkey is more mature at birth than
man has been presented above. Taking both these factors into
consideration it seems likely that the marked drop in sensitivity
for altering the balance between the two eyes occuring at 2 to 3
months of age in the monkey would correspond to about 1 year to 18
months of age in man, although, as in the monkey, there would still
be a degree of sensitivity to this parameter for considerably
longer. As judged by changes in LGN cells, sensitivity to binocular
disruption would be relatively high up to about 3 years of age in
man with a decline over the following 3 or 4 years. This is in
reasonable agreement with studies of disruption of the development
of binocular function in man (Banks et al. 1975, Hohmann and
Creutzfelt 1975), although there is also evidence for the initial
development of stereopsis at earlier ages (Held et al. 1980).
The development of binocular function during the late critical

period has not received as much attention in animal experiments as
has the early critical period and yet many patients are first seen
during this later period. There is no evidence at present as to how
these two aspects of visual development may interact, but it may
well be that the abnormal binocular development impedes the
correction of early abnormalities in the balance between the eyes.
If the balance between the two eyes is restored before the second
critical period then it is possible that binocular development will
proceed normally. However great caution is needed because of the
ease with which amblyopia can be induced by covering the undeprived
eye at these early ages and particularly in view of the report of
bilateral amblyopia following early reverse suture in kittens

(Murphy and Mitchell 1986) and its ocurrence in one of a similarly
treatoo group of infant monkeys (Sloper et al. unpublishoo).
Finally, the failure of normal binocular development may be

important not only in regard to visual perception but also in those
aspects of eye movement control in which binocular function is
involvoo. This may be important itself in the development of squint
and equally a squint may in turn lead to further disruption of
binocular visual function.
REFERENCES
BANKS,M.S., ASLIN,R.N. and LETSON,R.D. (1975). Sensitive period for

development of human binocular vision. Science 190: 675-677.
De muRTEN,C. and GAREY,L.J., (1982). Morphology of neurons in the
human lateral geniculate nucleus and their normal development.
Exp.Br.Res. 47: 159-171.
GAREY,L.J.,FISKEN,R.A. and POWELL,T.P.S.,(1973). Observations on
the growth of cells in the lateral geniculate nucleus of the cat.
Brain Res. 52: 359-362.
GAREY,L.J. and SAINI,K.D., (1981 ). Golgi studies of the normal
development of neurons in the lateral geniculate nucleus of the
monkey. Exp. Brain Res. 44:117-128.
GUILLERY,R.W., (1972). Binocular competition in the control of
geniculate cell growth. J.comp.Neurol. 144: 117-130.
GUILLERY,R.W. and STELZNER, D.J., (1970). The differential effects
of unilateral lid closure upon the monocular and binocular segments
of the dorsal lateral geniculate nucleus in the cat. J.comp.Neurol.
139: 413-422.
HEADON,M.P. and POWELL, T.P.S., (1973). Cellular changes in the
lateral geniculate nucleus of infant monkeys after suture of the
eyelids. J.l\Ilat. 116: 135-145.
HEADON,M.P. and POWELL, T.P.S., (1978). The effect of bilateral eye
closure upon the lateral geniculate nucleus in infant monkeys.
Brain Res. 143: 147-154.
HEADON,M.P., SLOPER,J.J., HIORNS,R.W. and POWELL, T.P.S., (1985a).
Sizes of neurons in the the primate lateral geniculate nucleus
during normal development. Dev. Brain Res. 18: 51-56.
HEADON,M.P., SLOPER,J.J., HIORNS,R.W. and POWELL, T.P.S., (1985b).
Effects of monocular closure at different ages on deprivoo and
undeprivoo cells in the·primate lateral geniculate nucleus. Dev.
Brain Res. 18: 57-78.
HELD,R., BIROf,E. and GWIAZDA,J. (1980). Stereoacuity of human
infants. Proc.Natl.Acad.Sci USA 77: 5572-5574.
HICKEY,T.L.,(1977). Postnatal development of the human lateral
geniculate nucleus: Relationship to a critical period for the
visual system. Science 198: 836-838.
HICKEY,T.L., SPEAR,P.D. and KR.ATZ,K.E.,(1977). Quantitative studies
of cell size in the eat's dorsal lateral geniculate nucleus
following visual deprivation. J.comp.Neurol. 172: 265-282.
OOHMANN,A. and OillJTZFELDI',O.D. (1975). Squint and development of

binocularity in humans. Nature 254: 613-614.
OOR'IOO,J.C. (1984). Cytochrome oxidase patches: a new
cytoarchitectonic feature of monkey visual cortex. Phil.Trans.
Roy.Soc.~ 304: 199-254.
HOR'IOO,J.C. and HEDLEY-WHITE (1984). Mapping of cytochrome oxidase
patches and ocular dominance columns in human visual cortex.
Phil.Trans.Roy.Soc.B 304: 255-272.
HUBEL,D.H., WIFSEL,T.N. and LEVAY ,S. (1977). Plasticity of ocular
dominance columns in monkey striate cortex. Phil.Trans.R.Soc.B 278:
377-409. -
LEVAY,S., WIESEL,T.N. and HUBEL,D.H., (1980). The development of
ocular dominance columns in normal and visually deprived monkeys.
J.comp.Neurol.191: 1-51.
MURPHY,K.M. and MITCHELL D.E., (1986). Bilateral amblyopia after a
short period of reverse occlusion in kittens. Nature 323: 536-538.
SHERMAN,S.M. and WILSON,J.R.,(1975). Behavioral and morphological
evidence for binocular competition in the postnatal development of
the dog's visual system. J.comp.Neurol. 161: 183-196.
SLOPER,J.J., HEAOON,M.P. and POWELL,T.P.S. (1987). Changes in the
size of cells in the monocular segment of the primate lateral
geniculate nucleus during normal development and following visual
deprivation. Dev. Brain Res. 31: 267-276.
VITAL-DURAND,F., GAREY,L...J. and BLAKEMORE,C., (1978). Monocular and
binocular deprivation in the monkey: Morphological effects and
reversibility. Brain Res. 158: 45-64.
VON N<X>RDEN,G.K., (1973a). Experimental amblyopia in monkeys.
Further behavioral observations and clinical correlations. Invest.
Ophthal. 12: 721-726.
VON N<X>RDEN",G.K., (1973b). Histological studies of the visual
system in monkeys with experimental amblyopia.
Invest. Ophthal. 12: 727-738.
VON NOORDEN,G.K. and CRAWFORD,M.I...J., (1978). Morphological and
physiological changes in the monkey visual system after short term
lid suture. Invest. Ophthal. 17: 762-768.
VON NOORDEN,G.K., CRA~'lFORD,M.L.J.and MIDDLEDITCH,P.R., (1976). The
effects of monocular visual deprivation: disuse or binocular
interaction. Brain Res. 111: 277-285.
VON NOORDEN,G.K. and MIDDLEDITCH,P.R., (1975). Histology of monkey
lateral geniculate nucleus after unilateral lid closure and
experimental strabismus: further observations. Invest. Ophth. 14:
674-683.
WIFSEL,T.N. and HUBEL,D.H., (1963a). Effects of visual deprivation
on morphology and physiology of cells in the eat's lateral
geniculate body. J.Neurophysiol. 26: 978-993.
WIESEL,T.N. and HUBEL,D.H., (1963b). Single-cell responses in
striate cortex of kittens deprived of vision in one eye.
J.Neurophysiol. 26: 1003-101 7.
16
THE INFLUENCE OF THE PERIOD OF DEPRIVATION ON
EXPERIMENTAL REFRACTIVE ERRORS
EARLL. SMITH III, RONALD S. HARWERTH, M.L.J. CRAWFORD

and GUNTER K. von NOORDEN
INTRODUCTION
The eye 1 s optical and axial components normally grow in a
regulated manner so that the eye maintains an approximately
emmetropic refractive condition throughout development. But, if the
potential for a clear retinal image is prevented during development
by an ocular abnormality (or by experimental manipulation in
laboratory animals), the coordinated growth of the eye is disrupted
resulting in an anomalous refractive error (Rabin et al., 1981; von
Noorden and Lewis, 1987; also see Criswell and Goss, 1983; and
Yinon, 1984, for recent reviews). Since these resulting refractive
errors can aggravate existing amblyopiogenic factors, optimal
treatment and management procedures for amblyopia require an
understanding of the vision-dependent mechanisms that influence the
emmetropization process.
Form deprivation produced by lid suture, the most common
manipulation employed to degrade the retinal image in experimental
animals, has consistently been shown to disrupt the normal
emmetropization process in young monkeys. There is, however, a
substantial amount of variability in the type and magnitude of the
refractive error produced by form deprivation in young monkeys
(Raviola and Wiesel, 1985; von Noorden and Crawford, 1978; Smith et
al., 1987). Since form-deprived monkeys potentially could serve as
useful animal models in investigations of the vision-dependent
mechanisms that influence the emmetropization process, it is
important to determine the primary reasons for the intersubject
variability in the effects of form deprivation on the monkey 1 s
refractive status. Identifying the experimental and subject factors
responsible for this variability should also provide insight into
the mechanisms regulating emmetropization. In the present report,
we have analyzed existing data on form-deprived monkeys in order to
*Supported in part by grants EY 03611, EY 01120, and EY 01139 from
the National Eye Institute.
197
198 E.L. SMITH III et al.
examine the influence of the period of deprivation (i.e., the age

at the onset and duration of deprivation) on experimentally induced
refractive errors.
EFFECTS OF FORM DEPRIVATION ON THE MONKEY'S REFRACTIVE STATUS

Type of Induced Refractive Error
In their initial publication on the effects of anomalous visual
experience on the refractive status of young monkeys, Wiesel and
Raviola ( 1977) reported that form deprivation initiated early in
life caused deprived eyes to become axially myopic. In virtually
all subsequent investigations, myopia of varying magnitude has been
observed in the majority of form-deprived monkeys. There have,
however, been a significant number of exceptions to this general
rule; several laboratories have found that some form-deprived
monkeys exhibit hyperopic refractive errors (von Noorden and
Crawford, 1978; Smith et al., 1987).
The variability in the type and magnitude of refractive errors
demonstrated by form-deprived monkeys is illustrated in Figure 1
which compares the refractive-error frequency distributions for
normal and form-deprived monkeys. The distribution of refractive
errors for the population of form-deprived monkeys (Figure 1B),
which includes data for monocularly deprived, binocularly deprived
and reverse-deprived monkeys from five laboratories, deviates
substantially from that for a population of normal monkeys (Figure
1A; both free-ranging and laboratory-reared monkeys are included).
Like humans, the great majority of normally reared monkeys exhibit
either no refractive error or a small degree of hyperopia. In
contrast, the population of form-deprived monkeys exhibits a higher
prevalence of large refractive errors. There is a particularly high
number of subjects with relatively large myopic refractive errors.
And although a minority of deprived monkeys have hyperopic
refractive errors, the magnitudes of these hyperopic errors are
also generally larger than those observed in normal monkeys.
The large range of refractive errors observed in the population
of form-deprived monkeys can be attributed to several factors. For
instance, the higher prevalence of moderate hyperopic refractive
errors in the refractive-error distribution for the population of
form-deprived monkeys can be attributed, in part, to the fact that
some of these animals were very young at the time that their
refractive errors were assessed. Many of the potentially
confounding variables associated with this population can be
minimized, however, by restricting the analysis of the effects of
form deprivation only to monkeys that were subjected to monocular
form deprivation. In these animals 1 the nondepri ved eye's
refractive error can be. used as a reference and 1 thus, assuming
that the nondeprived eye's refractive status has not been
influenced by form-depriving its fellow eye, interocular
EXPERIMENTAL MYOPIA 199
DISTRIBUTlON OF REFRACnVE ERRORS
A . NORMAL MONKEY EYES B. DEPAIYED MCN(EY EYES N . 74
, o-............
A
A ,
.............""""'
A
SPHERICAL EQUrvALENT REFRACnYE ERROR (DIOPTER$)
Figure 1. Refractive-error frequency distributions for the eyes of

normally reared monkeys (panel A; from Young, 1964) and for the
deprived eyes of lid-sutured monkeys (panel B). The following
letter codes are used in panel B to indicate the sources of the
data: A = von Noorden and Crawford ( 1977); B = Harwerth et al.
(1983); C = Raviola and Wiesel (1985); D =Smith et al. (1987); E =
Green and Guyton (1986); F = Smith et al. (unpublished); G =Wiesel
and Raviola (1977). The treatment paradigm for individual animals
is indicated by the appropriate shading.
comparisons of refractive-error differences between the deprived

and nondeprived eyes should minimize a number of important
interlabortory and intersubject variables (eg., genetic factors,
the housing environment, age, etc.).
In Figure 2, the interocular differences in refractive error
are shown for individual monkeys that were monocularly form
deprived before two years of age. By specifying the treated eye's
refractive status with respect to that for the nondeprived eye, the
type of refractive-error alteration produced by form deprivation is
consistent between subjects. In 24 of the 26 monkeys that were
monocularly lid sutured early in life, the deprived eyes were more
myopic or less hyperopic than the nondeprived eyes.
Influence of the Period of Form Deprivation
Although Figure 2 illustrates that monocular form deprivation
almost always causes the deprived eye to become relatively more
myopic than the nondeprived eye, it also demonstrates that there is
a high degree of intersubject variability in the magnitude of the
induced myopia. This variability has been considered to be due to
differences in the lid-fusion procedures employed by different
INTEROCULAR REFRACTIVE ERROR DIFFERENCES
·-·· 1
- 10
r-1
u
0
W€SE.l ..,~AAio'IOLA
(19711
.,~~.=~~~.~:,o
l.::1::
UJ G HARW'E.ATH lit ... f 1N.l)
u 0
z
UJ
-80 SMITH~.._ (111187)
cr
~ ~
~ ~ -60
cr
~ ii -40
i
UJ •
~ -~0
i= B ·.·.
u
<(
cr
lL
UJ
a: •2 0
0 5 10 15 20 2.5 30
MONOCULARLY DEPRIVED MONKEYS
Figure 2. Interocular refractive-error differences for monocularly

lid-sutured monkeys that were form deprived before two years of
age.
laboratories and the resulting inconsistencies in the nature of the

anomalous visual experience (von Noorden and Crawford, 1978; Yinon,
1984; Raviola and Wiesel, 1985). Several observations, however,
indicate that the intersubject variability may be attributed, in
part, to differences in the age at the onset of deprivation and the
duration of the period of deprivation. First, there appears to be a
"sensitive period" within which form deprivation can alter the
eye's refractive status. Form deprivation during early development
causes the eye to become relatively myopic, but similar durations
of form deprivation in adult monkeys fail to produce changes in the
deprived eye's refractive status (Wiesel and Raviola, 1977; von
Noorden and Crawford, 1978). Since the susceptibilities of other
parts of the visual system to the deleterious effects of form
deprivation decline during the sensitive period, it seems
reasonable to expect that the magnitude of the induced refractive
error should also decrease as the age at the onset of deprivation
is delayed. Nevertheless, previous investigations have failed to
find a clear relationship between the age at the onset of
deprivation and the magnitude of the induced myopia (Sommers et
al. , 1978). Second, Raviola and Wiesel ( 1985) have demonstrated
that when form deprivation was discontinued in a young monkey, the
progression of relative myopia in the deprived eye was halted and
the formerly deprived eye assumed a normal growth rate. Moreover,
in a recent study primarily involving binocularly deprived monkeys,
Greene and Guyton ( 1986) followed the changes in their monkeys'
refractive errors during the period of form deprivation and found
that there were progressive increases in the degree of myopia (or

decreases in the degree of hyperopia) as a function of time. These
results suggest that the magnitude of the induced myopia should
also vary as a function of the duration of the period of
deprivation.
The relationship between the age at the onset of deprivation
and the magnitude of the deprivation-induced refractive error is
illustrated in Figure 3. Interocular differences in refractive
errors are plotted as a function of the monkey's age at the onset
of monocular form deprivation. If only the subjects that developed
a relative myopia in their deprived eyes are considered (i.e., the
two subjects plotted below the dashed line are excluded from the
analysis), there is a significant relationship between the age at
onset and the magnitude of the relative myopia (linear regression
analysis; r=0.65; df=19; p<0.01). The earlier the form deprivation
is initiated, the larger the relative myopia. Previous studies
(eg., Sommers et al., 1978) apparently did not identify the age at
the onset of deprivation as an important variable because of the
small sample sizes and limited range of ages that are typically
employed in individual studies involving monkeys.
The data in Figure 3 can also be used to estimate the end of
the sensitive period for environmentally induced refractive errors
in monkeys. If it is assumed that an interocular refractive-error
difference of less than one diopter is insignificant, then by
extrapolation, the sensitive period for refractive errors would end

MONOCULARLY FORM-DEPRIVED MONKEYS
DURATION OF MD"" 2 MONTHS

-14.0
.6. WIESEL AND RAVIOLA (1977)
"'w
u -12.0
•
•
VON NOOROEN & CRAWFORD (1978)
HARWERTH et al (1963)
z e SMITH et ul
w
a:
~ ~ -10.0
~~ ~ -8.0
• •
....
owu
~~0
w~~
•
..
we~
> a: -4.0
.
t=
u
...
a:
t:
-2.0
w
a: 0 - - - - - - - - - - - - - - - - - - - - - - _ _ _ _. _ _ _
•2.0 •
•
30 100 300 1000
AGE AT ONSET (DAYS)

lid-sutured monkeys plotted as a function of the age at the onset
of form deprivation. Only data for monkeys that were deprived for a
duration of 2 months or longer have been included.
202 E.L. SMITH III eta!.
by about 7 years of age. Young ( 1961 ) has reported that it is

possible to produce small refractive-error alterations (about 0.75
D) in four- to six-year-old monkeys by restricting the animals 1
visual space. Although the sensitive period for experimental
refractive errors appears to be relatively long, experimental
manipulations initiated beyond about 2 to 3 years of age would have
a relatively small effect on a monkey 1 s final refractive status.
The one animal from the study by von Noorden and Crawford (1978)
that was lid sutured at 24 months of age, for a period of 30
months, illustrates the low degree of plasticity in two-year-old
animals. This monkey did not develop a significant myopia in its
deprived eye. These results suggest that the sensitive period for
the effects of form deprivation has effectively ended by 2 years of
age. The work of Greene and Guyton (1986) also provides an
indication of the the most sensitive period for experimentally
induced refractive errors. Their experiments indicate that about
1• 9 years are required to achieve approximately 98% of the total
refractive-error changes produced by neonatal lid suture in
monkeys. If it is assumed that the sensitive period for
experimental myopia coincides with the period of time required to
achieve the maximum refractive-error effect, then their data also
suggests that the sensitive period for deprivation-induced
refractive errors extends until about 2 years of age.
Tae influence of the duration of monocular form deprivation on
the magnitude of the induced myopia is shown in Figure 4. To avoid

MONOCULARLY FORM-DEPRIVED MONKEYS
Ci)
a:
,_w
0.
• VON NO ORDEN & CRAWFORD (1977)
.WIESEL a RAVIOLA (1917)

•
0 • HARWERTH •t aL (1983)
§
"'0wz ::i,_0a: •
.~
w z
a:
w 0
• •
,_
Ci w
•
l--__·_-_ _._ _-----------------:--1

a:
0
a:
a:
"'w
a: • •
t:
w
w
>
;:
• •
."'
0
a:
w
a:
+4.0 I I I
DURATION OF DEPRIVATION (DAYS)

lid-sutured monkeys plotted as a function of the duration of form
deprivation. Only data for monkeys that were deprived before 6
weeks of age are included.
the confounding effects associated with differences in the ages at

the onset of deprivation, only animals that were lid sutured before
6 weeks of age have been included. Although the resulting sample
size is relatively small, there is an indication that the magnitude
of the relative myopia increases as a function of the duration of
deprivation. If the two monkeys that manifested a relative
hyperopia in their deprived eyes are excluded from the analysis,
there is a significant relationship between the duration of
deprivation and the degree of relative myopia (r=-0.65; df=10;
p<O. 05). However, again, it nrust be emphasized that this
relationship only holds for animals that developed a relative
myopia in their deprived eyes.
Since both the age at onset and the duration of the period of
deprivation appear to influence the magnitude of the induced
relative myopia, we performed a nrul tiple regression analysis to
determine the extent to which the combination of these variables
could account for the range of refractive-error alterations
observed in monocularly form-deprived monkeys. A total of 23
animals from three different laboratories were included in the
A AXIAl LENGTH vt. REFRACTIVE ERROR
•
B
·f .o-· ..... · i
INTERQC.UlA" AlCIAL UIIIG1H Ra,1iOS
....
•
•
...........
• WIII!KL AIOID ~WIOU 11• 11)
""'"'' ~"'
10 1,02 Hl4 1.01 lOa 1.10 1.12 1.1 4
AXIAL LENGTH' RAT10S fOEPAIVEDs CONTROL,

1. 1. !.I a 120
. .
I.K)I'fOCI.A.AFIL. V DEPRI\IEO I.AONK£"1"$
Figure 5. A. Interocular axial-length ratios for monkeys that were

monocularly form deprived before two years of age. B. Interocular
axial-length ratios plotted as a function of the interocular
refractive-error differences for individual monocularly
form-deprived monkeys. The solid line was derived from a
least-squares linear regression analysis of the data. The broken
line represents the predicted relationship between the refractive
error differences and the axial-length ratios based on the
following assumptions: the nondeprived eye had a 19 mm axial length
and was ermnetropic (the optics of the eye were represented by a
single spherical refracting surface 1.32 mm behind the cornea) and
that the alterations in the deprived eye were restricted to the
depth of the vitreous chamber.
analysis. Only monkeys that were lid sutured before 2 years of

age (i.e., within the most sensitive portion of the critical
period) and only monkeys that developed a relative myopia in
their deprived eyes were considered. The duration of deprivation
was calculated with respect to the end of the most sensitive
portion of the critical period. Specifically, it was assumed that
the sensitive period ended at two years of age and that periods
of deprivation extending beyond the end of the sensitive period
were not included as part of the duration of deprivation. The
results of the analysis indicated that both the age at onset and
the duration of deprivation had a significant impact on the
magnitude of the relative myopia and that, together, these two
variables accounted for 30% of the variance in our population of
monocularly deprived monkeys.
Nature of the Deprivation-induced Refractive Errors
The refractive-error alterations produced by form deprivation
in the developing monkey's eyes are believed to be axial in nature
(Wiesel and Raviola, 1977; Thorn et al., 1982; Smith et al., 1987).
Interocular comparisons of axial-length differences in monkeys
monocularly deprived before two years of age indicate that form
deprivation almost always causes the deprived eye to develop a
longer axial length than its fellow nondepri ved eye (Figure 5A).
The relationship between the axial-length and refractive-error
alterations produced by monocular lid suture in young monkeys is
illustrated in Figure 5B. There is a significant relationship
between the interocular differences in refractive error and the
interocular ratios of axial length (r=0.87; df=15; p<0.01).
Moreover, there is good agreement between the experimental data and
the predicted relationship between the axial-length and
refractive-error alterations (dashed line, Figure 58). The close
relationship between the predicted and the experimental data
supports the conclusion that the deprivation-induced
refractive-error alterations are caused by an axial elongation in
the deprived eyes.
DISCUSSION
The results of our analysis demonstrate that when an
appropriate reference is available (eg., the nondepri ved eye of
monocularly lid-sutured monkeys), the effects of form deprivation
on the developing monkey's refractive status are extremely
consistent. In 45 of the 49 individual cases (92%) presented in the
literature, monocular lid suture caused the deprived eye to develop
a longer axial length and/or to become more myopic/less hyperopic
than the nondepri ved control eye. The magnitude of this relative
axial myopia depends on both the age at onset and the duration of
deprivation. It appears that much of the apparent intersubject
variability in the type and magnitude of these deprivation-induced
refractive errors can be attributed to fundamental experimental
variables and the difficulty in establishing appropriate controls.

The extent that these factors account for the observed intersubject
variability reinforces the idea that lid-sutured monkeys can be
readily used to evaluate how vision-dependent mechanisms affect the
emmetropization process.
Although the great majority of form-deprived monkeys develop a
relative axial myopia, exceptions, i.e., deprived eyes that
developed a relative axial hyperopia, have been noted by several
laboratories (von Noorden and Crawford, 1978; Harwerth et al. ,
1983; Thorn et al., 1982; Smith et al., 1987). As yet, we have not
been able to uncover any significant events in the history of these
monkeys that would explain why they failed to develop a relative
myopia in their deprived eyes. However, it is important to note
that exceptions have also been found in humans who have experienced
form deprivation early in life (von Noorden and Lewis, 1987). It
will be important in future investigations to take these exceptions
into account, but this parallel in observations on man and monkey
further increases the confidence level with which the data from
monkeys can be extrapolated to the human condition.
The results of this survey leave little doubt that anomalous
visual experience early in the life of a developing monkey can
cause significant refractive anomalies, even when the episode is as
brief as two weeks. Although it is difficult to compare the
constant occlusion associated with lid suture in monkeys to the
intermittent occlusion typically employed in orthoptic treatment
regimens, one should be aware of the possibility that therapeutic
procedures for amblyopia might result in refractive error
abnormalities that could subsequently disrupt binocular vision and
promote the development of amblyopia. In this respect, the specific
aspects of form deprivation (i.e., decreased luminance, reduced
image contrasts at specific spatial frequencies, etc.) that disrupt
the normally regulated growth of the eye and result in anomalous
refractive errors must be determined. This information should be
useful in the design of the optimal treatment and management
procedures for amblyopia.
REFERENCES
Criswell, M.H. and Goss, D.A. (1983). Myopia development in
nonhuman primates -- a literature review. Am. J. Optom. Physiol.
Opt., 60, 250-268.
Greene, P.R. and Guyton, D.L. (1986). Time course of rhesus
lid-suture myopia. Exp. Eye Res., 42, 529-534.
Harwerth, R.S., Smith, E.L., Boltz, R.L., Crawford, M.L.J., and von
Noorden, G.K. (1983). Behavioral studies on the effect of abnormal
early visual experience in monkeys: spatial modulation sensitivity.
Vision Res., 23, 1501-1510.
206 E.L. SMITH III eta!.
Rabin, J., Van Sluyters, R., and Malach, R. (1981).

Emmetropization: a vision-dependent phenomenon. Invest. Ophthalmol.
Vis. Sci., 20, 561-564.
Raviola, E. and Wiesel, T.N. (1985). An animal model of myopia. N.
Engl. J. Med., 312, 1609-1615.
Smith, E.L. III, Harwerth, R.S., Crawford, M.L.J., and von Noorden,
G.K. (1987). Observations on the effects of form deprivation on the
refractive status of the monkey. Invest. Ophthalmol. Vis. Sci., 28,
(in press). -
Sommers, D., Kaiser-Kupfer, M.I., and Kupfer, c. (1978). Increased

axial length of the eye following neonatal lid suture as measured
with A-scan ultrasonography. Invest. Ophthalmol. Vis. Sci.
(Suppl.), ..1..§., 158.
Thorn, F. , Doty, R. W. , and Gramiak R. ( 1982). Effect of eyelid
suture on development of ocular dimensions in macaques. Cur. Eye
Res., _l, 727-733.
von Noorden, G.K. and Crawford, M.L.J. (1978). Lid closure and
refractive error in macaque monkeys. Nature, 272,53-54.
von Noorden, G.K. and Lewis, R.A. (1987). Axial length in
unilateral congenital cataracts and blepharoptosis. Invest.
Ophthalmol. Vis. Sci. 28, 750-752.
Wiesel, T. N. and Raviola, E. ( 1977). Myopia and eye enlargement
after neonatal lid fusion in monkeys. Nature, 266, 66-68.
Yinon, U. (1984). Myopia induction in animals following alteration
of the visual input during development: a review. Cur. Eye Res., 3,
677-690. -
Young, F. A. ( 1961). The effect of restricted visual space on the
primate eye. Am. J. Ophthalmol., 52, 799-806.
Young, F.A. (1964). The distribution of refractive errors in
monkeys. Exp. Eye Res.,}, 230-238.
17
ELECTROPHYSIOLOGY OF CORTICAL NEURONS UNDER
DIFFERENT CONDITIONS OF VISUAL DEPRIVATION*
M.L.J. CRAWFORD
It is now well known that the monkey visual

system, at birth, is highly structured and functionally
organized to carry out the needs of vision. Examination
of the stimulus sensitivity of neurons in the newborn
visual cortex demonstrates every class of neuron seen in
the adult (Wiesel and Rubel, 1974). While all of the
neurons encountered at birth may not be as sensitive to
visual stimulation as in the adult, there seems to be a
'starter set' of neurons possessing adult properties
(Crawford et al., 1975). In the course of a few days of
normal visual experience, the population of neurons in
the infant monkey's visual cortex attains the high
sensitivity and precision seen in adult animals; the
apparent role of early normal visual experience is to
fine-tune the system as a whole for adult functions (see
Blakemore and Vital-Durand, 1986a).
The early months of visual development are

characterized by a period of malleability, where
compensation may be made for genetic or developmental
errors in the system, e.g. a severe strabismus, or a
ptosis. This is achieved by exclusion of the input from
the faulty eye from the visual processor, the visual
cortex. The result of this self-correcting compensation
is often the clinical condition of amblyopia. It is
also during this period that the clinician has the
opportunity to apply his skills to intervene in this
natural corrective procedure, to repair the error and
prevent, or to reverse a functional impairment. The
clinician has at his disposal a number of methods of
intervention, but is faced with the decision of what to
do within considerable time constraints. The better
understood the underlying developmental processes, the
*Supported by NIH grant EY01120.
207
208 M.L.J. CRAWFORD
better chance that the prompt and proper application of

therapy will be effective.
It is during this pliable period that conditions

simulating these clinical entities may be produced
experimentally in normal animal models in order to study
and understand the underlying processes. The past 25
years have seen exponential growth in experimentation
directed toward this end. As the topic of visual
development, normal and abnormal, has been reviewed
several times in the last few years (e.g., Movshon and
Van Sluyters, 1981, and Boothe et al., 1985), I shall
describe only some of the changes in cortical
electrophysiology of the visual system with different
forms of experimentally induced abnormal visual
experience. The electrophysiology to be considered is
restricted here to the behaviors of cortical neurons as
measured by extra-cellular microelectrodes. In most
cases I will present and discuss experiments performed
on young monkeys, pointing out the similarities and
differences between the monkey and cat when it is
important to the point. The young monkey is the model
of choice for confidence in the extrapolation to the
problems seen in the clinic.
The newborn monkey striate cortex is highly

organized at birth, and within a few days or weeks
appears no different from that of the adult. It is
during this period that the spatial domains for the two
eyes (eye-dominance columns) become more clearly defined
(morphologically, at least) in the striate cortex, even
though the basic electrophysiological pattern is evident
(Wiesel and Hubel, 1974). A microelectrode passed
through the striate cortex of the infant monkey (and as
well in the normal adult monkey) reveals the following
organization. Perpendicular electrode penetrations most
frequently encounter large numbers of binocular neurons
( > 85%) but each being dominated by one eye or the
other. Widely spaced, and parallel perpendicular
penetrations through the superficial cortical layers
reveal similar eye dominance, right or left, of
binocular cells. In each penetration, the best
orientation of the stimulus (edge, stripe, bar etc.)
remains essentially the same.
To the contrary, an oblique electrode penetration

(the most frequent case) makes clear the eye-dominance
domains as sequentially encountered neurons slowly
change in eye-dominance from being right- to left-eye
dominated in alternating fashion (see Hubel, 1977,
fig.7, and Blakemore et al., 1978).
As the electrode passes through the cortex, the

CORTICAL NEURONS AND VISUAL DEPRIVATION 209
eye-dominance of cells slowly changes back and forth

with increasing electrode depth; most neurons having
degrees of binocular input. A cytochemical correlate of
this normal eye-dominance arrangement is revealed in the
distribution of the metabolic enzyme, cytochrome oxidase
(CO; Wong-Riley, 1976,1979; Wong-Riley and Carroll,
1984). When normal afferent input to the cortex is
blocked for a few days, CO concentration is reduced in
the cortical zone controlled by the affected eye and,
when seen in the coronal sections, appear as alternating
CO-rich and CO-poor columns (see Wong-Riley and Carroll,
1984) .
If the electrode passage is tangential to the brain

surface (and directed at right angles to the eye-
dominance slabs), the zones of eye-dominance are even
better indicated (see LeVay et al., 1981, fig. 2) •
In layer 4C, the principal afferent recipient zone,

the neurons encountered in passage are largely
monocular, with the few binocular neurons being
encountered at the zone of transition between monocular
cells. Similar tangential penetrations above or below
layer 4C result in similar progressions of alternation
in eye-dominance with the principal difference being
that the extent of the binocular zone is widened, and
eye-dominance changes more slowly with electrode
excursion. Again, CO staining of comparable tangential
sections reveals a series of alternating CO-rich and CO-
poor stripes in layer 4C which have been shown to
represent the extent of the eye-dominance zones; CO-rich
for the normal and CO-poor for the blocked eye. With
monocular blockade, the CO density differences are seen
throughout the six layers of cortex, being characterized
in supragranular layers 2 and 3 by strings of 'blobs'
which alternate in CO-rich and CO-poor stripes.
In summary, the highly structured system of

cortical receptive field organization and eye-dominance
domains for neurons is laid down, in the rough, at birth
and becomes only more refined in degree with normal
visual experience.
Abnormal visual experience during early life can

dramatically alter the functional domains of the two
eyes in the cortex, and there are several indicators
of these dramatic changes. Assuming some monocular
defect, e.g., experimental lid closure, a ptosis, or
cataract, the foremost electrophysiological effect
seen in the striate cortex is the loss in control that
the defective eye has over cells. In experimental
animals so treated, the probing electrode fails to
encounter as many neurons influenced by the defective
210 M.L.J. CRAWFORD
eye, while the relative encounter rate for the normal

companion eye increases. With long-term deprivation,
the progression in the imbalance in cortical continues
to the virtual exclusion of the defective from the
cortex (see Hubel et al., 1977 and Blakemore et al.,
1978).
...
, ,
•-....u
_.., .
c __ ......................
_ _ _ c . ...... . _ ... , ,. ruo .. ,,
·~(.11
~ .. ::....~ :;'.."':,'~.:::.!'~~~·:. o 1 o 01ta Uo o l
;,
j
j ...
• !
'
!. L
:· .·...
~ • 4 • •• "
o•rOittOC• IC 11110 1
..
l~l!oOl
cs....-c, o
,.,,..,._,...,.,. , ,,,...r••
<1.1 . _ _ , _ . . . . . .. .. - •• • •• ••, , . ., .
........c,.,
, ,--.
~t
. . ..... ..........,
. ............... lo•WIO ....... . 0 00 I I to Uoo l
L
'
' ...
11 Jl 1111 • ~ •
o•o,ool ORIDIIII•C.• I( OUII
Figure 1. Threshold contrast sensitivity functions

for four macaque monkeys deprived at 30 days of age
by monocular occlusion of one eye for 14, 30, 60,
and 90 days duration. All were then reverse
deprived for 120 days, and then tested at two years
of age. The sensitivity of normal monkeys is
indicated by the shaded area (mean,+,- lsd).
Virtually complete recovery to normal is seen for
Subjects C and S having had up to 30 days of initial
deprivation, but to the complete disadvantage of the
reverse deprived eye. Initial deprivations of 60
and 90 days resulted in compromise of function for
both eyes, the initially closed eye not recovering
completely, as well as a deprivation effect
occurring in the most recently closed eye (from
Crawford et al., in preparation).
Normal
~
79%
A % D %
B
%tl=n E
~ Ofo
-LI::.n
0
1 4 7
F-LJjT-
60% 90/120 60% ED
c 40%
200fo
20% 11-N
1 2 3 4 5 6 7
ED F M B •
Figure 2. Cortical eye-dominance (A-C,E) and

orientation tuning (F) for 212 neurons recorded in
13 electrode penetrations into the striate cortices
of three reverse-deprived and one normal monkey (D).
The open bars represent dominance of cells by the
reverse-deprived eye: the solid ones represent the
initially deprived eye. It is clear that most
cortical cells were dominated by the initially
deprived eye, which had the better vision (re Fig.
1). Less than 5% of the neurons had binocular input
(E) whereas 79% of the cells from the normal monkey
had binocular influence (D). There were no defects
seen for orientation tuning (OT;F) in these animals
relative to normal. About the same percentage of
cells had fine (F=+-10deg.) moderate (M=+-20deg.)
broad (B=any preference) and no stimulus orientation
preference(*). (Crawford et al. in preparation).
An impressive characteristic of the primate visual
system is the rapidity and extent to which the
deprivation effects can be reversed (Blakemore and Van
Sluyters, 1974 (kitten): Blakemore et al., 1978ab: and
LeVay et al., 1981 (monkey)) by reverse deprivation
initiated in the early months of life. We have used the
reverse deprivation paradigm in young monkeys after
different durations of monocular deprivation for the
principal purpose of determining the precision of the
reversal process for several psychophysical functions
(Crawford et al., In Preparation). Examples of tha
212 M.L.J. CRAWFORD
extent of recovery in psychophysical contrast

sensitivity are shown in Figure 1, and the recovery in
cortical electrophysiology in Figure 2.
The contrast visual sensitivity for the initially

closed eye (open data points) is completely recovered
to the normal range if the companion eye (filled data
points) is closed before 60 days of age and for 120
days duration (Fig. lA & B). However, at 90 and 120
days of age, the same duration of reversal becomes
progressively less effective, so that with 90 days of
initial deprivation, and 120 days of reverse
deprivation, begun at 120 days of age, the two eyes
end up with a significant, but balanced, bilateral
amblyopia.
The cortical eye dominance in these monkeys showed

a dramatic absence of binocular cells ( <5%, Fig. 2E) and
the majority of cortical cells were driven by the
initially closed eye (Fig. 2A-C), i.e., there was a
complete reversal in cortical eye-dominance. As the
electrode passed obliquely through the cortex of these
animals, successive neurons were usually monocular with
no stimulus-elicited background response heard from the
stimulation of the opposite eye. With electrode
advancement, the switch in eye-dominance to the opposite
monocular control usually came within a short distance
and was complete, with only one or two binocular neurons
being encountered. Moreover, these few binocular
neurons were always weakly responsive and showed poor
receptive field specificity. Overall, however, the
receptive fields for the monocular neurons driven by
either eye differed in no significant way from normal,
ie. they showed no defect in high orientation tuning,
field size, or motion sensitivity. For example, when
precision of stimulus orientation tuning was used as a
criterion, the reverse-deprived monkeys had the full
complement of sensitive cells which did not differ from
normal (Fig. 2F)
An interesting property of the cortical
reorganization seen in the results of reverse
deprivation in kittens (Van Sluyters, 1978; however, see
Movshon, 1976 for a contrary result) and monkey
(Blakemore et al., 1978) is that the orientation
preference of successively encountered neurons seems to
change smoothly with electrode travel, even at the point
of rather sharp shifts in eye dominance, suggesting that
when the initially deprived eye regains control over the
cortical cell, the new receptive field has the same
stimulus orientation preference as that originally
there. It has been suggested (Blakemore et al., 1978)
that the "seeds" of the original receptive field and
ocular-dominance pattern are retained and "sprout" with

reverse deprivation to restore the original cortical
organization. Our data are consistent with such an
interpretation.
Models of these dramatic changes in the control

which an eye has over cortical neurons following natural
or experimental monocular deprivation have central to
them the concept of binocular competition. In the
absence of any convincing electrophysiological evidence
of a deprivation effect in the retina or lateral
geniculate nucleus (all the more amazing considering the
enormous change which occurs in LGN cell size: see the
paper by Headen et al., 1987) it is clear that the major
site of the effect of monocular deprivation must be in
the striate cortex (Blakemore and Vital-Durand, 1986).
We have proposed a dual mechanism for these effects:
disuse, in that retinal and LGN cells are not exercised
by light pattern stimulation during the first weeks of
life, which in turn leads to their ineffectual
competition (LGN terminals) in the striate cortex (von
Noorden et al, 1976). Originally proposed by Wiesel and
Rubel (1963ab, 1965) and promoted by Guillery and
associates (Guillery and Stelzner, 1970: Guillery, 1972:
Sherman et al., 1974), binocular competition is seen as
a pushing and shoving match of the LGN axonal terminals
at the point of convergence upon the dendrites of the
first-order cortical neurons. Thought to be well placed
at birth (Wiesel and Rubel, 1974), synaptic terminals of
afferents of the right and left eye tussle in retaining
their respective patterns: the defective or deprived eye
losing in the competition. Continuing the metaphor, as
the competition progresses and synaptic contacts are
lost, the terminal arborization of the afferents either
"shrink" or, they are "pruned back" resulting in
imbalanced spatial domains in the cortex, and
secondarily, in the eventual reduction in the soma size
of the LGN cells. Neither the mechanism or the site of
this competition is known.
Correcting the monocular defect in the primate has
little effect upon restoration of visual function.
Indeed, it may well be that a brief monocular
disadvantage sets in motion the deprivation process
which continues to deepen even though clear images are
again presented to the retina. It is only when the
companion dominant dye is placed at some disadvantage
(patching, blurring, reverse-patching, etc.) that
significant recovery is seen in the amblyopic eye. This
sequence of events and manipulations are consistent with
the binocular competition model for the inducement and
for the recovery from monocular amblyopia, in that the
shrunken terminal fields of parvocellular LGN afferents
214 M.L.J. CRAWFORD
appear to re-expand to normal size in layer 4C of the

striate cortex (Blakemore et al., 1980~ Crawford et al.,
In Preparation).
On the other hand, binocular deprivation also

produces severe bilateral amblyopia (see paper by
Harwerth~ this volume). The principal cortical effect
of bilateral eyelid closure occurring early in life
occurs in the loss of striate binocular cells (Wiesel
and Hubel, 1974). Examples of eye-dominance changes are
shown in Figure 3.
The data from two monkeys having had bilateral lid

suture from 30 days of age for 14 and 92 days duration
are shown in comparison to a control monkey. We
examined the eye-dominance and degree of orientation
sensitivity for 212 neurons encountered on 10 oblique
penetrations through foveal striate cortex of these
CORTICAL EYE DOMINANCE FOR BINOCULARLY

DEPRIVED AND NORMAL MONKEYS
50
IIIllL853B090
45
D 0328MBD14
:
B NCliMI 01 = 440)
~
I
1!5 25
.o
... 15
10
EYE ~ANCE n.E = 1; BIN= 2-6; RE = 7)
Figure 3. Eye dominance histograms for 440 cortical

neurons recorded from two monkeys having had 14 and
90 days of binocular deprivation, compared to a
control monkey. The eyes were closed at 30 days of
age, opened at 45 and 120 days of age, respectively.
These recordings were made at about three years of
age. The obvious effect of binocular deprivation
was in the loss of binocular neurons {16%) relative
to normal (80%). (Crawford et al., in preparation).
experimental animals. Each eye was found to be well

represented in the monocular eye-dominance groups (1 and
7) while the binocular categories were markedly reduced
from a normal 80% to only 16% in each case. This was
most evident as the electrode traversed the boundaries
between eye-dominance zones where, as we had seen with
reverse-deprived animals, there was a sharp transition
from monocular cells controlled by the right eye to
monocular ones connected with the left eye. The few
binocular cells, however, showed no reduced orientation
tuning such as. we had seen in the reverse-deprived
animals. Surprising as well, there was no significant
loss in the degree of orientation tuning among the
monocular cells following binocular deprivation.
Therefore, both monocular and binocular deprivation have
devastating and permanent effects upon binocular cells,
while the orientation tuning of the monocular cells are
less affected.
What then is the mechanism accounting for the loss

in binocular connections and visual contrast sensitivity
with bilateral deprivation? Obviously, it cannot be the
dominance of one input over another in competition, as
accepted for monocular deprivation. The most reasonable
explanation is that without simultaneous coherent
binocular stimulation, the cortical system loses
functional binocular connections through disuse,
resulting in an obvious loss in binocular function. In
the absence of a robust competition (as with monocular
deprivation) the effects of simple lack of exercise
(disuse) results in a relatively less severe amblyopia
than that seen with comparable periods of monocular
deprivation.
R E F E R E N C E S
Blakemore, c., Garey, L. J., Henderson, z., Swindale, N.

v., & Vital-Durand, F. (1980). Visual experience can
promote rapid axonal reinnervation in monkey visual
cortex. J. Physiol. Lond., 307, 25P-26P.
Blakemore, c., Garey, L. J., & Vital-Durand, F. (1978).

The physiological effects of monocular deprivation and
their reversal in the monkey's visual cortex. J.
Physiol. Lond.,283, 223-262.
Blakemore, c., & Van Sluyters, R. c. (1974). Reversal of

the physiological effects of monocular deprivation in
kittens: further evidence for a sensitive period. J.
Physiol., 237, 195-216.
Blakemore, c., Vital-Durand, F., & Garey, L. J. (1981).

216 M.L.J. CRAWFORD
Recovery from monocular deprivation in the monkey. I.

Reversal of physiological effects in the visual cortex.
Proc. R. Soc. Lond., 213, 399-423.
Blakemore, c. & Vital-Durand, F. (1986a). Organization

and post-natal development of the monkey's lateral
geniculate nucleus. J. Physiol., 380, 453-491.
Blakemore, c. & Vital-Durand, F. (1986b). Effects of

visual deprivation on the development of the monkey's
lateral geniculate nucleus. J. Physiol., 380, 493-511.
Boothe, R. G., Dobson, v., & Teller, D. Y. (1985).

Postnatal development of vision in human and nonhuman
primates. Ann. Rev. Neurosci.,~, 495-545.
Crawford, M. L. J., Blake, R., Cool, s. J., & Von
Noorden, G. K. (1975). Physiological consequences of
unilateral and bilateral eye closure in macaque
monkeys: some further observations. Brain Res., 84,
150-154.
Guillery, R. w. Binocular competition in the control of

geniculate cell growth. (1972). J. Comp. Neurol., 144,
117-130.
Guillery, R. w., & Stelzner, D. J. (1970). The
differential effects of unilateral lid closure upon the
monocular and binocular segments of the dorsal lateral
geniculate nucleus in the cat. J. Comp. Neurol.,l39,
413-422.
Harwerth, R. s., Smith, E. L., Duncan, G. c., Crawford,
M.L.J., & Von Noorden, G.K. (1986a). Effects of
enucleation of the fixating eye on strabismic amblyopia
in monkeys. Invest. Ophthalmol. Vis. Sci., 27, 246-254.
Harwerth, R. s., Smith, E. L., Duncan, G. c., Crawford,

M. L. J., & Von Noorden, G. K. (1986b). Multiple
sensitive periods in the development of the primate
visual system. Science, 232, 235-238.
Hubel, D. H. (1979). The visual cortex of normal and

deprived monkeys. American Scientist, 67, 532-543.
Hubel, D. H., & Wiesel, T. N. (1963b). Receptive fields
of cells in striate cortex of very young, visually
inexperienced kittens. J. Neurophysiol., 26, 994-1002.
Hubel, D. H., & Wiesel, T. N. (1977). Functional

architecture of macaque monkey visual cortex. Proc. R.
Soc. Lond., 198, 1-59.
Hubel, D. H., Wiesel, T. N., & LeVay, S. (1977).

Plasticity of ocular dominance columns in monkey
striate cortex. Phil. Trans. R Soc. Lond.,278, 377-409.
LeVay, s., Wiesel, T. N., & Hubel, D. H. ( 1981). The

postnatal development and plasticity of ocular-dominance
columns in the monkey. In, The organization of the
cerebral cortex. (eds. F. o. Schmitt, F. G. Worden, G.
Adelman & s. G. Dennis). MIT Press, Boston.
Movshon, J. A. (1976). Reversal of the physiological

effects of monocular deprivation in the kitten's visual
cortex. J. Physiol. Land., 261, 125-174.
Movshon, J. A., & Van Sluyters, R. c. (1981). Visual

neural development. Ann. Rev. Psycho!., ~. 477-522.
Sherman, s. M., Guillery, R. w., Kaas, J. H., &

Sanderson, K. J. (1974). Behavioural, electrophysio-
logical and morphological studies of binocular
competition in the development of the geniculocortical
pathways of cats. J. Camp. Neurol.,l58, 1-18.
Van Sluyters, R. c. (1978). Reversal of the

physiological effects of brief periods of monocular
deprivation in the kitten. J. Physiol., 284, 1-17.
Vital-Durand, F., Garey, L. J., & Blakemore, (1978). c.

Monocular and binocular deprivation in the monkey:
morphological effects and reversibility. Brain Res.,l58,
45-64.
Von Noorden, G. K., Crawford, M. L. J., & Middleditch,

P. R. (1976). The effects of monocular visual
deprivation: disuse or binocular interaction?. Brain
Res. 111, 277-285.
Wiesel, T. N., &Hubel, D. H. (1963). Single-cell

responses in striate cortex of kittens deprived of
vision in one eye. J. Neurophysiol., 26, 1003-1017.
Wiesel, T. N., & Hubel, D. H. (1963b). Effects of visual
deprivation on morphology and physiology of cells in
the eat's lateral geniculate body. J. Neurophysiol.,
26, 978-993.
Wiesel, T. N., & Hubel, D. H. (1965). Comparison of the

effects of unilateral and bilateral eye closure on
cortical unit responses in kittens. J. Neurophysiol.,
28, 1029-1040.
Wiesel, T. N., & Hubel, D. H. (1974). Ordered

arrangement of orientation columns in monkeys lacking
218 M.L.J. CRAWFORD
visual experience. J. Camp Neural., 158, 307-318
Wong-Riley, M., & Carroll, E. w. (1984). Effect of

impulse blockage on cytochrome oxidase activity in
monkey visual system. Nature, 307, 262-264.
Wong-Riley, M. T. T. (1976). Endogenous peroxidatic

activity in the brain stem neurons as demonstrated by
their staining with diaminoabenzidine in normal squirrel
monkeys. Brain Res., 108, 257-277.
Wong-Riley, M. T. T. (1979). Changes in the visual

system of monocularly sutured or enucleated cats
demonstrable with cytochrome oxidase cytochemistry.
Brain Res., 171, 11-28.
18
THE SENSITIVE PERIODS OF THE MONKEY'S
VISUAL CORTEX
COLIN BLAKEMORE
INTRODUCTION
The mechanisms of differentiation and interaction of

developing tissues remain deeply mysterious. Presumably, in most
cases, the crucial events underlying the timing of developmental
changes are determined by the expression of genes. In advanced
mammals, most of the major structural developments, in the brain no
less than in the rest of the body, take place before birth.
However, although little if any multiplication of neurons occurs
after birth, nerve cells certainly can continue to grow,
differentiate and form new dendrites, axons and synapses (see
Purves and Lichtman, 1985). Not only do these positive changes
contribute to post-natal development of the brain but so do a
number of important negative events. There is growing evidence
that, in many parts of the nervous system, there is an initial
overproduction of neurons and/or axonal connections, and that much
of the refinement of connectivity in the developing brain involves
the death of unwanted cells and the elimination of "inappropriat~'
axons and synapses (see Cowan et al., 1984). There are now several
examples of such regressive developmental phenomena occurring
postnatally.
The timing of growth and interaction after birth can, then, be

seen as a natural continuation of a carefully orchestrated
developmental programme that has largely been completed before the
moment of birth. However, any changes occurring postnatally have a
much increased probability of being influenced by functional
activity in the nervous system of the young animal. It is, then,
of special interest to consider whether any of the late events in
development, delayed until after birth, might take advantage of
information about the animal's environment conveyed to the brain,
by means of the activity of sensory systems.
In 1963, Wiesel and Hubel reported that the balance of input

from the two eyes on to individual neurons in the eat's striate
219
220 C. BLAKEMORE
cortex can be grossly modified by a period of monocular deprivation

early in life. Although the majority of cortical neurons can be
excited through either eye as soon as they start to respond to
visual stimuli (at about the time of natural eye opening), the
physiological input from one eye effectively disappears almost
completely if that eye is deprived of vision. Hubel and Wiesel
went on to show that deprivation of both eyes does not silence
cortical cells totally but leaves the majority of them still
responsive to visual stimuli and still binocularly driven. Thus,
the dramatic loss of input caused by monocular deprivation is not a
result of the lack of visual stimulation alone. Subsequently,
Guillery and and his colleagues (see Guillery, 1972) went on to
elaborate the concept of "binocular competition", on the basis of
their work on the control of growth of neurons in the lateral
geniculate nucleus (LGN).
In 1970 Hubel and Wiesel published results that established

that there is a quite distinct "sensitive period" during which the
X
w
0
z
z
0
i= 05
u
6z 04
02
-SENSITIVE PERIOD-
0 1
OL-------L-------L-------L-------L-------L---
0 5 10 15 20 25
AGE (weeks)
geniculate
This diagram summarizes Hubel and Wiesel's experiments
demonstrating a sensitive period for the modification of ocular
dominance of cells in area 17 of the kitten. Each bar represents
results from one animal. The length of the bar shows the period of
deprivation of one eye (always the contralateral eye). The
"induction index" is the ratio of the number of neurons dominated
by the experienced eye to the total number of responsive cells.
The interrupted line indicates the value of this index for a large
sample from normal cats. (Reproduced, by permission, from
Blakemore, 1974).
SENSITIVE PERIODS OF THE MONKEY'S VISUAL CORTEX 221
binocular input to cortical neurons in kittens is modifiable, on

the basis of the relative activity in the two eyes. They showed
that monocular deprivation before about three weeks of age or after
about three months of age had very little influence on the ocular
dominance of cortical cells but that deprivation as short as a day
or so starting at about four weeks of age could cause a very
dramatic change in dominance. Fig. 1 (a graphical representation
of the effects on the eat's visual cortex of various periods of
monocular deprivation starting at various ages) shows the way in
which sensitivity rises rapidly during the fourth week and then
falls gradually until the end of the third month. The effects on
cortical ocular dominance were generally quite well correlated with
relative retardation of growth in the deprived laminae of the LGN
(at least within the binocular segment of the nucleus) and with
degradation in visual performance and visual acuity through the
deprived eye (Mitchell and Timney, 1984).
Thus, from this work in the cat, it seems that the balance of
input to the cortex from the axons of right-eye and left-eye
geniculate cells is controlled by the relative activity of those
axons during a fairly distinct but quite long sensitive period
after birth. At the height of this period, as little as a few
hours of monocular deprivation has a detectable effect on cortical
ocular dominance.
THE SENSITIVE PERIODS OF THE PRIMATE STRIATE CORTEX
The Period of Plastid ty of Layer IVc
Evidence for a simple anatomical basis for binocular

plasticity first began to emerge from work on the pattern of
termination of thalamo-cortical afferent fibres in layer IVc of the
monkey. In the adult animal, axons carrying signals from the two
eyes terminate mainly in the lower part of layer IV in a
remarkable, segregated pattern. This takes the form of more-or-
less parallel alternating "stripes" of right-eye and left-eye
termination; the overall pattern is rather like a finger print and
the stripes tend to run into the boundaries of area 17 at right
angles to the border. The pattern and dimensions of these bands of
termination are very similar in all parts of the binocular portion
of area 17, each individual stripe being about 0.3 mm in width
(Rubel et al., 1977).
Now, geniculate axons begin to invade the cortical plate of

the monkey some weeks before birth, but at the time of birth the
density of terminal axons still appears very low and, most
significantly, right-eye and left-eye geniculate axons are
intermingled fairly uniformly within layer IVc (Rakic, 1977). At
about the time of birth only a slight waxing and waning of density
for each eye's set of terminals can be discerned across layer IVc.
Then, over the 6- 8 weeks after birth, in the normal monkey, the
222 C. BLAKEMORE
density of thalamic termination appears to increase and the two

sets of terminating axons become progressively segregated from each
other to form the characteristic ocular dominance stripes (Rubel et
al, 1977).
It turns out that monocular deprivation in the monkey modifies

this process of natural, postnatal segregation of termination, so
that the terminals from the deprived laminae of the LGN come to
occupy less than half the territory of layer IVc (Rubel et al,
1977; Swindale et al., 1981). Thus, the sensitive period for the
modifiability of terminal distribution in layer IV corresponds well
to the natural period of segregation of terminals. Moreover, these
anatomical results agree well with physiological changes in the
binocularity of cells recorded in layer IVc itself. In very young
monkeys, cells in IVc are often binocularly driven, whereas in
normal adults, they are mainly monocular with ocular dominance
corresponding to the stripe of termination in which they lie (Rubel
et al., 1977). A period of monocular deprivation during the first
two months of life leads not only to a change in the relative
distribution of terminals in layer IVc but also a corresponding
change in the proportion of cortical neurons functionally dominated
by non-deprived and deprived eyes in that layer (Rubel et al.,
1977; Blakemore et al., 1978, 1981).
Here, then, there seemed to be a rather simple account of the

nature of the sensitive period. Axons from cells in the right-eye
and left-eye layers of the LGN were thought to be establishing
synapses on first-order cortical cells during the first two months
of life and to be competing with each other for this synaptic
space. As long as the inputs from the two eyes remain balanced,
the two sets of terminals segregate into their normal regular
pattern. Any change in relative activity in the two sets of axons
leads to an anisotropy in terminal terri tory and a change in the
distribution of physiological input (Rubel et al., 1977). The
shift in ocular dominance for cells outside layer IVc (which are
mostly binocularly driven in the normal monkey) could then be seen
as a passive reflection of the change in terminal disttibution and
ocular dominance of cells in layer IVc.
Certainly the physiological sensitive period for the effects

of monocular deprivation on layer IVc corresponds very well with
the early anatomical period of segregation. Fig. 2 shows examples
of ocular dominance histograms for cells recorded in layer IVc of
three representative animals (Blakemore et al., 1978). In each
case the right eye was deprived by lid-suture for the period of
time shown. Early deprivation (2 - Si weeks) caused a very clear
imbalance in the fraction of cells in layer IVc dominated by the
two eyes, reflecting the change in the relative size of ocular
dominance bands. Late deprivation, starting at 11 months, had no
obvious effect at all on the physiologically-determined right-eye
and left-eye regions in layer IVc.
MONOCULAR DEPRIVATION
100
2-5 112 WEEKS

so
D CJ
20
11-16 MONTHS
10
ADULT FOR 10
61;2 MONTHS
LEFT _ _ RIGHT LEFT _ _ RIGHT
LAYER !Vc NON LAYER !Vc
geniculate
In these histograms, the ocular dominance of neurons in the
primate striate cortex is expressed according to the seven-group
system of Rubel and Wiesel (1962), arranged in such a way that
cells monocularly dominated by the left eye fall in the first bin
of the histogram, while those totally dominated by the right eye
lie in the seventh bin. Between these two extremes there are five
categories of binocular neurons, covering a spectrum of relative
dominance between the two eyes. The histograms with unfilled
blocks on the left refer to samples of non-oriented cells recorded
within layer IVc. Those on the right are for cells recorded in all
other layers (filled blocks for orientation selective cells,
unfilled for non-oriented), The ordinate represents the number of
neurons in each bin. The three monkeys whose data are represented
by these histograms were each monocularly deprived by suturing the
lids of the right eye during the period of time indicated on the
left. The eye was re-opened at the end of the period of
deprivation and recordings were immediately made from the cortex.
Very few cells were recorded in layer IVc of the adult animal and
therefore no histogram is shown for that unrepresentative sample.
(Results of C. Blakemore, L.J. Garey and F. Vital-Durand.)
224 C. BLAKEMORE
Results like these led Rubel et al. (1977) to suggest that the
plasticity of ocular dominance during the sensitive period is
dependent on the displacement from cells of layer IVc of terminals
carrying reduced activity from the non-deprived eye. Since much of
layer IVc is initially contacted by both sets of axons, the
exclusion of the less active fibres can be viewed as a passive and
irreversible degenerative loss.
The results of reverse deprivation established that the

relative activity of the two sets of axons has a positive as well
as a negative effect on synapse formation (Blakemore and Van
Sluyters, 1974; Rubel et al., 1977; Blakemore et al., 1978, 1981;
Swindale et al., 1981). Re-opening a deprived eye and closing the
initially non-deprived eye, if performed early enough in life,
causes an actual re-expansion of the shrunken stripes in layer IVc
occupied by the initially deprived axons and a concomitant increase
in the fraction of cells in layer IVc dominated by the formerly
deprived eye. Fig. 3 shows results from four monkeys whose right
eyes were initially deprived, until the age shown, at which time
the left eye was closed and the right re-opened. The results on
the left-hand side show that early reverse suturing (at Si weeks)
causes a shift of dominance in layer IVc in favour of the initially
deprived right eye. Reversal at the end of the normal period of
segregation of axons in layer IVc (8 - 9 weeks) still has some
effect but is capable only of restoring left eye and right eye
regions to roughly equal size. Very late reverse suturing (at 3~
weeks) leaves layer IVc physiologically dominated by the formerly
non-deprived, left eye. Blakemore et al. (1981) went on to show
that the shift of ocular dominance in layer IVc precipitated by
reverse deprivation is very rapid and is almost complete a week or
so after reverse suturing following an initial period of
deprivation from birth to four weeks of age. The anatomically-
defined re-expansion of the portions of layer IVc innervated by
the initially deprived eye proceeds at the same pace and, with some
sensible assumptions, this morphological change implies that a very
rapid rate of new synapse formation (about 100 new terminals per
hour per axon) is produced by the reversal in the balance of
activity between the two eyes (Swindale et al., 1981).
Neither the anatomical nor the physiological consequences of

reverse deprivation depends solely on the restoration of impulse
activity in the initially deprived-eye pathway: simple re-opening
of the deprived eye without closure of the other causes no
detectable change in the relative stripe size in layer IVc or in
the ocular dominance of cells in the cortex (Blakemore et al, 1981;
Swindale et al, 1981). Thus, the mechanism for regulating the
establishment of synaptic territory in layer IVc does not seem to
be dependent simply on the absolute impulse traffic in the two
pathways but is influenced by the ratio of activity. It seems
probable that post-synaptic neurons in the cortex, which
potentially have access to both sets of fibres during the period of
segregation, might in some way be regulating their relative success
in establishing synapses.
The Sensitive Period outside Layer IVc
In the cat, the anatomical segregation of afferent terminals

in layer IV is also complete by about 6 - 8 weeks after birth
(Shatz and Stryker, 1978) but physiological changes in ocular
dominance can be produced by monocular deprivation starting during
the third month of life (Fig. 1). In monkeys, the disparity of
sensitive periods within and outside layer IVc is even more
pronounced. Monocular deprivation starting as late as the second
REVERSE SUTURING
5112 WEEKS
D
8 WEEKS
D D
9 WEEKS
0 D
L
38 112 WEEKS
0
LEFT----+ RIGHT LEFT-RIGHT
LAYER !Vc NON LAYER !Vc
geniculate
Ocular dominance histograms, arranged as in Fig. 2, are shown
for four monkeys, whose right eyes were initially deprived until
the age indicated and then were re-opened while the left eye was
deprived for at least a further sixteen weeks. (Modified from
Blakemore et al, 1978),
226 C. BLAKEMORE
half of the second year of life can produce a detectable shift in

ocular dominance (Rubel et al., 1977; Blakemore et al., 1978). The
ocular dominance histograms on the right side of Fig. 2 show the
effects of periods of monocular deprivation at different ages on
the ocular dominance of neurons recorded outside layer IVc in the
monkey striate cortex. Clearly deprivation from 11 - 16 months,
which has no obvious effect on the organization of layer IVc,
causes a substantial bias of ocular dominance towards the non-
deprived eye. Fig. 6 shows an equally clear shift in ocular
dominance in an animal deprived from 14 - 24 months.
The clear difference in the duration of the period of

sensitivity to monocular deprivation outside and within layer IVc
strongly suggests that the underlying mechanisms for sensitivity to
deprivation are different. Whereas the physiological changes in
layer IVc seem to involve a simple, explicit anatomical
redistribution of afferent terminals, those outside layer IVc may
depend on a redistribution of the terminals of cortical
interneurons (perhaps belonging to spiny stellate cells receiving
direct, monocular input in layer IVc) on to cortical neurons in the
layers above and below. This secondary, activity-dependent
redistribution of functional input may also follow "competitive"
rules but with a time-scale different from the process of
segregation in layer IVc. In addition, some of the changes in
physiological influence outside layer IVc produced by periods of
monocular deprivation may depend on subtle changes in synaptic
efficiency (through the competitive regulation of relative synaptic
gain), rather than the frank redistribution of axon terminals
(Blakemore et al., 1982).
Interestingly, the effects of reverse suturing at different

ages in monkeys show a less distinct difference between layer IVc
and the other layers of the cortex. The right-hand side of Fig. 3
shows ocular dominance histograms for cells recorded outside layer
IVc in animals reverse sutured at different ages. The degree of
functional re-innervation by the initially deprived right-eye seems
roughly similar both within and outside layer IVc, whatever the age
of reverse suturing. Reverse deprivation at 38! weeks, or even 9
weeks, seems to have little restorative effect on the ocular
dominance of cells outside layer IVc, even though monocular
deprivation starting at those ages would have a dramatic effect.
DEVELOPMENT OF SPATIAL RESOLUTION AND THE BASIS OF AMBLYOPIA
It is obviously tempting to take the effects of monocular

deprivation on the primate visual cortex as a model of amblyopia -
at least of deprivation amblyopia. After all, amblyopia can be
produced by deprivation of form vision in one eye and there is
clearly a sensitive period for its induction. However the analogy
between the clinical state of amblyopia and the anatomical and
physiological results already described is a tenuous one.
Amblyopia is defined in terms of a reduction in visual acuity (even
though it is usually characterized by a general impairment of

vision through the affected eye). There is no unequivocal reason
why a modest decrease in the territory of innervation in layer IVc
or even a gross shift in ocular dominance of cortical neurons
should lead to a decrease of visual acuity through the eye in
question. In order to start to understand the basis of amblyopia
we need to know the factors that limit visual acuity, the way in
which they develop and the manner in which they are affected by
monocular deprivation and other conditions that seem to precipitate
amblyopia.
Visual acuity determined behaviourally, in both human and

monkey babies, improves dramatically (by at least a factor of 30)
over a long period of time after birth (see Boothe et al., 1985).
Even in monkeys, where the rate of improvement seems to be about
four times more rapid than in human infants, visual acuity and
contrast sensitivity do not reach fully mature levels until about
the end of the second year of life.
Blakemore and Vital-Durand (1983, 1986a) have determined the

spatial resolution and contrast sensitivity of individual neurons
in the LGN and striate cortex of monkeys, from the day of birth
onwards (see Fig. 4). In both structures, there is a general
improvement in these spatial properties, which parallels the
increase in behavioural performance (although it appears that
perceptual acuity, determined by the preferential-looking method,
is substantially worse than the performance of cortical cells for
the first few weeks of life, suggesting that either this
behavioural method underestimates the true ability of young animals
or that acuity in such animals is limited by some immaturity of
spatial properties beyond the striate cortex).
The spatial structure of the receptive fields of neurons in

LGN and striate cortex take about two years to become fully mature
in monkeys. The ocular dominance of cells outside layer IVc of the
cortex is, likewise, vulnerable to the effects of monocular
deprivation until about two years of age. This might lead to the
simple conclusion that monocular deprivation (and perhaps other
conditions that predispose the animal to amblyopia) interfere with
the maturation of a neural process that underlies the improvement
of spatial performance. Unfortunately, this simple hypothesis
cannot be true.
First, the fact that there is a very considerable improvement

in the spatial performance of LGN neurons during the first two
years of life suggests that the factors that determine the gradual
improvement of visual acuity lie peripheral to LGN neurons,
probably in the eye itself (Blakemore and Vital-Durand, 1986a).
However, even very prolonged monocular deprivation from birth
(which certainly causes severe amblyopia in monkeys) seems to cause
little if any retardation of the development of spatial properties
at the level of the LGN (Blakemore and Vital-Durand, 1986b). This
228 C. BLAKEMORE
suggests, then, that amblyopia is not simply due to interference

with the maturation of factors underlying the normal development of
visual acuity.
In addition, examination of the various components that might

contribute to the improvement in the spatial structure of the
receptive fields of geniculate cells suggests that changes in the
properties of neurons are likely to play a minor role in the
developmental process and that these peripheral "neural factors"
operate only between about three weeks and six months of age (D.S.
Jacobs and C. Blakemore, in preparation). Much of the improvement
in spatial resolution of LGN cells seen before six months of age
and all of the subsequent increase are probably accounted for by
passive changes in the linear packing of foveal cones and growth of
the eye (Fig. 4).
AGE (days)
10 20 50 100 200 500 1000
50 50
.
"'
~
~
"!
0 20 20
>-
0
z
0 10 10
t=
3
0
(j)
UJ
a:
...J
<(
t=
<( Behavioural acuity
"'-
(j)
10 20 50 100 200 500 1000

AGE (days)
geniculate
These curves represent limits to the resolution of spatial
detail in normal monkeys of different ages. In each case
resolution (in cycles per deg. of visual angle) is plotted against
age. The filled circles plot estimates of the absolute limit to
resolution imposed by the foveal cone mosaic, based on Hendrickson
and Kupfer's (1976) measurements of the diameter of the rod-free
area in pigtail macaque monkeys, combined with Blakemore and Vital-
Durand's (1986) measurements of the growth of the monkey eye and
hence the increase in retinal magnification. The unfilled
triangles (upright for the LGN; inverted for the striate cortex)
show the maximum spatial frequency of a high-contrast drifting
grating that produced a just-detectable response from the best-
performing neurons. (Data of Blakemore and Vital-Durand, 1983,
1986.) Finally, the filled squares plot behavioural determinations
of visual acuity, based on the preferential looking technique, for
a single pigtail macaque (F2) from the study of Teller et al
(1978). (Unpublished analysis of D.S. Jacobs and C. Blakemore.)
Although deprivation of form vision seems to have little if

any effect on the development of geniculate neurons, it grossly
interferes with the maturation of spatial properties at the level
of the striate cortex, causing cells to have spatial resolution and
contrast sensitivity similar to (or even worse than) those seen on
the day of birth. Since even the passive improvement in the
photoreceptor limit to resolution predicts almost a five-fold
increase in spatial resolution, without any change in neural
factors, deprivation must presumably be causing an actual
reorganization of the synaptic convergence at the level of afferent
input to the cortex, which leads to a coarsening of the spatial
structure of the receptive fields of cortical cells.
This work on the spatial properties of visual neurons and on

the effects of visual deprivation provides a new approach to the
question of the sensitive period, in relation to the cause of
amblyopia. It appears that the activity in the visual pathway
evoked by well-focused visual images is necessary to promote the
normal maturation of spatial properties at the level of first-order
neurons in the striate cortex. Although the entire time-scale of
spatial maturation is very prolonged, there might be a much briefer
period of time during which cortical neurons require activity in
their input in order to "select" amongst geniculate axons so as to
minimize synaptic convergence. (This might be achieved by one of a
variety of different rules of synaptic plasticity capable of
detecting correlations amongst inputs.)
Recent work by C. Blakemore and F. Vital-Durand suggests that

there might indeed be a specific sensitive period during which
cortical neurons take advantage of activity in the input in order
to "tune" their spatial properties, after which they are relatively
resistant to the effects of deprivation and their spatial
properties continue to improve passively, following the changes
occurring in the periphery. Even though there is little
improvement in the spatial structure of receptive fields before the
age of 45 days, vision during those first few weeks of life seems
to be especially important for "stamping in" the synaptic
connections between LGN axons and cortical cells. Perhaps the
clearest demonstration of this comes from the results of reverse-
deprivation in which one eye of a monkey is deprived until 45 days
and then re-opened, while the other eye enjoys normal vision until
45 days and is then deprived for a further year or more (see Fig.
5). In such an animal, most cortical neurons are monocularly
driven, with roughly equal populations dominated by the initially-
experienced and initially-deprived eves. Surprisingly, neurons
with receptive fields in the eye that had vision until only 45 days
and then was continuously deprived for many months had spatial
resolution and contrast sensitivity quite similar to cortical cells
in normal adult monkeys; on the other hand, cells connected to the
eye that had no visual experience until 45 days but then had
continous visual stimulation were extremely poor in spatial
performance. This strongly suggests that pattern vision before
230 C. BLAKEMORE
100
50
10
1t
5
0
.~
'--y-
0.5
-- .. ~ 10
"""
SPATIAL FREQUENCY (c/deg)
geniculate
For the results shown in this diagram individual neurons in
the monkey striate cortex were stimulated with drifting gratings of
different spatial frequency and different contrast to determine a
contrast threshold at each spatial frequency. The ordinate is
contrast sensitivity (the reciprocal of contrast threshold). Each
cu~e is the envelope of a substantial number of individual
contrast sensitivity curves for different cells. The thin upper
curve represents data from a sample of cells from a normal adult
monkey (Adult N). The two lowest, thin curves refer to a normal
45-day-old animal (45 d N) and a 45-day-old animal that was
deprived of vision in both eyes from the day of birth (45 d BD).
The two thick cu~es refer to responses through the right and left
eyes of a monkey that had been deprived of vision in its right eye
until the age of 45 days and then reverse sutured (restoring vision
to the right eye and depriving the left) until more than a year of
age. The results show that contrast sensitivity and spatial
resolution were higher through the eye that had had vision only
until 45 days of age. (Unpublished results of C. Blakemore and F.
Vital-Durand.)
about seven weeks of age (the period of establishment of synaptic

input to layer IVc) is crucially important in regulating the degree
of synaptic convergence between LGN axons and cortical neurons.
Lack of correlation between binocular plasticity and precipitation

of amblyopia
Taken together, the results described here suggest that the

ocular dominance of cortical neurons is not a reliable indicator of
the visual status of the animal and that shifts in ocular dominance
provoked by monocular deprivation or defocus, or by artificial
strabismus, are not necessarily correlated with amblyopia. After
all, even prolonged monocular deprivation around the peak of the
sensitive period still leaves a small population of neurons with
input from the deprived eye: indeed, in layer IVc of the monkey,
the most exaggerated reorganization of ocular dominance stripes yet
described still leaves about 20% of cortical cells dominated by
input from the deprived eye. As far as the behavioural capacity of
the animal is concerned, surely the properties and performance of
neurons through a deprived or deviating eye are at least as
important as the mere fraction of cells still responding through
that eye. Indeed, very recent work by c. Blakemore, I. Vital-
Durand and M.J. Hawken shows that shifts in ocular dominance are
sometimes not correlated with changes in the spatial performance of
cortical cells in the monkey. For instance, brief monocular
deprivation very early in life (1 - 8 days, with subsequent re-
of the deprived eye for more than a year) hardly causes a
openinggeniculate
detectable shift in ocular dominance yet leaves the contrast
sensitivity of cells distinctly lower through the deprived eye than
through the non-deprived eye.
MC V3
MD 14-24 months
j 30
w
u
"- 20
0
"'~
a:
10
::;;
:::>
z
0
1234567 10
0.1 0.5
OCULAR DOMINANCE
geniculate
These diagrams show results for a sample of neurons recorded
in the striate cortex of a monkey whose right eye was deprived by
closure of the lids from 14-24 months of age. The ocular dominance
histogram on the left (see Fig. 2) shows a substantial shift in
favour of the contralateral (non-deprived, left) eye. However, the
envelopes of samples of invidual contrast sensitivity curves shown
on the right (of the type shown in Fig. 5) suggest that spatial
performance through the deprived right eye was not severely
affected, certainly at the high spatial-frequency end of the
spectrum, where vision is normally very poor through an amblyopic
eye. The continuous line is the envelope of contrast sensitivity
curves for neurons when stimulated with drifting gratings shown to
the left, non-deprived eye. The interrupted line refers to the
sensitivity of cells when stimulated through the deprived, right-
eye. (Unpublished results of C. Blakemore, M.J. Hawken and F.
Vital-Durand.)
232 C. BLAKEMORE
An even more impressive dissociation between the plasticity of

ocular dominance and the spatial performance of cells is seen in
Fig. 6, which shows results from a monkey that was monocularly
deprived very late in the sensitive period for shifts in ocular
dominance. The right eye was closed between 14 and 24 months of
age, and the neurophysiological experiment on the cortex was
performed directly after this period of deprivation. The ocular
dominance histogram on the left side shows a substantial shift in
favour of the non-deprived, contralateral left eye. However, the
spatial resolution of neurons was indistinguishable through the two
eyes, as seen in the graph on the right. Sensitivity to contrast
at different spatial frequencies was determined for samples of
cortical neurons, through the normal eye and the deprived eye.
Smooth curves were fitted to the individual contrast sensitivity
cu~es and the compound curves shown here are the envelopes
covering all the individual sensitivity functions determined
through the normal eye (continuous line) and through the deprived
eye (interrupted line). There were differences in sensitivity at
very low spatial frequencies (below about 3 c/deg.), but such
variations at the low-frequency end of the scale are not uncommon
in samples of cells from normal animals and they reflect the great
variability of low-frequency attenuation amongst neurons. At the
high spatial-frequency end of the spectrum, however, performance of
cells through the two eyes was indistinguishable and the deriveci
acuities (the intercepts of the functions on the abscissa) were
virtually identical. If an animal like this simply used the
performance of its best cortical neurons to determine its acuity
(and its contrast sensitivity, at least above 3/deg., where the
effects of amblyopia are normally most pronounced) it would not be
amblyopic at all, despite the clear shift in ocular dominance of
its cortical cells.
In summary, caution must be exercised in the application of

the term "sensitive period" to the postnatal development of the
monkey striate cortex. The period for plasticity of ocular
dominance is clearly very different inside and outside layer IVc.
There is yet another, prolonged period of ROrmal maturation of
spatial properties, but this is largely limited by passive and
neural changes peripheral to the cortex itself. Finally, there is
a period of activity-dependence, which may coincide with the period
of terminal segregation in layer IVc, during which the spatial
performance of cortical cells is vulnerable to the effects of
deprivation.
Acknowledgements
The work described here was done in collaboration with L.J.

Garey, M.J. Hawken, Z. Henderson, D.S. Jacobs, N.V. Swindale and F.
Vital-Durand. It was supported by Programme Grants from the
Medical Research Council, London, by grants to INSERM, Unite 94,
Bran, France and by a NATO grant for international collaboration.
References
Blakemore, C. (1974). Development of functional connexions in the

mammalian visual system. Brit. Med. Bull., 30, 152-157.
Blakemore, C., Garey, L.J. and Vital-Durand, F. (1978). The

physiological effects of monocular deprivation and their reversal
in the monkey's visual cortex. J. Physiol., 283, 223-262.
Blakemore, C., Hawken, M.J. and Mark, R. (1982). Brief monocular

deprivation leaves subthreshold synaptic input on neurones of the
eat's visual cortex. J. Physiol., 327, 489-505.
Blakemore, C. and Van Sluyters, R.C. (1974). Reversal of the

physiological effects of monocular deprivation in kittens further
evidence for a sensitive period. J. Physiol., 237, 195-216.
Blakemore, C. and Vital-Durand, F. (1983). Development of contrast

sensitivity by neurones in monkey striate cortex. J. Physiol.,
334, 18-19P.
Blakemore C. and Vital-Durand, F. (1986a). Organization and

postnatal development of the monkey's lateral geniculate nucleus.
J. Physiol., 380, 453-491.
Blakemore, C. and Vital-Durand, F. (1986b). Effects of visual

deprivation on the development of the monkey's lateral geniculate
nucleus. J. Physiol., 380, 493-511.
Blakemore, C., Vital-Durand, F. and Garey, L.J. (1981). Recovery

from monocular deprivation in the monkey. I. Reversal of
physiological effects in the visual cortex. Proc. R. Soc. Lond.
B., 213, 399-423.
Boothe, R.G., Dobson, V. and Teller, D.Y. (1985). Postnatal

development of vision in human and nonhuman primates. Annual
Review of Neuroscience, ~. 495-545.
Cowan, W.M., Fawcett, J.W., O'Leary, D.D.M. and Stanfield, B.B.

(1984). Regressive events in neurogenesis. Science, 225,
1258-1265.
Guillery, R.W. (1972). Binocular competition in the control of

geniculate cell growth. J. comp. Neurol., 144, 117-130.
Hendrickson, A. and Kupfer, C. (1976). The histogenesis of the

fovea in the macaque monkey. Investigative Ophthalmology, ~.
746-756.
234 C. BLAKEMORE
Hubel, D.H. and Wiesel, T.N. (1962). Receptive fields, binocular

interaction and functional architecture in the eat's visual cortex.
J. Physiol., 160, 106-154.
Hubel, D.H. and Wiesel, T.N. (1970). The period of susceptibility

to the physiological effects of unilateral eye closure in kittens.
J. Physiol., 206, 419-436.
Hubel, D.H., Wiesel, T.N. and LeVay, S. (1977). Plasticity of

ocular dominance columns in monkey striate cortex. Phil. Trans. R.
Soc. Ser. B., 278, 377-409.
Mitchell, D.E. and Timney, B. (1984). Postnatal development of

function in the mammalian visual system. In Handbook of
Physiology - The Nervous System III, 507-555. -
Purves, D. and Lichtman, J.W. (1985). Principles~ Neural

Development. Sinauer, Sunderland, Mass.
Rakic, P. (1977). Prenatal development of the visual system in

rhesus monkey. Phil. Trans. R. Soc. Ser. B., 278, 245-260.
Shatz, C.J. and Stryker, M.P. (1978). Ocular dominance in layer IV

of the eat's visual cortex and the effects of monocular
deprivation. J. Physiol., Lond. 281, 267-283.
Swindale, N.V., Vital-Durand, F. and Blakemore, C. (1981).

Recovery from monocular deprivation in the monkey. III. Reversal
of anatomical effects in the visual cortex. Proc. R. Soc. Lond.
B., 213, 435-450.
Teller, D.Y., Regal, D.M., Videen, T.O. and Pulos, E. (1978).

Development of visual acuity in infant monkeys (Macaca nemestrina)
during the early postnatal weeks. Vision Research, 18, 561-566.
Wiesel, T.N. and Hubel, D.H. (1963). Single-cell responses in

striate cortex of kittens deprived of vision in one eye.
J. Neurophysiol., ~. 1003-1017.
19
PSYCHOPHYSICAL STUDIES OF VISUAL SYSTEM
PLASTICITY DURING CRITICAL PERIODS
OF DEVELOPMENT*
RONALD S. HARWERTH and EARLL. SMITH III
INTRODUCTION
Early in life, visual experience has an influence on the development

of neural mechanisms involved in the processing of visual information
(Hubel and Wiesel, 1970; Wiesel, 1982). During this "sensitive" or "critical"
period, an infant's visual system requires adequate stimulation for neural
information processing mechanisms to progress toward their normal adult
characteristics (Hubel and Wiesel, 1970; Mitchell and Timney, 1984). There
is considerable nervous system plasticity throughout the critical period and
adverse environmental factors can disrupt the normal developmental
processes, but after the critical period is over, abnormal sensory
environments are not effective in modifying the response properties of
visual system neurons.
The clinical literature in pediatric vision clearly illustrates that the

concepts of critical periods of development and neural plasticity have had a
fundamental impact on the diagnosis and management of visual disorders in
children (e.g., von Noorden and Crawford, 1979; von Noorden, 1985; Duane,
1986). At their most basic level, these concepts emphasize the importance
of early detection and treatment of conditions that prevent either the
formation of clear retinal images (e.g., congential cataracts or
anisometropia) or the simultaneous formation of retinal images on
corresponding points, (e.g., strabismus). Beyond this basic application, it
seems obvious that a more complete understanding of neural plasticity
during critical periods may have important implications for the
optimization of treatment strategies for children with many types of vision
disorders. To pursue this possibility, the present report will describe some
of our psychophysical investigations of visual system plasticity using a
well-established primate model for the human visual system, the rhesus
monkey. Two aspects of our studies will be considered: 1) A comparison of
the durations of critical periods of visual development for several different
*Supported in parts by grants EY01139, EY03611, and EY01120 from the

National Eye Institute.
235
236 R.S. HARWERTH and EARLL. SMITH III
psychophysical functions (Harwerth et al., 1986), and 2) The role of

binocular competition in the development of amblyopia during the early
portion of the critical period for form vision (Harwerth et al., 1987).
BEHAVIORAL STUDIES OF CRITICAL PERIODS IN MONKEYS.
Several studies have shown that there is not a unitary critical period
for the whole visual system, but rather, there are different critical periods
of development for various levels of the visual pathway (Blakemore et al.,
1978; LeVay et al., 1980; Jones et al., 1984) and, at a given level, different
response characteristics of visual neurons may have different critical
periods (Berman and Daw, 1977). On the basis of these physiological
findings, it might logically be predicted that psychophysical measures of
visual system function would exhibit different critical periods for
processing various types of visual stimuli. Therefore, we have conducted a
psychophysical assessment of critical periods of development in monkeys
through measurements of the alterations produced by monocular form
deprivation on four visual functions (scotopic spectral sensitivity, photopic
increment-threshold spectral sensitivity, spatial contrast sensitivity, and
binocular summation). The age of onset of form deprivation was varied
between animals, but its duration was held constant at 18 months.
Data for three of the psychophysical functions are presented in i'"Ig.

1. The shortest critical period was found for scotopic spectral sensitivity
(Fig. 1 A). In panel A, the logarithm of the ratio of the scotopic sensitivity
of the nondeprived eye to the sensitivity of the deprived eye is plotted for
each of the subjects lid sutured between the ages of 1 and 6 months. We
found that the dark-adapted spectral sensitivity functions for both eyes of
all of the monkeys were equally well-fit by a rhodopsin absorption curve
modified for the pre-retinal light losses of the monkey eye (Smith et al.,
1986) but, the sensitivities of the deprived eyes of the subjects lid sutured
at either 1 or 2 months of age were considerably depressed with respect to
the sensitivities for their nondeprived eyes. In contrast, all of the monkeys
lid sutured at 3 months of age or later demonstrated equal sensitivities for
their two eyes. Therefore, these data indicate that form deprivation
instituted early in life causes substantial sensory deficits for visual
functions mediated by the rod receptors, however, the critical period for
these deficits is relatively short and has ended by three months of age.
Fig. 1B illustrates the critical period data for the neuro-sensory

interactions involved in the photopic increment-threshold spectral
function. Three sets of interocular sensitivity ratios are presented to show
the sensitivity alterations at the short-, middle-, and long-wavelength
regions of the visible spectrum. The faning out of the sensitivity ratios for
subjects treated during the first two months of life illustrates that early
monocular form deprivation results in an alteration in the basic shape of
the spectral sensitivity function in addition to a large decrease in absolute
sensitivity. However, as the age of onset for form deprivation was further
delayed, the spectral sensitivity differences between the nondeprived and
deprived eyes systematically decreased until all of the animals lid sutured
at 6 months of age or later exhibited identical spectral sensitivity functions
CRITICAL PERIODS OF DEVELOPMENT 237
1A 18
'\
-~---------------
~ ....____________________ _
4 6 8 10 12 14 16 2 14 16 18 20 22
IIG~ AT TIMf' MONOCULAR DC PRIVATION !MONTHS AGE AT TIME- MONOCULAR DEPRIVATION !MONTHS'•
1C
~ .,.*':--~;\
11.0 \
g
0.0 ---------~~--~---;:=:--:::--:::--:=:--~--~--~--~--=--""--
6 e •o '' ,. 1c •s zo 22 ?•
AGE AT TIMf MONOCUI AH DEPRIVATION 1MONTHS1
Fig. 1. Determination of critical periods for three psychophysically

measured functions. The logarithms of the ratio of the sensitivities of the
nondeprived eye to the deprived eye as a function of the age of the monkey
at the time monocular form deprivation was initiated are presented for
each function. (A) lnterocular sensitivity ratios for the scotopic spectral
sensitivity mechanisms. (B) Interocular sensitivity ratios for the photopic
increment-threshold spectral sensitivity mechanisms using a 3000 td
achromatic adaptation field. Data are presented for the short-wavelength
(crosses), middle-wavelength (pluses), and long-wavelength (diamonds)
sensitive mechanisms. (C) Interocular sensitivity ratios for four
parameters which define the spatial modulation sensitivity functions
(diamonds: low-spatial-frequency slope, diamonds: peak sensitivity, crosses:
peak spatial frequency, and pluses: high-spatial-frequency slope).
for their treated and untreated eyes. Therefore, based on these

measurements, we have concluded that there are essentially two
components to the critical period for photopic increment-threshold spectral
sensitivity. With very early monocular form deprivation (2 months of age
or less), there is a large decrease in sensitivity and the deprived eye's
spectral sensitivity appears to be determined by scotopic mechanisms even
at what are traditionally considered to be photopic adaptation levels
(Harwerth et al., 1981). Later in the critical period, between 3 and 5
months of age, monocular form deprivation causes a relative sensitivity
deficit for the deprived eyes, but at these ages the sensitivity of both eyes
is determined by photopic mechanisms. Finally, the critical period for this
visual function appears to have ended by 6 months of age since lid suture at
this age or later has no effect on increment-threshold spectral sensitivity.
The deprivation-induced alterations in the monkey's spatial vision

(Fig. 1 C) have been characterized by the interocular ratios of four
independent parameters that describe the animals' spatial modulation
sensitivity functions (Klein, in preparation). The four parameters evaluated

were the peak sensitivity, the peak spatial frequency, the high spatial-
frequency slope, and the low-spatial frequency slope. In this analysis, the
ratios of the slopes of the high-spatial-frequency portion of the contrast
sensitivity functions are equal in magnitude to the ratios of the cut-off
spatial frequencies or resolution limits of the subjects. Thus, these ratios
reflect the depth of amblyopia resulting from the experimental treatment.
In comparison to the critical periods for either of the spectral

sensitivity functions, the critical period for spatial vision is quite long. The
relative effects of stimulus deprivation on the high spatial frequency slope,
the peak spatial frequency, and the peak sensitivity are quite similar and
suggest that the critical period of development for form vision lasts until
about 25 months of age. The slope of the low-spatial-frequency roll-off, in
those subjects where sufficient data were obtained to determine it, does
not provide an indication of a critical period since it does not appear to be
affected by monocular form deprivation to any great extent. Overall, the
contrast sensitivity experiments show that during the first five months of
the monkey's life the visual system is highly plastic and that the effects of
monocular stimulus deprivation are quite severe. Between 5 months and 1
year of life the amount of plasticity rapidly declines, but there is an
additional year during which abnormal visual experience can cause high-
spatial-frequency deficits in form vision. At 25 months of age, the visual
plasticity of the spatial-vision processing mechanisms has ended and even
relatively long-term form deprivation does not result in abnormal
properties for these mechanisms.
The last visual function investigated in these monkeys was binocular

summation, i.e., the improvement in visual performance with binocular
viewing compared to monocular viewing. In human or monkey subjects with
normal binocular vision, the expected improvement in sensitivity with
binocular viewing is typically 40% (0.15 log units) for spatial modulation
sensitivity measurements. Our previous experiments have also shown that
an absence of binocular summation is associated with a lack of stereopsis
and a deficiency of binocularly driven neurons in the striate cortex
(Crawford et al., 1983; 1984). Therefore, the binocular summation
experiments provide an indication of the critical periods of development
for the binocular interactions underlying sensory fusion and stereopsis.
Binocular summation data for complete spatial and temporal modulation
sensitivity functions were obtained for the monkeys whose monocular
contrast sensitivity functions were similar enough to make the binocular to
monocular comparison reasonable, i.e., those treated at 12, 18, or 25
months age. However, none of the experimental animals exhibited
binocular summation ratios that were significantly different from zero
which indicated a complete lack of benefit from viewing the stimuli with
both eyes rather than one. Therefore, it may be concluded that the
mechanisms involved in binocular summation had remained sensitive to the
effects of monocular form deprivation through the age of 25 months and
that our treatment series failed to determine the limit of the critical
period for these mechanisms.
In the context of the present workshop, there are several aspects of

these critical period studies that apply to strabismus and amblyopia in

children. The first general concept that emerges from these studies is that
the critical period of development for each unique visual function is
probably different. In addition, the overall organization of critical periods
appears to be hierarchial in the sense that functions primarily requiring
information processing in the peripheral portions of the visual system have
shorter critical periods than those requiring more central processing. For
example, most of the adaptation and light transduction processes involved
in regulating the sensitivity of scotopic vision are thought to be located in
the peripheral part of the visual system and the critical period for scotopic
visual deficits is quite short. In contrast, the sensitive period for binocular
summation, which is believed to reflect interneuronal interactions in the
visual cortex, is much longer. While these psychophysical data do not
suggest a mechanisms for critical period differentiation, the organization
of critical periods substantiates the clinical practice of treating conditions
that obstruct retinal image formation, such as congenital cataract or
ptosis, very early in the child's life. However, even conditions occuring
much later in life must be treated quickly if good visual acuity and
binocular vision are to be preserved.
A second consideration that arises from these experiments concerns

the applicability of these data, obtained from form-deprived animals, to
the more frequently encountered clinical conditions of anisometropia or
strabismus. In this respect, stimulus deprivation may be considered a
"worst case" condition, because it involves two powerful factors, i.e.,
constant, severe image degradation and anomalous binocular competition,
that produce modifications of visual functions when plasticity is present.
Because of these factors, monocular deprivation may be the best way to
determine the limits of the periods of plasticity. Other types of abnormal
visual experience, such as anisometropia or strabismus, may show different
age-dependent effects if these conditions provide periods of adequate
stimulation for normal development. There is, however, considerable
evidence that there are some neural adaptations resulting from strabismus
or anisometropia during early childhood which are substantially different
from those resulting from form deprivation (Levi and Klein, 1982; Bedell et
al., 1985; Smith et al., 1986). Therefore, while monocular form deprivation
is a suitable method for the determination of many properties of visual
system plasticity during critical periods, age-dependent effects produced
by other types of early abnormal visual experience must be thoroughly
studied in order to understand the potential variations in this basic process.
The third consideration that is important in the present context, is
the prediction of critical periods of development for humans from the data
on monkey subjects. The most important factor in the application of
developmental data from monkeys to human children is the compensation
for the relative rates of development for the two species. A comparison of
the development of visual acuity in macaque and human infants has
indicated an approximate 4:1 ratio for their rates of development (Boothe
et al., 1985). Using this ratio, it would be predicted that the critical period
for spatial vision extends to about 8 years of age in children and that the
critical period for binocular vision is somewhat longer. The estimate of 8
years as the end of spatial-vision plasticity is in reasonable agreement with
240 R.S. HARWERTH and EARL L. SMITH III
estimates of the critical period for spatial vision obtained from

assessments of visual acuity of children suffering traumatic cataract
(Vagen and Taylor, 1979; von Noorden, 1981).
BINOCULAR COMPETITION IN THE DEVELOPMENT OF AMBLYOPIA.
The experiments in the previous section have shown that abnormal

visual experience during the critical period can result in deficits in several
different visual capabilities. Many of the deficits produced by unilateral
form deprivation can be attributed to anomalous binocular competitive
interactions between the afferent inputs from the two eyes during early
development. According to the binocular competition model, inadequate
pattern stimulation of the deprived eye gives the nondeprived eye a
competitive advantage, and as a result, its neurons gain control of central
connections at the expense of those from the deprived eye (Wiesel and
Hubel, 1974; Guillery, 1972; Sherman and Spear, 1982). However, in
addition to these competitive mechanisms, noncompetitive or direct
deprivation effects also contribute to the total deficits observed in these
studies. A direct deprivation effect would be a factor if the closed eyelid
prevented sufficient retinal stimulation to maintain the amount and/or the
pattern of neural activity that is required for the maintainance of normal
neural developmental processes (Guillery, 1972; Sherman and Spear, 1982).
The experimental differentiation of competitive and noncompetitive
effects and the study of the contribition of each to the development of
amblyopia should have direct revelance to the treatment of ocular
disorders in children.
The most common experimental technique employed to differentiate

between the age-dependent effects of binocular competition and direct
deprivation has involved a comparison of the deficits produced by unilateral
form deprivation to those produced by bilateral form deprivation. With
unilateral form deprivation, both competitive and noncompetitive factors
are operational, while balanced bilateral form deprivation presumably
eliminates binocular competition and allows only noncompetitive factors to
be operational. Physiological investigations of bilateral form deprivation in
monkeys have shown that the induced alterations are less severe than those
caused by unilateral form deprivation (Wiesel and Hubel, 1974; Crawford,
this volume). Based on these physiological studies, it would be predicted
that the psychophysical consequences of bilateral form deprivation would
aJso be less severe than those produced by unilateral form deprivation.
However, it is important to empirically test this prediction in order to
completely understand normal developmental processes and, potentially, to
apply the data to clinical situations. In a recent study (Harwerth et al.,
1987), we have investigated the psychophysical effects of direct
deprivation in five monkeys reared with bilateral lid suture, starting at 4
weeks of age, for treatment periods ranging from 2 to 18 weeks.
The first question regarding the results of these experiments was

whether form deprivation by bilateral lid suture caused similar deprivation
effects for each of the subjects' eyes. Our data comparing the relative
sensitivities of the two eyes of each of the experimental subjects are
2.0 . - - - - - - - - - - - - - - - - , 2.0
1.8 2A 28
0 1.6
>=
1.4 <(
a: 1.2
>-
f-
1.0 > 0.8
>=
Ci5
z 0.4
06 Ul
rn
(!)
0.2 '3 0.0
0.0
-0.4
-0.2 -<---,-,---,---.-,-----.---.------,---r'
.12 .25 .50 1.0 2.0 4.0 8.0 16.0 32.0 .12 .25 .50 1.0 2.0 4.0 8.016.0 32.0
SPATIAL FREQUENCY (c/deg) SPATIAL FREQUENCY (c/deg)
o.3 ____________________________________________ 2C_

0
~
a: 0.2 A BD-2
i: o BD-6
+ BD-7
~0.1 ___________~------------------------------------ + BD-13
asrn o.o .. BD-18
s -0.1 \. •
X
•
MD-2
MD-8
12 .25 .50 1.0 2.0 4.0 8.0 16.0 32.0

Fig. 2. The logarithm of interocular contrast sensitivity ratios as a

function of spatial frequency for bilaterally form-deprived (BD) and
unilaterally form-deprived (MD) monkeys. The symbol table at the top of
panel A indicates the type of treatment and duration (in weeks) for each
subject. (A) Interocular contrast-sensitivity differences between the two
eyes of each of the BD and MD subjects. The sensitivity ratios for the eye
with higher sensitivity to the eye with lower sensitivity as a function of
spatial frequency are presented. (B) The contrast sensitivity differences
between the mean monocular sensitivity of eight normal control subjects
and the eye with higher sensitivity for each of the BD and MD subjects. (C)
Binocular summation ratios i.e., the ratios for the sensitivity with binocular
viewing to the sensitivity with monocular viewing as a function of spatial
frequency, for each of the BD subjects.
presented in Fig. 2A. In this figure the logarithm of the ratio of the
sensitivity of the eye with higher sensitivity to that of the eye with lower
sensitivity was plotted as a function of the spatial frequency of the test
stimulus. For comparison purposes, the 95% confidence limits for the
population variability of interocular sensitivity differences for normal
control animals is shown by the dashed lines. The data demonstrate that
only one subject, the subject treated for six weeks (BD-6), had interocular
sensitivity ratios within the range of control subjects. The sensitivity
ratios of all of the other subjects were outside the normal range, but the
magnitudes of the deviations were not ordered with respect to the duration
of form deprivation. For example, the sensitivity ratios for subject BD-18,
the subject treated for 18 weeks, were closer to the normal range than
those of the subject treated for 13 weeks (BD-13). However, in every case,
the interocular sensitivity differences for the bilaterally form-deprived

monkeys were considerably smaller than the differences caused by
equivalent periods of monocular form deprivation. The interocular
sensitivity ratios for two monocularly deprived monkeys (MD-2 and MD-8 in
Fig. 2A), illustrate that extremely large sensitivity losses occur when both
binocular competition mechanisms and direct deprivation influences are
allowed to act on the function of one eye.
Of the bilaterally deprived subjects, the data for subject BD-2 show
the largest deviations from the normal range and require special
consideration. Subject BD-2 was form deprived for two weeks, which was
the shortest treatment period used in the study, and based on the data for
subject BD-6, BD-2 would have been expected to have approximately equal
and normal sensitivity functions. However, at the end of the treatment
period, when the lids were parted, the animal developed an eyelid infection
of the left eye and he voluntarily held his left eyelids closed for about five
days while the infection was treated. Although, in a sense, this subject was
compromised for the bilateral form deprivation experiment, his data are
presented to demonstrate the clinically relevant effects of a short-term,
incomplete, monocular form deprivation during a relatively early portion of
the critical period of development. On the other hand, it is appropriate to
include BD-2 in the study since the monocular form deprivation followed
the completion of the bilateral-form-deprivation treatment and the direct-
deprivation effects of the left, unifected eye are still important to the
purposes of the present experiments. In fact, the evaluation of direct
deprivation effects was restricted to the eyes with higher sensitivities for
all of the subjects, because all but one of them developed bilateraly
asymmetric losses of contrast sensitivity.
The reason for the occurence of interocular asymmetries in the

contrast sensitivity functions of these animals is not clear. A possible
factor, i.e., a post-treatment effect from an anisometropia produced by the
period of deprivation, was ruled out because the refractive errors were
small and nearly equal in the two eyes of each of the animals. Strabismus
would not seem to be a factor, since all of the monkeys appeared to have
normal binocular fixation and unrestricted eye movements. Among the
possible causes for the interocular asymmetries in the contrast sensitivity
functions that can not be ruled out are: eye preference and natural or
surgically induced differences in the optical densities of the lids.
In order to evaluate direct deprivation effects in the presence of the

bilaterally asymmetric sensitivity deficits, the sensitivities of the less-
affected eye of each of the experimental animals were compared to the
mean monocular sensitivities of eight control animals. The spatial contrast
sensitivity ratios from this comparison are shown in panel B of Fig. 2. As
in panel A, the dashed lines represent the 95% confidence limits for the
control population variability. The data show that the sensitivity ratios for
the nondeprived eyes of both of the unilaterally deprived animals (MD-2
and MD-8) and the more sensitive eyes of the two animals bilaterally
deprived for short durations (BD-2 and BD-6) fall within the range of
normal variability at all spatial frequencies. In contrast, each of the other
bilaterally deprived animals have relative losses in sensitivity that
monotonically increases as a function of spatial frequency for stimuli above

about 1 c/deg although the magnitude of the effect does not appear to be
strongly influenced by the duration of bilateral form deprivation. This
absent or weak correlation between the treatment effect and duration is
also shown by a comparison of the data for subjects BD-6 and BD-7. Even
though BD-7 was treated only one week longer than BD-6, BD-7 shows a
large sensitivity loss while BD-6 was within the range of normal
variability. Therefore, it may be concluded that while there are direct
deprivation effects resulting from bilateral form deprivation, these effects
are somewhat variable and they are not highly correlated with the duration
of deprivation.
Two additional measures of binocular visual function, depth detection

in dynamic, random-dot stereograms (Crawford et al., 1984) and binocular
summation for grating stimuli (Fig. 2C), were also investigated in these
bilaterally form-deprived animals. The results from the two measures were
unequivocal because none of the animals demonstrated binocular
performance that was above chance or monocular performance levels. The
most important subject in this series of experiments was BD-6 who
exhibited normal monocular functions for each eye, but failed to show
binocular vision by either of these measures of binocular function.
Therefore, these results, in agreement with ocular dominance data (Wiesel
and Hubel, 1974; Crawford, this volume), demonstrate that binocular vision
mechanisms are affected by bilateral lid suture even though the deprivation
procedure has attempted to minimize the influences of binocular
competition.
In the context of the present workshop, the experimental results from

the bilaterally form-deprived monkeys are important in at least two
respects. First, the data show that in the development of functional
amblyopia, anomalous binocular competition produces more severe
alterations than direct deprivation. Although direct deprivation
mechanisms may produce spatial-vision deficits, the deficits are
considerably milder than those produced in conjunction with anomalous
binocular competition. In addition, the data suggest that if the deprivation
could be perfectly balanced, the developing visual system may tolerate
brief periods of bilateral occlusion without permanent spatial vision
defects. This situation is quite different from the effects of monocular
occlusion where, as demonstrated by the results of subject BD-2, even brief
periods of incomplete monocular deprivation can cause substantial degrees
of amblyopia. Therefore, these experiments suggest that when ocular
conditions in young children require occlusion as part of the treatment
(except for patching treatment for amblyopia) both eyes should be
occluded, even if the specific condition affects only one eye.
The second aspect of these data that may be interesting from a

clinical perspective is the extreme vulnerability of sensory binocular vision
functions to early abnormal visual experience. Our studies of visual system
plasticity using a variety of early rearing strategies, including bilateral
occlusion, have shown that the purely binocular functions such as stereopsis
and binocular summation are adversely affected even when all of the
monocular functions for each eye appear to be completely normal. The
adverse effects on binocular sensory functions cannot be easily explained

on the basis of oculomotor factors since none of the monkeys appeared to
have a strabismus or restricted eye movements. In the final analysis, it
appears that almost any type of abnormal visual experience during early
childhood will diminish the sensory binocular vision capabilities of that
child and it will be a challenge to develop treatment procedures for
amblyopia and strabismus that preserve these binocular functions.
CONCLUSIONS
The principal conclusions that may be drawn from our studies of

visual system plasticy in monkeys during the critical periods of
development are: 1) There are multiple, partially overlapping critical
periods of development and the critical period for each specific visual
function is probably different, and 2) During the critical period, deprivation
in conjunction with anomalous binocular competition is considerably more
effective in disrupting the normal properties of visual system functions
than direct deprivation.
REFERENCES
Bedell, H.E., Flom, M.C. and Barbeito, R. (1985). Spatial aberrations and
acuity in strabismus and amblyopia. Invest. Ophthalmol. Vis. Sci.,~ 909-
916.
Berman, N. and Daw, N.W. (1977). Comparison of the critical periods for
monocular and directional deprivation in cats. J. Physiol., 265, 249-259.
Blakemore, c., Garey, L.J. and Vital-Durand, F. (1978). The physiological

effects of monocular deprivation and their reversal in the monkey's visual
cortex. J. Physiol., 283, 223-262.
Boothe, R.G., Dobson, V. and Teller, D.Y. (1985). Postnatal development of

vision in human and nonhuman primates. Ann. Rev. Neurosci., ~ 495-545.
Crawford, M.L.J., von Noorden, O.K., Meharg, L.S., Rhodes, J.W.,

Harwerth, R.S., Smith, E.L. and Miller, D.D. (1983). Binocular neurons and
binocular function in monkeys and children. Invest. Ophthalmo1 Vis. Sci.,
1!, 491-495.
Crawford, M.L.J., Smith, E.L., Harwerth, R.S. and von Noorden, G.K.
(1984). Stereoblind Monkeys have few binocular neurons. Invest.
Ophthalmol. Vis. Sci.,~ 779-781.
Duane, T.D. (1986). Clinical Ophthalmology Volume 1. Harper and Row
Guillery, R. W. (1972). Binocular competition in the control of geniculate

cell growth. J. Comp. Neurol., 144, 117-130.
Harwerth, R.S., Crawford, M.L.J., Smith, E.L. and Boltz, R.L. (1981).
Behavioral studies of stimulus deprivation amblyopia in monkeys. Vision
Res., 1!_, 779-789.
Harwerth, R.S., Smith, E.L., Duncan, G.C., Crawford, M.L.J. and von
Noorden, G.K. (1986). Multiple sensitive periods in the development of the
primate visual system. Science, 232, 235-238.
Harwerth, R.S., Smith, E.L., Paul, A.D., Crawford, M.L.J. and von Noorden,
G.K. (1987). Psychophysical studies of binocular deprivation in monkeys.
Invest. Ophthalmol. Vis. (suppl), ~ 318.
Bubel, D.H. and Wiesel, T.N. (1970). The period of susceptibility to the
physiological effects of unilateral eye closure in kittens. J. Physiol., 206,
419-436.
Jones, K.R., Spear, P.D. and Tong, L. (1984). Period of susceptibility to

effects of monocular deprivation: Differences between striate and
extrastriate cortex. Soc. Neurosci., _!, 2583-2552.
LeVay, S., Wiesel, T.N. and Hubel, D.H. (1980). The development of ocular
dominance columns in normal and visually deprived monkeys. J. Comp.
Neurol., 191, 1-51.
Levi, D.M. and Klein, S.A. (1982). Differences in vernier discrimination for
gratings between strabismic and anisometropic amblyopes. Invest.
Ophthalmol. Vis. Sci.,~ 398-407.
Mitchell, D.E. and Timney, B. (1985). Postnatal development of function in

the mammalian visual system. In Handbook of Physiology, Section I: The
Nervous System, vol III. (eds. J.M. Brookhart, et al.,) American
Physiological Society, Bethesda. pp 507-555.
Sherman, S.M. and Spear, P.D. (1982). Organization of visual pathways in

normal visually deprived cats. Physiol. Rev.,~ 738-855.
Smith, E.L., Harwerth, R.S., Duncan, G.C. and Crawford, M.L.J. (1986). A
comparison of the spectral sensitivities of monkeys with anisometropic and
stimulus deprivation amblyopia. Behav. Brain Res.,~ 13-24.
Vagen and Taylor, D. (1979). Critical period for deprivation in children.

Trans. Ophthalmol. Soc. U.K.,~ 432-439.
von Noorden, G. K. (1981). New clinical aspects of stimulus deprivation

amblyopia. Am. J. Ophthalmol., _!!b 416-421.
von Noorden, G. K. and Crawford, M.L.J. (1979). The sensitive period.

Trans. Ophthalmol. Soc. U.K.,~ 442-446.
Wiesel, T.N. (1982). Postnatal development of the visual cortex and the
influence of environment. Nature, 299, 583-591.
246 R.S. HARWERTH and EARL L. SMITH III
Wiesel, T.N. and Rubel, D.H. (197 4). Ordered arrangement of orientatio
columns in monkeys lacking visual experience. J. Comp. Neurol., 158, 307
318.
20
NORMAL VISUAL DEVELOPMENT AND ITS DEVIATIONS
RICHARD HELD
Within the last decade, refined testing procedures have allowed

vision scientists to measure a number of aspects of the development
of human vision beginning in infancy. These new procedures have
been essential for extending our knowledge to the early stages of
the development of visual function when rapid rates of change occur.
One suspects that it is during this stage of rapid growth of func-
tion that the visual system may be particularly prone to aberrant
development resulting from either intrinsic or extrinsic influences.
Accordingly, this paper will review some of the major findings
which appear to bear on the developmental aspects of amblyopia.
They come from tests of grating, vernier, and stereo acuity; studies
of the development of binocularity including pre-binocular vision
and the onset ages of stereopsis, detection of binocular rivalry,
and susceptibility to the development of amblyopia. Finally, some
speculations are included on the neuronal mechanisms which may
underlie these developments.
The testing of infants requires the use of a non-verbal re-

sponse on the part of the subject. A number of methods have been
used with varying effectiveness. The most frequently used methods
have been the visually evoked potential (VEP) and preferential
looking (PL). In the former, electrodes are placed in the occipital
region of the scalp and potentials contingent upon variation of a
visual stimulus are recorded and analyzed. In the latter, two
stimuli are presented to the infant. They generally are equated in
all but one property, that which is being tested. If a significant
looking preference is shown by the infant (it prefers to look at one
stimulus more than the other), that indicates that the stimulus
property has been detected. The advantages and disadvantages of
these methods have been discussed elsewhere (Dobson & Teller, 1978;
Sokol et al, 1983; Norcia & Tyler, 1985). The following account
presents the results of their use.
247
248 R.HELD
GRATING ACUITY AND CONTRAST SENSITIVITY
Measurement
Soon after birth, infants without ocular or other pathology of

the visual system will prefer to look at a square-wave grating, if
the grating has bars no finer than one-half to one degree of visual
angle, over an otherwise equivalent stimulus fo~~d by a blank
region. Average Snellen acuity at that age is then in the range
between 20/500 and 20/1200. The very low value of acuity cannot be
attributed to optical factors for the following reasons. In the
absence of pathology the media are clear. Measures of refractive
errors do not have much effect on acuity measures (Powers and
Dobson, 1982) because the de-focusing effects of both common refrac-
tive and accommodative errors are small at the low soatial fre-
quencies to which young infants are sensitive (Banks, 1980).
Measures of acuity taken during successive months show a rapid

rise with age up to 6 months with average acuity approaching 20/100.
Reviews of these findings can be found in Atkinson and Braddick
(1981), Dobson and Teller (1978), and Held (1981). A somewhat
slower rise follows such that by one year of age acuity approaches
20/50 (Gwiazda et al, 1978, 1980). It does not reach adult levels
when tested by behavioral methods until the fourth or fifth year of
life (Birch et al, 1983A); Mayer and Dobson. 1980). When tested on
obliquely oriented gratings, the reduced acuity relative to the
vertical and horizontal, ordinarily found in adults, is not observed
in infants until approximately six months of age (Gwiazda et al,
1978, 1980).
VEP procedures yield higher measures of acuity at comparable

ages such that an approach to adult levels of acuity occurs by the
end of the first year of life, if not earlier. The cause of this
difference is not clear. For reasons mentioned below, the differ-
ence appears to result from a failure, at least in early infancy,
for the information available at the site of the evoked potential to
influence looking behavior. Aside from the changed time scale, data
on grating acuity obtained by this method have not yet revealed any-
thing radically different from those produced by the behavioral
methods including PL.
Contrast sensitivity measures are of interest because they

measure the sensitivity of the visual system over a range of spatial
frequencies. Losses in sensitivity can be restricted to low or
middle frequency bands without affecting acuity at the high cutoff
frequency. Moreover, contrast sensitivity across the frequency
spectrum can, in principle, determine the visibility of any form
which is harmonically analyzed into its component frequencies.
Measures show that soon after birth infants have contrast sensitiv-
ity of no more than one-thirtieth that of adults. They have a peak
sensitivity at a fraction of a cycle. Both measures show a steady
NORMAL VISUAL DEVELOPMENT AND ITS DEVIATIONS 249
increase over the first six months. A review of this literature can
be found in Banks and Dannemiller (1987). Measurements of contrast
threshold can provide sensitive indices of amblyopia. However,
psychophysical methods require large numbers of trials which make
their application to infants rather difficult. Perhaps new VEP
methods may make these measurements more accessible (Norcia and
Tyler, 1985).
Application to Amblyopia
As the infant's visual system increases in sensitivity to

higher spatial frequencies refractive error becomes of greater sig-
nificance. Significant degrees of ametropia can produce habitual
blurring and the well-known pathological consequences. The most
prevalent refractive error in infancy appears to be astigmatism.
Several groups of investigators have reported a high incidence of
astigmatism compared with that found in adults (reviewed by Gwiazda
et al, 1986). By two to three years of age that incidence is re-
duced to adult levels but not without having left its imprint upon
the visual nervous system, as is discussed below.
There is good reason to believe that a major factor accounting

for the rise in acuity during the early months is change occurring in
the retina. There is an increase both in density of cone receptors
in the fovea and in the length of the anterior segment of these
cones (Hendrickson and Yuodelis, 1984; Yuodelis and Hendrickson,
1986). However, this factor can hardly account for either the re-
duced acuity on the oblique axes or for that matter any appreciable
meridional differences in acuity. Although the photoreceptor array
does have some anisotropies of orientation because of its hexagonal
packing (Hirsch, 1984), these cannot explain either the oblique
effect or the occurrence of meridional amblyopia discussed below.
One suspects that significant variations in acuity resulting from
differences in edge orientation result from cortical processes
which are known to be orientationally selective.
Further evidence of cortical involvement comes from observa-

tions of interocular interaction produced by monocular occlusion and
disocclusion. Occlusion is produced either by pathology, such as
cataract, or therapeutic procedures such as patching. Measurements
of grating acuity before, during, and after such occlusion often
show that while the non-occluded eye increases in acuity, the goal
of such therapy, the acuity of the occluded eye will simultaneously
decrease. The reverse may happen upon removal of occlusion
(Jacobson et al, 1983). We termed this the tradeoff effect. In
some cases we observed an abnormally high acuity for the age (super-
acuity) resulting from these procedures (Mohindra et al, 1983).
250 R.HELD
VERNIER ACUITY
Because vernier acuity is not limited by the minimal distance

separating foveal cones, as is grating acuity, it is considered to
be a form of hyperacuity (Westheimer, 1979). As such, essential
processing must go on at higher levels of the visual system. Con-
sequently, it is of interest to test the development of vernier
acuity in infants since it reflects the maturation of such higher
processes. To our surprise, vernier acuity turned out to be less
than grating acuity (hence hypoacute) before the age of about 3
months (Shimojo and Held, 1987). However, it rises more rapidly so
that by four or five months it exceeds grating acuity (Shimojo et
al, 1984; Manny and Klein, 1984) and continues to rise at a more
rapid rate during the next few months.
Vernier acuity is a very sensitive measure of amblyopia ex

anopsia (Levi and Klein, 1982). Using this measure we recently dis-
covered that differences in habitual dioptric blur between horizon-
tal and vertical edges, resulting from early astigmatism, predict
later meridional amblyopia (Gwiazda et al, 1986). The period of
susceptibility to this form of amblyopia appears to occur during the
first two years.
While the rise in grating acuity may be accounted for in good

part by the increase in receptor density and sensitivity in the
fovea, the increase of vernier acuity to hyperacute levels requires
additional central processing. Wilson (in press) has argued that
the increasing effectiveness of inhibition in tuning the receptive
fields of cortical neurons may account for the rapid increase in
vernier acuity up to hyperacute levels. The effect of habitual blur
in producing amblyopia (either meridionally selective or ex anopsia)
is consistent with the interpretation that conditions of exposure
modulate the development of connections in the cortex either
inhibitory or otherwise. Neither orientational selectivity nor
binocular convergence occur prior to cortex.
Curiously enough, during the third, fourth, and fifth months

analysis of the data revealed a significant sex difference in ver-
nier acuity. During this period, female infants show higher acuity
than males on this test (Held et al, 1984). However, they do not
show this difference on the test of grating acuity which therefore
provides a control against the possibility that the difference re-
sults from general behavioral factors. Both forms of test use much
the same behavioral responses. As mentioned above, there is reason
to believe that vernier acuity involves an important component of
cortical processing and the sex difference appears related to that
fact. The period from 2 to 8 months of age is one of rapid synap-
togenesis in the visual cortex (Huttenlocher et al, 1982). We have
suggested that the sex difference may be caused by a difference
during this period in nerve growth influences such as might be
exerted by the presence of testosterone in male but not in female
infants (Held et al, 1984).
STEREOPSIS AND BINOCULARITY
Tests for stereopsis have been developed in a number of labora-

tories using both two-choice preference techniques (Atkinson and
Braddick, 1976; Birch et al, 1982; Fox et al, 1980; Held et al,
1980) and VEP recording (Petrig et al, 1981). All the data show
that stereopsis (the detection of coarse binocular disparity) has an
abrupt onset in most infants at ages ranging between two and six
months. This agreement between VEP and PL data stands in stark con-
trast to the difference obtained between these techniques in
measurements of grating acuity. The agreement in this case suggests
that the disagreement in the case of grating acuity results from a
real difference in neuronal mechanism and is not an artifact of the
measuring procedure. The onset of stereopsis could conceivably re-
sult from increased accuracy of vergence (Aslin, 1977). However, a
number of results argue against this simple interpretation. Perhaps
most convincing is the finding that testing with large redundant
stereograms, in which accurate vergence is unnecessary for stereop-
sis, does not appreciably alter either the age of onset or the
threshold (Birch et al, 1983B). After the onset of coarse stereop-
sis, stereoacuity improves rapidly so that within a few weeks it
approaches adult values (Held et al, 1980; Birch et al, 1982).
Since stereopsis is clearly a function of cortical processing,

its abrupt onset and perhaps its rapid rise in acuity call for
interpretation in terms of developmental changes in cortex. We have
speculated that the segregation of the ocular dominance columns in
layer 4 of the visual cortex could, in principle, account for this
abrupt onset (Held, 1985). The existing evidence suggests that such
segregation in the human visual cortex is only partial by four
months and not adult-like until six months of age (Hickey and
Peduzzi, 1987). Prior to such segregation in monkeys, thalamocort-
ical afferents from both eyes overlap in their projections to this
layer and even synapse on the same cells (Levay et al, 1980). This
arrangement implies a partial loss of eye-of-origin information at
the entry level of the visual cortex and precludes further process-
ing of the kind required for stereopsis and other functions such as
binocular rivalry. Only after segregation of the dominance columns,
with alternating columns representing inputs from the right and left
eyes, will eye-of-origin information be preserved for recombination
in disparity sensitive and mutually inhibitory circuits.
Two criticisms of this model have been made. First, the pro-
cess of segregation of the columns takes a considerable period of
time, one that varies with species. Can such a protracted process
account for the apparently sudden onset of stereopsis? To this we
can only answer that, like the onset of many other behaviors, an
underlying neuronal process which develops gradually must at some
point exceed an internal threshold thereby allowing the expression
of behavior. The second criticism says that some species, like the
tree shrew, have stereopsis but do not have segregated ocular domi-
nance columns. To this one can only reply that different species
252 R.HELD
have evolved different neuronal mechanisms. Moreover, there are

suggestions in the literature that ocular segregation may occur on a
different anatomical basis.
Confirming the model, tests for the detection of binocularly

rivaling stimuli have also shown an onset age approximately the same
as that for stereoacuity (Birch et al, 1985). In this test a non-
rivaling, binocularly fusable stimulus is paired with a binocularly
rivaling stimulus in the two-choice looking preference procedure.
The fusable stimulus becomes preferred, suggesting that the rivaling
stimulus is aversive.
We have taken a further step to test this model. One possible

consequence of the absence of eye-of-origin information would be a
simple superposition in some central representation of the images on
the two retinas. Thus if a vertical grating is presented to the
right eye and a horizontal one to the left eye, the superimposed
representation of the orthogonal gratings should appear as a plaid
or grid figure. This suggests a simple test. We know that infants
prefer to look at a grid over a grating at all ages. As mentioned
above, they also prefer to look at a fused over a rivalrous stimulus
beginning at the age of onset of stereopsis. Consequently, we pre-
sented the binocularly superimposed orthogonal grating stimuli
paired with binocularly superimposed parallel (hence fusable) grat-
ings in a two-choice looking preference arrangement. Prior to
segregation of the columns a gridlike structure should be seen and
preferred to the fused grating. However, once segregation is com-
pleted and binocularity is achieved, the orthogonal grating stimuli
should appear to rival and preference should shift to the fused
grating. The results showed that following an initial preference
for the orthogonal gratings of close to 100% during the first months
there was a sudden shift to a preference of close to 100% for the
parallel gratings (Shimojo et al, 1986). This shift occurred at the
approximate age of acquisition of stereopsis (average about 4
months). The shift is completed within an average time interval of
two weeks for each infant. The result not only clearly confirmed
the abruptness of the transition to adult-like binocularity, it is
also the first demonstration, known to us, of a complete reversal of
an infant visual preference. As such, it might provide an efficient
marker for testing the presence or absense of mature binocularity in
both the laboratory and the clinic. Finally, we found that the age
of this shift was significantly earlier in female than in male in-
fants (Bauer et al, 1986). This result is a confirmation of the
conjecture that the male-female difference is to be found for pro-
cessing at the cortical level.
The very rapid increase in stereoacuity, following the onset of

coarse stereopsis, may call for the same explanation as that accord-
ed the rapid rise in vernier acuity. The achievement of hyperacute
levels of reoslution in both vernier detection and stereopsis occurs
at roughly the same age. Both imply a rapid sharpening of spatial
tuning as the result of the presumed growth of inhibitory
connections. Moreover, the onset of detection of binocular rivalry

is obviously related to the creation of circuitry in the cortex that
allows of mutual inhibition between the two eyes.
AMBLYOPIA
How can the development of mature binocularity be related to

the amblyopia resulting from esotropia? Amblyopia, defined as a
significant difference in acuity between the eyes, first appears in
esotropic infants at about the same age as the onset of the other
binocular processes discussed above (Jacobson et al, 1981; Birch and
Stager, 1935). This suggests that adult-like binocularity is a pre-
cursor of susceptibility to amblyopia resulting from esotropia.
However, there is an alternative possibility, namely that rnisconver-
gence does not occur chronically until this age.
Vergence movements are not well coordinated for the first few
months after birth although newborns are said to converge appropri-
ately for targets at 25 and 50 ern (Slater and Findlay, 1975).
Aslin (1977) recorded vergence movements in 1-, 2-, and 3-rnonth olds
as they followed a target that continuously moved from a 50 ern to a
15 ern distance. Only the 3-month olds showed consistent convergent
following of the target. Moreover, following was better on slower
moving targets. On the other hand when convergence demand was in-
creased suddenly and stepwise by the application of base-out wedge
prisms, vergence following was effective only after an age between
4.5 and 6.0 months. Several possible explanations have been pro-
posed for this difference in results (Aslin, 1987). I should like
to propose a new one in line with the developmental model for binoc-
ularity discussed above.
During the period of overlap of the ocular dominance columns a

plausible cortical mechanism for maintaining convergence of the eyes
derives from matching the visual inputs from the two eyes in laver
4. Assuming that layer 4 is topographically organized, optimal ex-
citation could be produced by superimposing the input distributions
from both eyes. Such superposition norrnallv occurs unon correct
convergence on target regions in the visible field. This optimiza-
tion depends upon several conditions which should be met as a result
of the infant's view of its world. First, the eyes should be stimu-
lated by images that fill most of the visual field. A sparse set of
stimuli might produce maxima for many vergence angles. Second, the
images should have wide variations in contrast over the range of
detectable spatial frequencies so as to preclude saturation by input
from one eye. Third, the stimuli should not be periodic so that
shifts of integral numbers of wavelengths, produced by vergence
shifts, cannot produce the same level of excitation. Deviations
from superimposition, which should mirror correspondence of the two
retinas, would in general reduce net excitation. Assume that the
strength maintaining a particular convergent state is proportionate
to the summed excitation in layer 4. Then a small amount of hunting
254 R.HELD
could maintain convergence on either a stationary target or one

moving slowly in depth but not on a target that either moves rapidly
or makes stepwise jumps in depth. The ability to follow such a
target requires information about the depth increment and its
direction towards or away from the eyes. But this is just the in-
formation that can only be supplied by a functioning measure of
disparity which becomes available later when, according to this pro-
posal, the ocular dominance columns segregate and the disparity
mechanism functions.
Control of vergence eye movements is of course crucial to the

maintenance of normal binocular vision. The model suggests that
vergence is maintained by two different cortical mechanisms which
function in stages. First, the optimization of excitation in layer
4 is achieved by utilizing the overlap of thalamocortical afferents.
Then the mechanism of disparity sensitivity takes over as the
columns segregate and new binocular circuitry develops outside of
layer 4. Infantile esotropia is said not to appear at birth either
in human infants (Helveston, 1987) or monkey infants (Kiorpes et al,
1985). It does appear during the transition period when stereopsis
has its onset. Could it be that the transition period, when the
vergence-maintaining mechanisms switch control is the time at which
the eye movement system is most susceptible to that loss of align-
ment which characterizes infantile esotropia? If the transition
process runs into trouble from any source of deprivation, or other
factor that influences the development of the control of vergence
by the disparity system, we might anticipate some loss of vergence
control. Stereopsis does develop in these esotropes (Mohindra et al,
1985; Birch and Stager, 1985) but is, presumably, degraded as a
result of chronic vergence error. While speculative, the implica-
tions of these notions have at least the virtue of being testable by
recently developed methods some of which have been discussed above.
ACKNOWLEDGMENT
The preparation of this manuscript and research reported from

the author's laboratory has been supported by grants from the
National Institute of Health (No. 2R01-EY-1191, No. SP30-EY02621,
and BRSG 2S07RR0747-18). The author wishes to thank his colleagues
Joseph Bauer, Eileen Birch, Jane Gwiazda, Anthony Passera, Frank
Thorn, and Jeremy Wolfe for their helpful comments.
REFERENCES
Aslin, R.N. (1977). Development of binocular fixation in human

infants. J. Exp. Child Psychology, ~. 133-150.
Alsin, R.N. (1987). Motor aspects of visual development in infancy.

In Handbook of Infant Perception. (eds. P. Salapatek and L. Cohen).
Academic Press, New York, New York.
Atkinson, J. and Braddick, 0. (1976). Stereoscopic discrimination

in infants. Perception, ~. 29-38.
Atkinson, J. and Braddick, 0. (1981). Acuity, contrast sensitivity,

and accommodation in infancy. In Development of Perception. (eds.
R Aslin, J. Alberts, and M. Petersen). Academic Press, New York.
Banks, M. (1980). Infant refraction and accommodation. Inter-

national Ophthalmology Clinics, 1Q, 205.232.
Banks, M. and Dannemiller, J. (1987). Infant visual psychophysics.

In Handbook of Infant Perception. (eds. P. Salapatek and J.
Dannemiller). Academic Press, New York.
Bauer, J.; Shimojo, S.; Gwiazda, J.; and Held, R. (1986). Sex
differences in the development of binocularity in human infants.
Investigative Ophthalmol. Visual Sci.,~. 265.
Birch, E.; Gwiazda, J.; Bauer, J.; Naegele, J.; and Held, R. (1983A).
Visual acuity and its meridional variation in children aged 7 to 60
months. Vision Research, ~. 1019-1024.
Birch, E.; Gwiazda, J.; and Held, R. (1982). Stereoacuity develop-

ment ~or crossed and uncrossed disparities in human infants. Vision
Research, ~. 507-513.
Birch, E.; Gwiazda, J.; and Held, R. (1983B). The development of

vergence does not account for the onset of stereopsis. Perception,
g, 331-365.
Birch, E.; Shimojo, S.; and Held, R. (1985). Preferential looking
assessment of fusion and stereopsis in infants aged 1 to 6 months.
Invest. Ophthal. Visual Sci., ~. 366-370.
Birch, E.; and Stager, D. (1985). Monocular acuity and stereopsis

in infantile esotropia. Invest. Ophthal. Visual Sci., 1&, 1624-30.
Dobson, V. and Teller, D. (1978). Visual acuity in human infants:

a review and comparison of behavioral and electrophysiologic al
studies. Vision Research, ~. 1469-1483.
Fox, R.; Aslin, R.; Shea, S.; and Dumais, S. (1980). Stereopsis in
human infants. Science, 207, 323-324.
Gwiazda, J.; Bauer, J.; Thorn, F.; and Held, R. (1986). Meridional
amblyopia does result from astigmatism in early childhood. Clinical
Vision Sciences, l, 145-152.
Gwiazda, J. ; Brill, S. ; Mohindra, I. ; and Held, R. (19 7 8). Infant

visual acuity and its meridional variation. Vision Research, ~.
1557-1564.
256 R.HELD
Gwiazda, J.; Brill, S.; Mohindra, I.; and Held, R. (1980). Prefer-
ential looking acuity in infants from two to fifty-eight ~•eeks of
age. Amer. J. of Optometry and Physiol. Optics, 2l• 428-432.
Held, R. (1981). Development of acuity in infants with normal and

anomalous visual experience. In Development of Perception. (eds.
R. Aslin, J. Alberts and M. Petersen). Academic Press, New York.
Held, R. (1985). Binocular vision- behavioral and neural develop-

ment. In Neonate Cognition: Beyond the Blooming, Buzzing Confusion.
(eds. J. Mehler and R. Fox). L. Erlbaum Associates, Hillsdale, NJ.
Held, R.; Birch, E.; and Gwiazda, J. (1980). Stereoacuity of human

infants. Proc. Nat. Acad. Sci., 12, 5572-5574.
Held, R.; Shimojo, S.; and Gwiazda, J. (1984). Gender differences in

the early development of human visual resolution. Invest. Ophthal.
Visual Sci., ~. 220.
Helveston, E. (In Press). Origins of congenital esotropia. American

Orthoptics Journal.
Hendrickson, A. and Yuodelis, C. (1984). The morphological develop-

ment of the human fovea. Ophthalmology, 2l• 603.612.
Hickey, T. and Peduzzi, J. (1987). Structure and development of the

visual system. In Handbook of Infant Perception. (eds. P. Salapatek
and L. Cohen). Academic Press, New York.
Hirsch, J. (1984). Quality of the primate photoreceptor lattice and

limits of spatial vision. Vision Research, ~. 347-355.
Huttenlocher, P.; de Courten, C.; Garey, L.; and van der Loos, H.
(1982). Synaptogenesis in human visual cortex- evidence for synapse
elimination during normal development. Neuroscience, ]2, 247-252.
Jacobson, S.; Mohindra, I.; and Held, R. (1981). Age of onset of

amblyopia in infants with esotropia. Doc. Ophth., 30, 210-216.
Jacobson, S.; Mohindra, I.; and Held, R. (1983). Monocular visual

form deprivation in human infants. Doc. Ophth., 22· 199-211.
Kiorpes, L.; Boothe,R.; Carlson, M.; and Alfi, D. (1985). Frequency

of naturally occurring strabismus in monkeys. J. Pediat. Ophthal. &
Strabismus, ~. 60-64.
LeVay, S.; Wiesel, T.; and Rubel, D. (1980). The development of

ocular dominance columns in normal and visually deprived monkeys.
J. Comp. Neurol. 191, 1-51.
Levi, D. and Klein, S. (1982). Differences in vernier discrimination

for gratings between strabismic and anisometropic amblyopes. Inv.
Ophth. and Visual Sci., ll• 398-407.
Manny, R. and Klein, S. (1984). The development of vernier acuity in

infants. Current Eye Research, l• 453-62.
Mayer, D. and Dobson, V. (1980). Assessment of vision ia young child-

ren: a new operant approach yields estimates of acuity. Inv. Ophth.
and Visual Sci., 12_, 566-570.
Mohindra, I.; Jacobson, S.; Zwaan, J.; and Held, R. (1983). Psycho-
physical assessment of visual acuity in infants with visual disor-
ders. Behavioral Brain Research, ~, 51-58.
Norcia, A. and Tyler, C. (1985). Spatial frequency sweep VEP: visual

acuity during the first year of life. Vision Research, 25, 1399-1408.
Petrig, B.; Julesz, B.; Kropfl, W.; Baumgartner, G.; and Anliker, M.
(1981). Development of stereopsis and cortical binocularity in human
infants: electrophysiological evidence. Science, 213, 1402-1404.
Powers, M. and Dobson, V. (1982). Effect of focus on visual acuity

of human infants. Vision Research, ~, 521-528.
Shimojo, S.; Bauer, J.; O'Connell, K.; and Held, R. (1986). Pre-
stereoptic binocular vision in infants. Vision Research, ~, 501-510.
Shil'lojo, S.; Birch, E.; Gwiazda, J.; and Held, R. (1984). Develop-
ment of vernier acuity in infants. Vision Research, ~, 721-728.
Shimojo, S. and Held, R. (1987). Vernier acuity is less than grating

acuity in 2 and 3-month olds. Vision Research, ~, 77-86.
Slater, A. and Findlay, J. (1975). Binocular fixation in the newborn

baby. J. of Exp. Child Psychology, 20, 248-273.
Sokol, S.; Hansen, V.; Moskowitz, A.; Greenfield, P.; and Towle, V.
(1983). Evoked potential and preferential looking estimates of vis-
ual acuity in pediatric patients. Ophthalmology, 90, 552-562.
Westheimer, G. (1979). The spatial sense of the eye. Inv. Ophth.

and Visual Sci., 18, 893-912.
Wilson, H. (In Press). Development of spatia-temporal mechanisms in

the human infant. Vision Research;
Yuodelis, C. and Hendrickson, A. (1986). A qualitative and quantita-

tive analysis of the human fovea during development. Vision
Research, 26, 847-855.
21
NEURONAL MECHANISMS OF DEPRIVATION
AMBLYOPIA
WOLF SINGER
INTRODUCTION
Activity-dependent changes of neuronal connectivity are an
ubiquitous phenomenon during early ontogeny and a constituent
factor in the self-organization of the brain. During postnatal
development, these shaping processes become increasingly subtle and
dependent on a growing number of variables including sensory
experience.
In the following paragraph I shall concentrate on
activity-dependent modifications of cortical connections that are
required for the development of binocular receptive fields.
However, similar processes are likely to promote the sharpening of
orientation selectivity and the generation of selectively coupled
neuron ensembles subserving stereopsis and higher visual functions.
Before visual experience becomes effective in influencing the
ocularity of cortical neurons, most cells in the visual cortex of
cats and monkeys respond to stimulation of both eyes. This
condition is maintained with normal experience but also with dark
rearing. With monocular deprivation, however, the large majority of
cortical cells lose the ability to respond to the deprived eye
(Blakemore et al., 1978; Wiesel and Hubel, 1965a). The cells which
continue to respond to the deprived eye are located preferentially
within layer IV in the center of the shrunken termination fields of
the deprived eye (Shatz and Stryker, 1978). This suggests that only
those cells continue to respond to the deprived eye which have been
monocularly driven by this eye right from the beginning. With
reverse suture, if it occurs early enough, the previously closed
eye may recover control over most cortical neurons while the
previously open eye in turn becomes ineffective (Blakemore et
al., 1978; Wiesel and Hubel, 1965b).
259
260 W. SINGER
THE EFFECTS OF MONOCULAR OCCULSION ON INDIVIDUAL CORTICAL NEURONS

So far the changes resulting from monocular deprivation have
never been followed in individual neurons. All interpretations and
conclusions are based on the comparison of independently analyzed
samples of neurons which in most cases have been recorded from
different animals that were raised under different conditions. We
have, therefore, developed a method for chronic recording which
allowed us to determine repeatedly the receptive field properties
of cortical neurons in alert and only lightly restrained kittens.
This made it possible to study the effects of monocular deprivation
and reverse occlusion in indivisual cortical cells (Mioche and
Singer, in prep.).
The first effect of monocular deprivation was the
disappearance of binocular summation, i.e. the uperiority of
binocular over monocular responses. Overt changes of ocular
dominance have been observed as soon as 6 h after the beginning of
monocular deprivation and a complete loss of excitatory responses
to deprived eye stimulation was seen as early as 12 h after
occlusion. These changes in ocularity resulted mainly from a
gradual decrease of the excitatory response to deprived eye
stimulation and were only occasionally associated with a moderate
increase of responses to the normal eye. During this shift neurons
preserved their orientation and direction selectivity.
Interestingly, stimulation of the deprived eye continued to evoke
inhibitory responses even after excitatory responses had vanished
completely. The interval required for the manifestation of these
alterations appeared to be rather independent of the duration of
preceding visual exposure suggesting that the duration of visual
stimulation required for the induction of a change is shorter than
the time needed for subsequent expression of the change. Neurons
that had already been monocular prior to deprivation were found
resistant to monocular deprivation and showed neither changes in
responsiveness nor in ocularity.
The recovery of responses to the previously deprived eye in
the case of reverse occlusion had a considerably slower time
course: at least 24 h of reverse occlusion were required before the
responses to the deprived eye reappeared and growth of these
responses continued for up to 74 h. The first change after revers a I
was a reduction of the response to the newly deprived eye. In
numerous neurons responses to the newly open eye reappeared only
after the cell had ceased to be excitable from the previously open
eye. Thus, these neurons passed through a phase during which they
were entirely unresponsive to visual stimulation. In the remaining
cells responses to the newly open eye reappeared before responses
to the now deprived eye had disappeared. During this reversal of
ocular dominance all cells preserved their original orientation
preference, but there were indications for a reduction of
orientation tuning.
MONOCULAR DEPRIVATION 261
The results of the monocular deprivation experiments suggest

several conclusions: First, the down-regulation of the deprived
afferents £annot be considered as the result of a homeostatic
response to the up-regulation of inputs from the viewing eye. Our
results rather favor a mechanism which leads to primary
inactivation of deprived inputs. This may either be due to
weakening of excitatory or to strengthening of inhibitory inputs.
The observation that inhibitory responses persisted once the
excitatory response had disappeared is compatible with the latter
interpretation. Compatible is further the finding that blockade of
GABAergic inhibition reverses the early effects of monocular
deprivation (Sillito et al., 1983). Finally, evidence is now
available that the efficiency of GABAergic inhibition can increase
in a use-dependent way in visual cortex slices (Artola and Singer,
1987). However, the alternative that inhibitory connections remain
unmodified while excitatory transmission is weakened has to be
considered as equally liable.
The second conclusion is that the time required for the
manifestation of changes in ocular dominance does not depend solely
on the amount of visual stimulation but seems to be determined by
the time constants of a process that is independent of experience.
This has a bearing on the controversial discussion about the
existence of a consolidation period for the expression of ocular
dominance changes. Our data favor the notion of the involvement of
two different processes: an early, vision-dependent induction phase
and a subsequent vision-independent mani estation period. The
recent finding of Rauschecker and Hahn (1987) that anaesthesia if
induced within 1 h after monocular exposure prevents the
manifestation of ocular dominance changes also points in this
direction.
Finally, the results of the monocularly deprived kittens
showed that the breakdown of binocular summation precedes the first
signs of an ocular dominance shift. This suggests that binocular
cooperativity is a particularly vulnerable network property. The
earlier reports that short periods of monocular deprivation lead to
a reduction of binocular neurons rather than to a shift of ocular
dominance towards the open eye may be relevant in this context
(Freeman and Olson, 1982).
The results of the reversal experiments support the conclusion
that the reduction of transmission in deprived afferents is not a
direct consequence of increased efficiency of afferents from the
open eye since the former process seemed to precede the latter.
Differential gain changes in deprived and non-deprived afferents do
not appear to result from a normalizing process aimed at keeping
the overall excitatory input to a neuron constant. Rather it
appears as if there are two processes which have different time
constants and interact with each other through some still
unidentified links: one leads to a rapid inactivation of deprived
afferents while the other promotes a slow increase in the efficacy
262 W. SINGER
of the previously inactivated connections. The latter process

appears to take over only once competing excitatory inputs get
weakened.
Our finding that reverse occlusion leads in numerous cells to
a transitory loss of excitatory responses to stimulation of either
eye confirms and explains the observation that brief periods of
reverse occlusion increase the number of visually unresponsive
neurons (Blakemore et al., 1974, 1978). This observation may also
be relevant for clinicians who use occlusion techniques to improve
visual functions in amblyopic eyes of children. Our results explain
why it is so difficult to restore binocular functions in that they
show that the transition from one monocular activation state to the
other occurs only rarely via a transitional state during which the
neuron is excitable from both eyes. The finding that relatively
short periods of occlusion are sufficient to cause substantial
weakening of the efficiency of the previously normal eye should be
taken as a caveat and make frequent determinations of visual
functions imperative during occlusion therapy.
CONDITIONS REQUIRED FOR THE INDUCTION OF OCULAR DOMINANCE CHANGES

The role of postsynaptic activation
It has become clear that changes in ocular dominance do not
solely depend on the level of activity in the afferent connections
(Singer et al., 1977; Greuel et al., 1987; Cynader and Mitchell,
1977). It rather appears that the critical variable is the state of
activation of the postsynaptic neuron and in particular the degree
of temporal correlation between pre- and postsynaptic activation
(Rauschecker and Singer, 1979, 1981). The use-dependent
modifications of excitatory transmission actually seem to follow
rules which closely resemble those postulated by Hebb (1949), Stent
(1973), and Changeux (1976; 1984) for adaptive neuronal connections
(Rauschecker and Singer, 1981). The basic assumption is that the
direction of the change - increase or decrease of efficacy -
depends on the correlation between pre- and postsynaptic
activation. The efficacy of excitatory transmission appears to
increase when presynaptic afferents and the postsynaptic cell are
active in temporal contiguity and to decrease when the postsynaptic
target is active while the presynaptic terminal is silent. These
rules, when applied to circuits where two (or more) afferent
pathways converge onto a common postsynaptic target cell, have the
effect to selectively stabilize and hence associate pathways that
convey correlated activity. Likewise, these selection criteria lead
to competition between converging pathways if these convey
uncorrelated activity. In that case there is always one subset of
afferents inactive while the other is driving the postsynaptic
cell. Hence one subset will increase its gain on the expense of the
respective other. Eventually, those afferents will win which have
the highest probability of being active in temporal contiguity with
the postsynaptic target cell.
The duration of the integration interval for Hebbian matching

The modification rules predict that pathways consolidate only
when they are active in contingency with the postsynaptic cell. For
converging pathways, as e.g. those arriving from the two eyes, this
implies that they have to be active simultaneously if they are to
remain connected to a common target cell. As soon as they get out
of phase they will compete with each other and one pathway will
become repressed. The cut-off point beyond which asynchrony between
the activity patterns from the two eyes leads to disruption of
binocularity has now been determined using high speed solid state
shutters which allowed for rapidly alternating monocular occlusion
in freely behaving kittens (Altmann et al., 1987). The maximal
interval of asynchrony still compatible with the maintenance of
binocular c nnections was found to be in the order of 200 to
400 msec. This agrees with the psychophysical evidence that the
neuronal representation of a stimulus persists for 300 to 400 msec
and that binocular signals can be integrated over such long time
spans. This implies that the signal which is generated by one eye
persists, presumably in the cortical cells, for a few hundred msec.
The neuronal processes subserving persistence have not yet been
identified. Both, long-lasting postsynaptic conductance changes and
reverberation in reciprocally coupled neuron ensembles are likely
candidates. Of importance in the present context is that the
evaluation of binocular synchronicity cannot be based on the
contiguity of individual action potentials. This notion may be
important for clinical measures aimed at rehabilitating binocular
cooperativity. It implies that minor asynchronies in binocular
activation are presumably tolerable during training with dichoptic
stimulation paradigms.
The role of non-retinal signals in ocular dominance plasticity

The evidence reviewed above suggests that the activation of
postsynaptic target cells is a necessary prerequisit for the
induction of experience-dependent modifications. However, it has
now become clear that the generation of action potentials in the
postsynaptic cell is not sufficient to produce a change. Even when
contour vision is unrestricted and retinal signals readily elicit
responses in the neurons of the visual cortex, vision-dependent
modifications of excitatory transmission fail to occur in a variety
of rather different conditions. Thus, neurons of the kitten's
striate cortex may remain binocular despite monocular deprivation
when the open eye is surgically rotated within the orbit (Singer et
al., 1982), when large angle squint is induced in both eyes (Singer
et al., 1979) or when strabismus is induced by bilateral
cyclotorsion (Crewther et a l., 1980). In these cases contour vision
per se is unimpaired but the abnormal eye position and motility
264 W. SINGER
lead to massive disturbances of the kittens' visuo-motor

coordination. Initially the inappropriate retinal signals cause
abnormal visuo-motor reactions and during this period are effective
in influencing cortical ocular dominance. Subsequently, however,
the kittens rely less and less on visual cues and develop a near
complete neglect of the visual modality. In this phase, retinal
signals no longer modify ocular dominance and they also fail to
support the development of orientation-selective receptive fields.
These latter results suggest that retinal signals only
influence the development of cortical functions when the animal
uses them for the control of behavior. Thus, retinal signals do not
lead to changes of cortical functions when the kittens are
paralyzed and/or anaesthetized while exposed to visual patterns
(Buisseret et a!., 1978; Freeman and Bonds, 1979; Singer, 1979;
Singer and Rauschecker, 1982). Vision-dependent modifications also
failed to occur in kittens in which a sensory hemineglect was
induced by unilateral lesions of the intralaminar nuclear complex
of the thalamus (Singer, 1982). Following monocular deprivation
neurons in the visual cortex of the normal hemisphere had become
monocular as is usual with monocular deprivation. However, neurons
in the visual cortex of the hemisphere with the lesion had remained
binocular. Thus, although both hemispheres had received identical
signals from the open eye, these signals induced the expected
modifications only in the normal hemisphere and remained rather
ineffective in the hemisphere which - because of the diencephalic
lesion - "attended" Jess to retinal stimulation.
The chemical nature of permissive gating signals.

Kasamatsu and Pettigrew (1979) had shown that neurons of the
kitten striate cortex remain binocular despite monocular
deprivation when cortical norepinephrine (NE) is depleted by local
infusion of the neurotoxin 6-hydroxydopamine (6-0HDA). Since they
could demonstrate that microperfusion of the depleted cortical
tissue with NE reinstalls ocular dominance plasticity (Kasamatsu et
a!., 1979), these authors have proposed that normal NE levels are a
necessary prerequisite for ocular dominance plasticity.
Subsequently, however, several independent investigations have
indicated that ocular dominance changes can be induced despite NE
depletion. In these investigations 6-0HDA was either injected prior
to monocular deprivation or the noradrenergic input to cortex was
blocked by other means than local 6-0HDA application (Adrien et
a!., 1986; Bear and Daniels, 1983; Bear et a!., 1983; Daw et
a!., 1984; 1985). This apparent controversy may now have been
resolved by the demonstration that ocular dominance plasticity is
influenced both by noradrenergic and cholinergic mechanisms (Bear
and Singer, 1986). Ocular dominance;Dlasticity is abolished only if
both the noradrenergic pathway from locus coeruleus and the
cholinergic jection from the basal forebrain are lesioned.
Disruption of either system alone is not sufficient to arrest
plasticity. These results suggest that both neuromodulators

acetylcholine and norepinephrine have a permissive function in
cortical plasticity, either system being capable of substituting
the other.
Involvement of NMDA-receptor activation and Ca++-fluxes

in use-dependent plasticity
The lesion and stimulation experiments reported above
indicated that use-dependent modifications of synaptic transmission
require a certain amount of cooperativity between retinal input and
additional, internally generated signals. This suggested that the
process mediating cortical plasticity has a threshold which is
reached only when retinal signals are coincident with additional
facilitatory input. As a possible substrate for such a threshold
process the high threshold voltage-dependent Ca++-channels were
considered. Their voltage threshold is higher than that of the
Na+-channels which mediate the r generative action potential and
they appear to be located in dendritic rather than in somatic
membranes (Llinas, 1979). Moreover, there is evidence from
developing Purkinje cells and from binding studies in the kitten
visual cortex (Bode-Greuel and Singer, 1987) that these channels
are particularly numerous in the membranes of developing neurons
(Llinas and Sugimori, 1979). An attempt was made, therefore, to
determine with ion-selective electrodes whether stimuli known to
modify cortical response properties also led to an activation of
Ca++-channels (Geiger and Singer 1986).
The results revealed a close correlation between the
occurrence of changes in neuronal response properties and the
activation of Ca++-fluxes from extra- to intracellular
compartments.
The hypothesis that Ca++-currents are involved in
activity-dependent long-term changes of neuronal excitability
receives further support from the recent finding that the
activation of the NMDA-receptor, a subtype of the receptors for
excitatory amino acids, is also crucial for the induction of ocular
dominance changes (Kleinschmidt, Bear, Singer, 1987). Infusion of
APV, a selective antagonist of the NMDA-receptor blocks completely
any experience-dependent modifications of striate cortex functions
without, however, abolishing neuronal responses to light. The
NMDA-receptor is activated by excitatory amino acids and has a
selective affinity for the synthetic agonist N-methyl-0-aspartate.
It is coupled to a ionophore which is permeable for Ca++-ions. A
special property of this NMDA-sensitive channel is its voltage
dependence. It is activated only when the receptor is occupied by
its endogeneous ligand and when the membrane is sufficiently
depolarized. Thus, this mechanism is ideally suited to "evaluate"
contingency of pre- and postsynaptic activation. It contributes to
the cells' depolarization only if there is sufficient cooperativity
266 W. SINGER
between converging afferents. Hence the NMDA-mechani sm could be the

substrate for the threshold that needs to be trespassed in order to
obtain long-term modifications of neuronal excitability.
Experiments in slices of the visual cortex seem to confirm
this interpretation. They show that the conventional single spike
responses to orthodromic stimulation of the optic radiation do not
depend on the activation of the NMDA-mechani sm. However,
NMDA-channels seem to be involved in triggering the high frequency
bursts that may also be evoked by orthodromic stimulation,
especially when stimuli are applied at higher frequency. These
burst discharges are likely to depend on the activation of
Ca++-conductances and are blocked by application of the
NMDA-receptor blocker APV. In a substantial fraction of cortical
cells prolonged application of burst generating stimuli induces a
long-term enhancement of excitatory transmission that is similar to
long-term potentiation in the hippocampus. When the NMDA-receptors
are blocked by APV these same stimuli no longer generate burst
discharges and also no longer induce changes in synaptic
transmission (Artola and Singer, 1987). Thus, the activation of the
NMDA-dependent EPSP and the resulting burst discharge appear to be
necessary conditions for the induction of long-term modifications.
Sine both phenomena are most likely associated with postsynaptic
CA++-fluxes through NMDA-dependent ionophores and/or high threshold
Ca++-channels these results support the hypothesis that
activity-dependent modifications require a surge of intracellular
Ca++-ions. As APV also blocked experience-dependent modifications
of cortical response properties in vivo it is likely that these
adaptive processes are based on similar mechanisms as those
observed in vitro.
Supporting evidence for this possibility comes from
experiments in which attempts have been made to induce lasting
changes of neuronal response properties by combining retinal
stimulation with direct pharmacological modulation of the
excitability of cortical neurons (Greuel et al., 1987). In
anesthetized, paralyzed kittens cortical cells were activated from
the retina with light while at the same time they were exposed to
iontophoretically applied pulses of NMDA and the neuromodulators
ACh and NE. In numerous neurons 30 min of conditioning were
sufficient to induce a selective enhancement of the responses to
the light stimulus used for conditioning and to produce a
concomitant depression of responses to nonconditioned stimuli.
Thus, with appropriate light stimulation, drastic changes in ocular
dominance and/or orientation and direction preference could be
obtained that outlasted the end of the conditioning sequence by
more than 30 min. These results support the notion that induction
of use-dependent long-term modifications of response properties
requires cooperativity between retinal input and further,
internally generated, facilitatory influences. Especially the
permissive effects of NMDA are suggestive of an involvement of
Ca++-currents in these processes.
THE FUNCTIONAL SIGNIFICA~CE OF OCULAR DOMINANCE PLASTICITY

It has been proposed that activity-dependent reorganization of
interocular connections is part of the normal developmental process
that serves to select among the many possible combinations of
afferents those subsets which originate from corresponding loci in
the two retinae. This selection has to be based in part on
functional criteria and cannot be achieved with positional markers
alone because it is in principle impossible to anticipate with
sufficient precision which retinal loci in the two eyes will
actually be corresponding in the adult animal. The reason is that
correspondence depends on variables such as interocular distance
and size and shape of the eye balls, parameters which vary during
development and are influenced by non-predictable epigenetic
factors. Afferents arriving from corresponding retinal loci can,
however, be identified on the basis of their activity. When the
animal fixates an object with both eyes, corresponding loci in the
two eyes are exposed to similar luminance distributions. Hence, the
activity patterns in afferents coming from corresponding retinal
loci are likely to be correlated. The two retinotopic maps can thus
be matched by evaluating contingencies using patterns in the
outside world to close the loop. Obviously, the adaptive mechanisms
whose effect is disclosable with monocular deprivation, are ideally
suited to serve such selection processes. They stabilize
selectively afferents conveying correlated activity. For such
selection to be successful, however, it is required that the animal
is actually fixating a non-ambiguous visual target with both eyes.
This necessitates control over eye movements and some evaluation of
the best match of retinal signals in the innate and coarsely
specified binocular connectivity. Circuit selection must not occur,
however, when the eyes are movingor not properly aligned or when
the visual signals are ambiguous. Circuit selection must also not
occur in response to self-generated patterned activity such as
occurs during certain sleep stages. It appears therefore
well-adapted that ocular dominance plasticity is gated by
proprioceptive signals from extraocular muscles and by central core
projections which modulate cortical functions in a state-dependent
way.
The adaptive mechanisms disclosed by monocular deprivation and
the resulting neuronal changes can be related directly to the
pathophysiology of deprivation amblyopia. Most likely,
however,theye are relevant also for other developmental
disturbances of binocular functions. No further assumptions need to
be made to account for the disruption of binocular cooperativity in
the (ase of alternating strabismus. Due to the squint the activity
pattern conveyed by the afferents from the two eyes show no
systematic correlation. Hence, afferents from one eye stabilize
always on the expense of the respective afferents from the other
268 W. SINGER
eye and cortical cells become monocular.

Only one additional mechanism has to be postulated to account
for the development of strabismic amblyopia. One has to assume that
central selection processes introduce a consistent bias in favor of
one eye. As discussed above, retinal signals are only capable of
promoting experience-dependent shaping of neuronal connectivity if
they are attended to and used for the control of behavior. If, as
is the case with the development of amblyopia, the brain attends
consistently only to activity from one eye, only the neuronal
chains connected to this eye will mature properly. The neuronal
ensembles driven by the other eye will not only fail to profit from
experience-dependent pruning of connectivity, but they will
predictably be affected also by active competition that is exerted
by the good eye via inhibitory interactions. The mechanisms of
input selection could be the same which normally serve to suppress
activity from one eye to avoid diplopia of objects located outside
Pannum's area. lnterocular inhibitory pathways exist both at the
thalamic and cortical level. Recent electrophysiological
experiments in adult cats have demonstrated that these can cause a
substantial suppression of visual responses already in the lateral
geniculate when the two eyes are stimulated with dissimilar
patterns that cannot be fused (Varela and Singer, 1987). Moreover,
there is evidence from slice experiments that GABAergic inhibition
antagonizes very effectively the NMDA receptor dependent EPSP
component in cortical neurons (Artola and Singer, 1987). Thus,
inhibition interferes in a very potent way with a mechanism, whose
activation appears to be indispensable for use-dependent pruning of
neuronal connectivity. Thus, activity conveyed by the non-attended
eye would be prevented at several levels from promoting the circuit
selection that appears necessary for the development of normal
cortical functions. Because selective attention is not directed to
the signals arriving from this eye there will be no enhancement of
corresponding responses. This will lower the probability that the
activation of neurons connected to this eye reaches the threshold
that needs to be trespassed for the induction of
experience-dependent modifications. The probability of reaching
plasticity threshold will be lowered further by inhibitory
interaction both at the thalamic and the cortical level which
further attenuates responses to the non-attended eye.
Support for such a pathophysiological mechanism of strabismic
amblyopia comes from psychophysical experiments. Sireteanu and her
coworkers have shown that strabismic amblyopia develops only in
those parts of the visual field where fusion of binocular signals
is not possible (Sireteanu and Fronius, 1981). This suggests local
interactions and competition between the activity patterns arriving
from the two eyes. Furthermore, it was possible to demonstrate that
interocular inhibition is particularly pronounced and to the
disadvantage of the amblyopic eye in regions of the visual field in
which the suppressed eye developed amblyopic deficits (Sireteanu et
a I., 1981).
The available neurophysiological data on activity-dependent

selective stabilization of binocular connections have some bearing
also on the development of anomalous correspondence. Within the
range of overlap between the afferents from the two eyes, that is
provided by the genetically prespecified Anlage, any correspondence
can in principle be established by contingency matching and
selective stabilization of afferents that convey correlated
activity. Obviously, when squint angles are large and exceed the
relatively small range of tolerance that is provided by overlap in
area 17, binocular integration can no longer be achieved in this
area and anomalous correspondence and fusion must occur in other
cortical areas. In some prestriate cortical areas receptive fields
are very large and retinotopic maps are less precisely arranged.
Thus, in these areas contingency matching could lead to anomalous
correspondence even with very large squint angles. Since these
areas subserve specific functions it would even appear conceivable
that anomalous correspondence and fusion occur not at all instances
but only when visual functions are executed which involve
prestriate areas with particularly large receptive fields. Such
would be the case e.g. for the processing of motion signals. This
predicts that in strabismic amblyopia certain binocular functions
might be preserved even if there is a complete loss of those
binocular functions that require binocular integration in primary
visual cortex.
We have been able to demonstrate that such is indeed the case.
The ability to compute from the comparison of movement vectors in
the two eyes the trajectories of objects moving in space is
preserved in most strabismic amblyopes even if conventional
stereopsis is completely abolished (Sireteanu et al., 1981).
Patients, by contrast, who have overcome the problem of diplopia by
developing alternating fixation have lost this ability. In that
case, binocular integration is obviously lost also in non-striate
visual areas. Thus, there appear to be two strategies to cope with
diplopia. One is to fixate with one eye only and to suppress
selectively the input from the other eye that cannot be matched.
With moderate esotropia this seems to disrupt binocular functions
only in some areas of visual cortex including striate cortex.In
that case conventional stereopsis is lost and amblyopia develops
for the central part of the visual field in the suppressed eye.
However, the ability to use dichoptic cues for movement analysis is
preserved. The second strategy is to develop alternating fixation
and to suppress globally all signals from the non-attended eye.
This disrupts all binocular functions but preserves normal
monocular vision in both eyes. The ability to analyze movement
trajectories is an important visual function but the price for its
preservation seems to be monocular amblyopia. Since the
ophthalmologist has the option to favor with his therapeutical
measures the development of alternating fixation this raises a
difficult ethical question: should one preserve normal acuity in
both eyes but sacrifice all binocular functions or should one let
one eye become amblyopic and preserve binocular functions in
270 W. SINGER
non-striate cortical areas? Clearly, before reaching a consensus on

this question, measurements of the ability to perceive motion in
depth must become a standard test in clinical ophthalmology and the
behavioral relevance of this function needs to be evaluated more
thoroughly.
If one accepts to generalize the neuronal mechanisms which
mediate the effects of monocular deprivation in cats to the various
forms of de elopmental disturbances of binocular functions in
humans, a few suggestions can be derived with respect to
rehabilitation strategies. To strengthen inactivated connections
the activation threshold has to be reached beyond which
consolidation and pruning processes become possible. This requires
first that central gating systems are active, i.e. the patient
needs to be aroused, very attentive and motivated to use the visual
modality. Thus, training paradigms should be designed which force
the patient to identify visual stimuli and to use these for the
control of behavior. If applied to children the visual stimuli
ought to be integrated as part of interactively structured video
games containing an element of competition. A second prerequisite
is that the stimulus provides maximal excitatory drive for a
maximal number of cortical neurons. This requires contoured stimuli
with high contrast. To optimize the activation of individual
orientation columns, the stimulus should contain a broad spectrum
of spatial frequencies and should be moving. Only one orientation
should be present at a time at any given location to minimize
intercolumnar inhibition. All orientations should, however, appear
sequentially at any location. Evoked potential responses to phase
reversing stimuli could perhaps help to individually adjust these
stimulus parameters to the respective functional state of the
patient's visual system. Since monocular patching often weakens the
input from the healthy eye before it enhances the responses to the
amblyopic eye it needs to be applied with as much caution as
possible. Repeated measurements of visual functions (acuity,
contrast sensitivity, binocular motion in depth perception) ought
to be made to detect as early as possible signs of impairment in
the normal eye and to determine the optimal period for binocular
stimulation. It might be worth trying to reduce the competitive
advantage of the good eye with gray filters and to resume dichoptic
training as soon as some residual functions can be recovered from
the amblyopic eye.
Clearly, more animal experiments are necessary that are
devoted directly to the problem of rehabilitation of binocular
functions. The increasing knowledge about the developmental
mechanisms and the improved techniques to record neuronal changes
as a function of experience will undoubtedly improve the design of
such experiments.
REFERENCES
Altmann, L., Luhmann, H.J., Greuel, J.M. and Singer, W. (1987)
Functional and neuronal binocularity in kittens raised with rapidly
alternating monocular occlusion. J. Neurophysiol., in print.
Artola, A. and Singer, W. (1987) Long-term potentiation and NMDA
receptors in rat visual cortex. Nature, submitted.
Bear, M.F. and Daniels, I.D. (1983) The plastic response to
monocular deprivation persists in kitten visual cortex after
chronic depletion of norepinephrine. J. Neurosci. l· 407-416.
Bear, M.F. and Singer, W.(1986) Modulation of visual cortical
plasticity by acetylcholine and noradrenaline. Nature 320, 172-176.
Blakemore, C. and Van Sluyters, R.C. (1974) Reversal of the
physiological effects of monocular deprivation in kittens:
Further evidence for a sensitive period. J. Physiol. 237, 195-216.
Blakemore, C., Garey, L.J. and Vital-Durand, F. (1978) The
physiological effects of monocular deprivation and their reversal
in the monkey's visual cortex. J. Physiol. 238, 223-262.
Bode-Greuel, K.M. an~ Singer, W. (1987) The distribution of binding
sites for organic Ca +-channel blockers in the visual cortex of
kittens and adult cats. Exp. Brain Res., submitted.
Buisseret, P., Gary-Bobo, E. and Imbert, M. (1978) Ocular motility
and recovery of orientational properties of visual cortical
neurones in dark-reared kittens. Nature 272, 816-817.
Changeux, J.P. and Danchin, A. (1976) Selective stabilization of
developing synapse as a mechanism for the specification of neuronal
networks. Nature 264, 705-712.
Changeux, J.P., Heidmann, T. and Patter, P. (1984) Learning by
selection. In The Biology of Learning, Dahlem workshop. (eds. P.
and H.S. Terrace) Springer-Verlag, Berlin, pp. 115-133.
Crewther, S.G., Crewther, D.P., Peck, C.K. and Pettigrew, J.D.
(1980) Visual cortical effects of rearing cats with monocular or
binocular cyclotorsion. J. Neurophysiol. 44, 97-118.
Cynader, M. and Mitchell, D.E. (1977) Monocular astigmatism effects
on kitten visual cortex development. Nature 270, 177-178.
Daw, N.W., Robertson, T.W., Rader, R.K., Videen, T.O. and Cosica,
C.J. (1984) Substantial reduction of noradrenaline by lesions of
adrenergic pathway does not prevent effects of monocular
deprivation. J. Neurosci. i· 1354-1360.
272 W. SINGER
Daw, N.W., Videen, T.O., Parkinson, D. and Rader, R.K. (1985} DSP-4
depletes noradrenaline in kitten visual cortex without altering the
effects of monocular deprivation. J. Neurosci. ~. 1925-1933.
Freeman, R.D. and Bonds, A.B. (1979) Cortical plasticity in
monocularly deprived immobilized kittens depends on eye movement.
Science 206, 1093-1095.
Freeman, R.D. and Olson, C. (1982) Brief periods of monocular
deprivation in kittens: Effects of delay prior to physiologcial
study. J. Neurophysiol. 47, 139-150.
Geiger, H. and Singer, W. (1986) A possible role of calcium
currents in developmental plasticity. Exp. Brain Res. Ser. 1!·
256-270.
Greuel, J.M., Luhmann, H.J. and Singer, W. (1987) Persistent
changes of single cell responses in kitten striate cortex produced
by pairing sensory stimulation with ionophoretic application of
neurotransmitters and neuromodulators. In Cellular Mechanisms of
Conditioning and Behavioral Plasticity. (eds. C. Woody et al.)
Plenum Press, in print.
Hebb, D.O. (1949) The organization of behavior. John Wiley &Sons,
New York.
Kasamatsu, T. and Pettigrew, J.D. (1979) Preservation of
binocularity after monocular deprivation in the striate cortex of
kittens treated with 6-hydroxydopamine. J. Camp. Neurol. 185,
163-162.
Kasamatsu, T., Pettigrew, J.D. and Ary, M.-L. (1979) Restoration of
visual cortical plasticity by local microperfusion of
norepinephrine. J. Comp. Neurol. 185, 163-182.
Kleinschmidt, M., Bear, M.F. and Singer, W. (1987) Evidence for an
involvement of NMDA receptor mechanisms in experience-dependent
plasticity of kitten striate cortex. Science, in print.
Llinas, R. (1979) The role of calcium in neuronal function. In The
Neurosciences Fourth Stud Pro ram, 4th. (eds. F. 0. Schmitt ana
F.G. Worden. MIT Press, Cambridge, MA, pp. 555-571.
Llinas, R. and Sugimori, M. (1979) Calcium conductances in Purkinje
cell dendrites: their role in development and integration. Progr.
Brain Res. ~. 323-334.
Mioche, L. and Singer, W. (1987) Long-term recordings and receptive
field measurements from single units of the visual cortex of awake
unrestrained kittens. in prep.
Rauschecker, J.P. and Hahn, S. (1987) Ketamine-xylazine anaesthesia
blocks consolidation of ocular dominance changes in kitten visual

cortex. Nature 326, 183-185.
Rauschecker, J.P. and Singer, W. (1979) Changes in the circuitry of
the kitten's visual cortex are gated by postsynaptic activity.
Nature 280, 58-60.
Rauschecker, J.P. and Singer, W. (1981) The effects of early visual
experience on the eat's visual cortex and their possible
explanation by Hebb synapses. J. Physiol. 310, 215-239.
Shatz, C.J. and Stryker, M.P. (1978) Ocular dominance in layer IV
of the eat's visual cortex and the effects of monocular
deprivation. J. Physiol. 281, 267-283.
Sillito, A.M., Kemp, J.A. and Blakemore, C. (1983) The role of
GABAergic ihibition in the cortical effects of monocular
deprivation. Nature 291, 318-322.
Singer, W. (1979) Central core control of visual cortex functions.
In The Neurosciences Fourth Stud Pro ram, 4th. (eds. F.O. Schmitt
and F.G. Worden , MIT Press, Cambri ge, MA, pp. 1093-1109.
Singer, W. (1982) Central core control of developmental plasticity
in the kitten visual cortex: I. Diencephalic lesions. Exp. Brain
Res. 47, 209-222.
Singer, W. and Rauschecker, J. (1982) Central core control of
developmental plasticity in the kitten visual cortex: II.
Electrical activation of mesencephalic and diencephalic
projections. Exp. Brain Res. 47, 223-233.
Singer, W., Rauschecker, J. and Werth, R. (1977) The effect of
monocular exposure to temporal contrasts on ocular dominance in
kittens. Brain Res. 134, 568-572.
Singer, W., Tretter, F. and Yinon, U. (1982) Central gating of
developmental plasticitty in kitten visual cortex. J. Physiol. 324,
221-237. -
Singer, W., von GrOnau, M. and Rauschecker, J. (1979) Requirements
for the disruption of binocularity in the visual cortex of
strabismic kittens. Brain Res. lZl· 536-540.
Sireteanu, R. and Fronius, M. (1981) Naso-temporal asymmetries in
human amblyopia: consequence of long-term interocular suppression.
Vis. Res. ~. 1055-1063.
Sireteanu, R., Fronius, M. and Singer, W. (1981) Residual binocular
interaction in the peripheral field of squint and anisometropic
amblyopes. Doc. Ophthal. Proc. Ser. 30, 224-229.
274 W. SINGER
Stent, G.S. (1973) A physiological mechanism for Hebb's postulate

of learning. Proc. Nat. Acad. Sci. ZQ, 997-1001.
Varela, F.J. and Singer, W. (1987) Neuronal dynamics in the visual
corticothalamic pathway revealed through binocular rivalry. Exp.
Brain Res. 66, 10-20.
Wiesel, T.N. and Hubel, D.H. (1965a) Comparison of the effects of
unilateral and bilateral eye closure on cortical unit responses in
kittens. J. Neurophysiol. 28, 1029-1040.
Wiesel, T.N. and Hubel, D.H. (1965b) Extent of recovery from the
effects of visual deprivation in kittens. J. Neurophysiol. 28,
1060-1072.
22
ABNORMAL BINOCULAR INTERACTION:
EVIDENCE IN HUMANS
GUNTER K. von NOORDEN
The current thinking on the role of binocular interaction

in the development of vision is based on experimental data from
animal models and may be summarized as follows:
The normal development of visual pathways depends not only
on the amount of afferent impulse activity but also on the
interrelationship of this activity. For instance, afferent
genicula-cortical paths are competing during development for
control over cortical synapses. A reduction in efficiency in one
set of synapses permits the other set to take over at the
expense of the first set. Cell bodies losing out in this
competition undergo retrograde atrophy which has been studied
extensively in the lateral geniculate nucleus (LGN).
Competitive interaction in the afferent visual system can
occur only wherever such interaction between the information
receive from the two eyes is anatomically possible. This
excludes the retina and the monocular segment of each LGN which
receives exclusively from the nasal retina of the contralateral
eye, corresponding to the temporal crescent of the visual field.
The disuse effect of visual deprivation causes LGN cells in the
monocular segment corresponding to a sutured eye to be smaller
than in the non-deprived monocular segment (von Noorden &
Middleditch 1975) (von Noorden, Crawford &Middleditch 1976)
(Sloper, Headon & Powell 1986) (Headon, Sloper & Powell 1987).
However, no difference in cell sizes exists between the
monocular segments of esotropic monkeys with strabismic
amblyopia (von Noorden & Middleditch 1975). We have interpreted
this observation as being supportive of the view that abnormal
Supported by grants EY 01120, EY 07001 and EY 02520 from

the National Institutes of Health
275
binocular interaction and competition play a predominant and

disuse from visual deprivation an insignificant role in the
etiology of strabismic amblyopia (von Noorden 1985).
With the competitive nature of factors regulating or
deregulating the growth of visual connections in experimental
animals well established, attention must be drawn to other modes
of binocular interaction and competition that are capable of
modifying the function of one eye or interfere with normal
binocular vision, temporarily or permanently, without
necessarily changing the structure of the afferent visual
pathways. These dynamic aspects of abnormal binocular
interaction have thus far received little or no attention from
those working on animal models. Yet, binocular interaction plays
a predominate role in the production and maintenance of
suppression and of amblyopia and remains effective long after
maturation of the afferent visual pathways. The purpose of this
paper is to discuss these dynamic phenomena which play a
fundamental role in the adaptation of the human visual system to
abnormal input.
It has been known since Tschermak (1900) that unequal
visual impressions from both eyes cause binocular interaction,
competition and inhibition of retinal image processing from one
eye. The clinical conditions that trigger abnormal binocular
interaction in humans are strabismus, anisometropia, unilateral
cataracts, corneal opacities or other causes of unfocused images
in one eye. All these conditions share one common denominator,
namely the presentation of the visual centers in the brain with
unequal excitation from both eyes (Jampolsky, 1978). In the
case of strabismus, the foveae in the two eyes are directed
towards different visual objects. As the foveae have a common
visual direction these visual objects are now perceived by the
patient as being superimposed in subjective visual space which
causes visual confusion. Moreover, the patient becomes aware of
diplopia from stimulation of corresponding retinal points with
different images. Visual inhibition or, as it is known to
clinicians, suppression, or amblyopia may ensue as consequences
of visual confusion or diplopia.
In anisometropia without strabismus the foveae receive
images from the same visual objects. However, the images of the
more myopic, hypermetropic or astigmatic eye are out of focus.
Competition between a sharply focused and a blurred image sets
in which is soon decided in favor of the more sharply focused
image. In addition there is a passive visual deprivation effect
from the unfocused image.
ABNORMAL BINOCULAR INTERACTION 277
Unilateral cataracts, or other opacities of the ocular

media cause retinal image degradation and thus, abnormal
binocular interaction in addition to a passive visual
deprivation effect. The resulting amblyopia is, as a rule, more
severe and resistant to therapy than in the case of strabismic
amblyopia.
Cataracts in both eyes cause no abnormal binocular
interaction, provided the reduced stimulation received by each
eye is balanced. In this instance, the amblyopia is present in
both eyes and is caused exclusively by visual deprivation. A
milder form of amblyopia also caused exclusively by visual
deprivation and present in both eyes occurs in patients with
bilateral uncorrected high hypermetropia.
Inhibition is the cause of two pathological states:
suppression and amblyopia. As we shall see, the difference
between these conditions is merely a quantitative one.
Suppression is limited to strabismus. It is a binocularly
elicited phenomenon, characterized by a temporary decrease of
visual functions in the non-fixating, deviated eye of the
strabismic patient. There are no monocular manifestations of
suppression. Upon changing fixation to the other eye, the
previously suppressed eye regains immediately its full
functional potential and does so within a remarkably short time.
According to Steinbach this trade of roles between the fixating
and suppressed eye takes place in only 80 msec.
(Steinbach, 1981). In strabismic amblyopia the inhibition from
the misaligned eye or in anisometropia, from the visually
deprived eye is carried over into monocular vision, i.e. it
persists upon occluding the fixating eye. Depending on the age
of onset, the duration of the amblyopiogenic condition and the
age at which treatment is begun amblyopia ·is completely or
partially reversible.
Aside from these differences the clinical features of
amblyopia and suppression are remarkably similar. Both
conditions develop only in childhood but may persist through
life. The end of the sensitive period for the development of
suppression and amblyopia is, as a rule, reached with completion
of the eighth year of life. The absence of double vision in
children or adults with an onset of strabismus after that time
is probably caused by a psychological and not a physiological
adaptation; the second image is ignored rather than actively
suppressed.
In suppression and amblyopia the involved eye is not
excluded entirely from vision. Suppression in most forms of
strabismus is "regional", i.e. affecting only parts of the

retina of the deviated eye. This causes two scotomas, one
corresponding to the fovea (central scotoma) and the other one
to that part of the peripheral retina that receives the same
image as the fovea of the fixating eye (fixation point scotoma
of Harms [1937]). In some cases of deep suppression the two
scotomas merge into one. These scotomas are relative, i.e.
retinal function is decreased but not totally abolished. In
strabismic amblyopia the functional defect is limited to the
center of the visual field. In anisometropic and stimulus
deprivation amblyopia, on the other hand, retinal sensitivity is
decreased in the fovea as well as in the periphery (Hess, 1980).
In his masterful thesis Harms presented numerous
observations that establish suppression as a dynamic rather than
as a static form of inhibition by showing that the depth of
suppression and the location of supression scotomas in the
retina undergo constant changes depending on the visual demands
made on the fixating eye (Harms, 1937). Bagolini developed a
filter ladder, containing a series of red filters of increasing
density with which the depth of suppression is determined
(Bagolini, 1957). The density of the filter which, when held
before the fixating eye elicits diplopia indicates the depth of
suppression.
It is less well known but equally well established by
numerous clinical studies that the functions of an amblyopic eye
are dependent in a similar manner upon the functions of the
sound eye. For instance, Pugh found that in many cases of
amblyopia the stimulation of the sound eye resulted in a marked
reduction of visual acuity of the amblyopic eye (Pugh, 1954).
Together with Leffler we investigated this phenomenon and showed
that the degree of inhibition of the amblyopic eye varies
proportional to the luminance of the stimulus reaching the
normal eye (von Noorden &Leffler, 1966). This inhibitory
defect decreased the visual acuity of the amblyopic eye
significantly, in drastic cases from 0.3 to as low as 0.1. No
such interocular defect was present in organic vision loss
caused by retinal damage where the visual acuity remains the
same regardless of what stimulation is received by the sound eye
(Awaya &von Noorden, 1972).
Moran and Gordon reported a peculiar form of binocular
competition in a patient with congenital monocular cataracts
removed at the age of 19 years (Moran &Gordon, 1982). The
visual input of the normal eye caused enormous increases in the
sensitivity threshold of the deprived eye. This is in contrast
to the behavior in normal eyes where the threshold of one eye is

not influenced by visual input to the other eye.
Pigassou and co-workers noted that these inhibitory effects
from the sound eye persist even after visual acuity of the
amblyopic eye is fully recovered by treatment (Pigassou, 1969).
This finding may be expected as long as the conditions giving
rise to amblyopia, such as a manifest ocular deviation or an
anisometropia continue to exist. However, the possibility
exists that the visual system becomes conditioned to binocularly
elicited inhibition during infancy and that this inhibition may
continue to be triggered by any form of binocular activity,
normal or abnormal, long after the input of the two eyes has
been equalized and the amblyogenic conditions have disappeared.
I have observed amblyopic patients who were orthotropic,
emmetropic and had no other discernible cause for their
amblyopia. Their visual acuity improved rapidly after occlusion
of the sound eye and a relapse of amblyopia occurred as soon as
this treatment was discontinued (von Noorden, 1985). Presumably,
the original cause of amblyopia in these patients was
anisometropia in infancy which was no longer present at the time
of the examination due to emmetropization of the more ametropic
eye.
That interocular inhibition in amblyopes reaches down to
basic visual functions was shown convincingly by Mackensen
(1959} and Aulhorn (1967}. Quantitative perimetry performed in
monocular and binocular conditions showed a significant increase
of relative light threshold of the amblyopic eye when both eyes
were open.
Finally, the fact that visual functions of the amblyopic
eye may occasionally recover partially or even completely in
adult patients (Vereeken, 1984) or in monkeys with experimental
strabismus (Harwerth et al, 1986) after loss of function of the
good eye provides additional evidence of the role of the sound
eye in maintaining the amblyopic defect. That such recovery
after enucleation of the sound eye does not occur in monkeys
with stimulus deprivation amblyopia from unilateral lid suture
(Harwerth et al, 1984) could mean that inhibition plays a lesser
role in monocular visual deprivation than in strabismus and that
the effect of lid suture is primarily one of passive disuse.
Indeed, the normalcy of the monocular lateral geniculate nucleus
segment in monkeys with strabismic amblyopia and the cell
shrinkage in the monocular segment of monkeys with lid closure
(von Noorden &Middleditch, 1975} would lend support to this
theory.
All of these findings imply that the amblyopic eye is at

its best when the sound eye is excluded from any form of visual
stimulation. The practical implication of this observation
relates to the therapy of amblyopia. In order to create optimal
conditions for training of the amblyopic eye, intraocular
inhibition must be reduced as much as possible. It becomes
obvious that only total occlusion with an opaque occluding
device is the most effective therapy and preferable over
semitransparent occluding devices, high plus lenses or
atropinization of the sound eye.
The reduction of visual functions of an amblyopic eye
through visual stimulation of the sound eye is also of
theoretical interest. First, it is incompatible with the view
that strabismic amblyopia is simply caused by an arrest of
visual development from habitual exposure to a defocussea-image
of low contrast and blurred detail. Other arguments against this
theory which is still widely held today, include the observation
that the fovea of the deviated eye does receive well-focussed
images and that visual acuity can be normalized in amblyopic
patients with conditions that were clearly present from birth.
If the arrest theory were correct the treated eye would have at
best a visual acuity of 20/600 which is the level of acuity
estimated for newborn children (Held 1987). The fact that
amblyopia can occur in children long after they have reached
adult level visual acuity is also incompatible with the arrest
theory.
Second, the dynamic nature of amblyopia, its rapid
reversibility in children and its occasional reversibility in
adults after loss of the good eye emphasize that there must be
consequences from abnormal binocular interaction other than a
redistribution of geniculo-cortical connections. It is
difficult to blame on neuronal sprouting and re-establishment of
new synaptic contacts the change of amblyopia within a few days
from one eye to the other in young children, the recurrence of
amblyopia in visually mature, older children and the rapid
switch of cortical dominance from one eye to the other after
reverse lid suture in kittens and infant monkeys.
We must also remember that inhibition triggered by abnormal
binocular visual input is not a mechanism especially created for
an abnormal stimulus situation but part of normal binocular
vision. According to Burian, all abnormal responses of
squinters are preformed in the normal process of vision and are
merely exaggerations of it. For example, retinal rivalry from
simultaneous excitation of corresponding retinal points by
dissimilar retinal images causes one or the other piece of
information originating in each eye to be temporarily

suppressed. Thus, suppression and, by extension, amblyopia may
be considered as a loss of the rhythm of binocular rivalry. To
take another example, physiologic diplopia is part of normal
binocular vision. It rarely becomes a clinical problem because
the double images lying in front of or behind the horopter are
ignored or perhaps even suppressed by most subjects. Inhibition
also plays an important role in Panum vision, because stereopsis
is based on the facultative unity of visual directions between a
retinal element in one eye and an element located within an area
surrounding the inhibited corresponding element in the other
eye. There are additional examples for interocular inhibition
in normal binocular vision such as the extinction phenomenon of
Aulhorn (1967) and the modification of monocularly evoked visual
potentials by the type of visual stimulation received by the
fellow eye (Lehmann &Morris, 1967) (Lennerstrand, 1978).
Inhibitory activity clearly is part of normal binocular vision
and is only exaggerated when visual input is unequal and a
competitive situation exists between the two eyes.
Where is the location and what is the nature of this
binocularly elicited inhibition? Clearly, it must originate in
that part of the afferent visual pathway where binocular
interaction becomes anatomically possible. This excludes the
pregeniculate portion. It is debatable whether inhibition can
occur as peripheral in the afferent visual pathways as the
lateral geniculate nucleus. Even though terminals from
corresponding retinal elements are located in adjacent layers we
do not know whether translaminar inhibition occurs in the
primate lateral geniculates. This places the origin of the
inhibitory process in the primary visual cortex even though the
consequences of this process extend as far peripheral as the
retina. We do not know how the brain makes a choice in the case
of strabismus or anisometropia which of the two images
originating from the fovea of the two eyes to accept and which
one to reject. A kind of selection process must take place that
is probably based on attention factors as well as on previous
visual experience. According to Singer (1984) this process must
occur beyond layer IVc, i.e. in the more superficial or deeper
layers of the striate cortex.
It is reasonable to speculate that the neurochemistry of
visual inhibition involves synaptic neurotransmitters. A
synaptic block would be rapidly reversible in cases of
alternating suppression. In amblyopia presumably a similar but
quantitatively different neurochemical process becomes
reversible only after prolonged occlusion of the sound eye, or
reversal does not occur at all. A neurochemical reactivation of
dormant visual connections in amblyopia or a protection of the

visual system against suppression or amblyopia in the presence
of strabismus are no longer entirely utopic therapeutic or
preventive possibilities for the future. However, we must
remember that suppression and amblyopia are also adaptive
mechanisms, "nature's way out of trouble" as Burian once stated
(Burian, 1953), that protect the patient from great visual
discomfort. On the other hand they also represent formidable
obstacles to restoration of normal binocular function in those
patients in whom a potential for cure exists.
Laboratory researchers in experimental animals would do
well in focusing on the mechanism of inhibition in amblyopia and
to also include the study of suppression in animal models for I
predict that a better understanding of suppression may well
provide some day the key that opens the black box of amblyopia.
SENSITIVE PERIODS FOR

UNILATERAL LID CLOSURE
(MACACA MULATTA)
Scotopic sensitivity Birth - 3 m

Photopic sensitivity Birth - 6 m
Form vision Birth - 25 m
Binocular summation Birth - 25 m
Stereopsis Birth - 25 m +
References
1. Aulhorn, E. (1967). Die gegenseitige Beeinflussung
abbildungsgleicher Netzhautstellen bei normalem und
gest8rten Binocularsehen. Doc. Ophthalmol., 23,
26-61. -
2. Awaya, S. and von Noorden, G.K. (1972). Visual
acuity of amblyopic eyes under monocular and
binocular conditions: further observations. In
Orthoptics. (eds. J.J.M. Bierlaagh and T.E.A.
Brummelkamp-Dons). Excerpta Medica, Amsterdam.
3. Bagolini, B. (1957). Presentazione di una sbarra di

filtri a densita scalare assorbenti i raggi
luminosi. Boll. Ocul., 36, 638-651.
4. Burian, H.M. (1953). Adaptive mechanisms. Trans. Am.
Acad. Ophthalmol. Otolaryngol., 57, 131-144.
5. Harms, H. (1937). Ort und Wesen der Bildhemmung bei
Schielenden. Graefes Arch. Ophthalmol., 138,
149-210. -
6. Harwerth, R.S., Smith, E.L., Duncan, G.C., Crawford,
M.L.J., von Noorden, G.K. (1986). Effects of
enucleation of the fixating eye on strabismic
amblyopia in monkeys. Invest. Ophthalmol. Vis. Sci.,
27, 246-254.
7. Harwerth, R.S., Smith, E.L. III, Crawford, M.L.J.,
and von Noorden, G.K. (1984). Effects of enucleation
of the nondeprived eye on stimulus deprivation
amblyopia in monkeys. Invest. Ophthalmol. Vis.
Sci., 25,10-18.
8. Headon-;-M.P., Sloper, J.J., Powell, T.P.S. (1987).
Effects of abnormal visual experience on the morphology
of the lateral geniculate neurons in infant monkeys. In
Strabismus and Amblyopia - experimental basis for advances
clinical management, Wenner-Gren Center Symposium,
Stockholm.
9. Held, R. (1987). Normal visual development and its
deviations. In Strabismus and Amblyopia - experimental
basis for advances in clinical management, Wenner-Gren
Center Symposium, Stockholm.
10. Hess, R.F., Campbell, F.W. and Zimmern, R. (1980).
Differences in the neural basis of human amblyopias:
the effect of mean luminance. Vision Res., 20,
295-305. --
11. Jampolsky, A. (1978). Unequal visual inputs and strabismus
management: a comparison of human and animal strabismus.
In Symposium on Strabismus, Transactions New Orleans
Acad. Ophthalmol. C.V. Mosby, St. Louis.
12. Lehmann, D. and Morris, A. (1967). Monocularly
evoked averaged EEG potentials influenced by
continuous visual input into the other eye. Invest.
Ophthalmol., 6, 547.
13. Lennerstrand,-G. (1978). Binocular interaction
studied with visual evoked responses (VER) in humans
with normal or impaired binocular vision. Acta
Ophthalmol., 56. 628-637.
14. Mackensen, G.-r1959). Monoculare und binoculare
statische Perimetrie zur Untersuchung der
HemmungsvorgMnge beim Schielen. Graefes Arch. Clin.
Exp. Ophthalmol., 160, 573-587.
15. Moran, J. and GordOO, B. (1982.). Longterm visual
deprivation in a human. Vision Res., 22, 27-36.
16. Pigassou, R., Garipuy, J., Corda, M. and Sforzin,

P. (1969). Amblyopie fonctionelle. Ann. Ocul.,
202, 30-46.
17. Pugh, M. (1954}. Foveal vision in amblyopia. Br. J.
18. Singer, W. (1984). Neurophysiologische Grundlagen
der Amblyopie. In: Pathophysiologie des Sehens,
Herzau V editor. (ed. V. Herzau). Bacherei des
Augenarztes, Heft 98, F. Enke, Stuttgart.
19. Sloper, J.J., Headen, M.P., Powell, T.P.S. (1986). Changes
in the size of.cells in the monocular segment of the
primate lateral geniculate nucleus during normal
development and following visual deprivation. Dev. Brain
Res., 31, 267-276.
20. Steinbach, M. (1981). Alternating exotropia;
temporal course of the switch in suppression.
Invest. Ophthalmol. Vis. Sci., 20, 129-133.
21. Tschermak, A. {1900) Ueber physTOlogische und
pathologische Anpassung des Auges. Leipzig; Veit &
Camp.
22. Vereeken, E.P. and Brabant, P. (1984). Prognosis for
vision in amblyopia after loss of the good eye.
Arch. Ophthalmol., 102, 220-224.
23. von Noorden, G.K. (1986). Amblyopia: a multidisciplinary
approach (Proctor Lecture). Invest. Ophthalmol. Vis. Sci.,
26, 1704-1716.
24. von Noorden, G.K. and Middleditch, P.R. (1975}.
Histology of the monkey lateral geniculate nucleus
after unilateral lid closure and experimental
strabismus: further observations. Invest.
Ophthalmol., 14,674-683.
25. von Noorden, G:K., Crawford, M.L.J. and Middleditch,
P.R. (1976). The effects of monocular visual
deprivation: Disuse or binocular interaction? Brain
Res., 111, 277-285.
26. von Noorden, G.K. and Leffler, M.B. (1966}. Visual
acuity in strabismic amblyopia under monocular and
binocular conditions. Arch. Ophthalmol., 71,
172-177. --
27. von Noorden, G.K. (1985). Idiopathic amblyopia. Am.
J. Ophthalmol., 100, 214-217.
23
DISCUSSION: NORMAL AND ABNORMAL
VISUAL DEVELOPMENT
Moderator: GUNTER K. von NOORDEN
von Noorden: Dr Garey has shown that development is more

complete in the monkey at birth than in the human,
especially as far as the lateral geniculate nucleus is
concerned. There may be clinical significance in this
finding. Maybe this is one of the reasons why naturally
occurring strabismus is limited to humans. If the human
visual system is as immature for the first 3-4 months as
Dr Held has shown us, it may be especially vulnerable to
peripheral influences, perhaps from the brainstem, that
may interfere with the development of normal ocular motor
coordination.
Held: I'm not sure you want to make that claim because
there are kittens that show natural strabismus as well as
monkeys. On the other hand I would agree that the lengthy
immaturity may make the visual system vulnerable.
von Noorden: In the strabismic cat there is abnormal

crossing of optic nerve fibres through the chiasm.
Held: That is true of Siamese species but there are
others which seem to be strabismic without the Siamese
anomaly to the best of my knowredge.
Singer: I find the hypothesis very attractive that faulty

development of binocular cooperativity might result from
inappropriate timing. If maturation of the fusion
mechanism lags behind the onset of the critical period
for ocular dominance plasticity, disruption of
binocularity would be an unescapable consequence. There
are diseases such as Down's syndrome and hypothyreoidism
285
where brain development is retarded. Is it known whether

the incidence of squint is increased in these patients?
Levi: The incidence of strabismus in the mentally

retarded and in Down's syndrome and other genetic defects
is very high compared to that on normal population.
Campos: I would like to know from Dr Garey whether

certain adaptive phenomena such as anomalous retinal
correspondence could find a morphological explanation in
pre-existing and available dendrites and synapses in the
visual cortex.
Garey: We do not know about production of spines,

dendrites and synapses in amblyopes. Normally there is
rapid production in the first weeks of life, followed by
a drastic "pruning". This is a possible substrate for the
sort of phenomenon you are asking about. We cannot be
sure that the morphological changes we observe have a
direct relationship with sensitive periods, although they
are correlated in time. However, Dr Held showed us the
remarkable functional changes between about 4 and 6
months that correspond well to the period of rapid
synaptogenesis we described.
Helveston: In a monkey simply doing a large recession,

resection or myotomy, results in a straight eyed monkey
in a very short time. In order to create a strabismic
monkey it is necessary to tie the muscle down and to
produce a restriction. Does this represent a
selfcorrecting motor system and does it fit with the
sensory responses of the monkey?
Garey: When you make monkeys strabismic you are

presumably dealing with a normal nervous system, whereas
this might not be the case in a strabismic human. We have
little idea of structural changes in human amblyopia and
so do not know whether a monkey whose muscles you resect
is a good model, but it seems the best we have at the
moment.
Hel vest on: I have seen several presumably normal humans

after laceration of an extra ocular muscle. Each has
become strabismic immediately and remained so until
appropriate surgery was done. After alignment the
patients reverted to their previously normal sensory
function.
von Noorden: We know of a few cases of naturally occuring
strabismus in Maccacca nemestrina but not in M. mulatta.
Also, unlike humans, blind monkeys do not develop
nystagmus. Are there any anatomical differences between
DISCUSSION: NORMAL AND ABNORMAL VISUAL DEVELOPMENT 287
different kinds of monkeys and between monkeys and humans

that would explain why congenital strabismus and
nystagmus are absent in the monkey?
Garey: When you compare humans with Old and New World
monkeys it is surprising that there are so few
differences in normal maturation, especially when you
remember that the structure of the visual cortex in New
World species lacks typical ocular dominance columns.
Baker: Dr Garey, could you tell us more precisely what

you mean when you used the expression "overproduction" in
your presentation of the development of spines and
synaptic connections?
Garey: By "overproduction " I mean that you find more

spines and synapses in young animals than in adults.
During development one sees several forms of exuberant
dendritic appendage, including growth cones, filopodia
and normal looking spines. The growth cones and filopodia
disappear before the final elimination of spines, so I
feel confident that there is overproduction of adult-like
spines, although I cannot be sure that they are
functional.
Levi: I wanted to follow up on Dr von Noorden's question

about the difference between strabismus in monkey and man
by asking whether any of the ocular motor people know
whether there are differences in the extraocular muscles
of monkey and man.
Spencer: As far as we can tell there are no differences

at all between Maccacca and human. The only possibility
is in the Maccacca who do have tendency to have some
extra muscles like the accessory lateral rectus muscle.
You don't see it in humans. Basically a section through
the lateral or medial rectus of monkey and human looks
identical.
von Noorden: Is there a discussion of the presentation of

Dr Headen and his co-workers?
Crawford: The paper by Dr Headen and co-workers makes

four main points:
(1) There is little or no growth in monkey LGN cell size
after one week of age, i.e., there is little difference
in cell size with age, which leads to the conclusion that
during monocular occlusion, LGN cells get smaller from
shrinkage rather than from some failure to grow
(2) During monocular deprivation, LGN cells connected
with both the deprived and normal eyes show transient
changes in size; a shrinkage of deprived cells, and a
transient hypertrophy of cells connected with the viewing

eye. The point that these deprivation effects should be
compared to normal animals, rather than to the sizes of
cells connected with the companion eye, is well taken.
The consequence of using cells in layers connected with
the viewing eye would be that the LGN cell shrinkage for
the closed eye would be exaggerated, and in fact, in
error, since most of the difference is in the effect seen
in layers connected with the open, rather than the closed
eye. The mechanism for such a hypertrophy, however, is
unknown.
(3} The changes in LGN cell size with monocular
deprivation depends upon whether or not there has been
visual experience prior to the beginning of the
deprivation. When initiated at birth, cell size
differences appear to be produced largely by a transient
hypertrophy of cells connected with the non-deprived eye
and not by shrinkage of cells connected with the deprived
eye. Monocular closure following some visual experience
is characterized by a concurrent shrinkage of all
parvocellular cells, with the differential cell-size
effect decreasing to one year of age. Magnocellular
layers seem to be unaffected.
(4} The monocular segments show similar, but smaller,
incrases and decreases in cell size.
Recall that it was the anatomical changes subsequent
to monocular form deprivation; the easily observed lack
of Nissl staining in adjacent LGN layers which induced
the enormous amount of study of the phenomenon of LGN
cell size. But, it is long overdue to question the
significance of LGN cell size. What do small LGN cells
indicate about visual function? It was noted in kitten
early on by Wiesel and Hubel that though smaller than
normal cells connected with the non-deprived eye,
shrunken LGN cells appeared functionally normal. This has
again been pointed out recently for monkey by Blakemore
and Vital-Durand who could find no functional deficit for
these diminutive cells.
Recall as well the relationship between the size of
the LGN cell soma and the spread (size} of it's axonal
arbor as proposed originally by Guillery. Soma size was
reduced with monocular deprivation, as the axonal
terminals in cortex failed to establish, or were
displaced from, functional synaptic contacts with cells
in layer 4C. The presumptive reason for this shrinkage
was some lessened metabolic demand for support of a
shrinking axonal arbor. It made intuitive sense, and has
been generally accepted as being the major determinent of
shrunken LGN cells; disuse from lack of afferent exercise
being the other factor.
In light of these new observations, a re-examination
of the relationship between LGN soma size and the size of
the axonal arbor in the cortex, as proposed by Guillery,

may suggest a variant on the original theme to account
for the results presented in this paper. The transient
hypertrophy results in LGN cells connected with the non-
deprived eye because of additional metabolic demands
placed on these cells in establishing exclusive synaptic
control over processes in layer 4C of the striate cortex
normally controlled by the closed eye. This process has
the concurrent effect of eliminating the innate
functional population of binocular cells which provide
reciprocal connections with the LGN via cells of cortical
layer 6, and all LGN cells begin to shrink. On this last
point, we can offer some corroboration of the necessity
of binocular function for maintaining normality of LGN
cell size.
Several years ago we began using the technique of
optical dissociation by prisms to create an artificial
strabismus and a condition of chronic diplopia in infant
monkeys (see the paper on the physiological basis for
stereopsis). In those experiments, we measured LGN cell
sizes, comparing differences between layers having input
320
!J N =8 (Crawford,
NORMAl MONKEYS
1987)
280 IZl N= 7 PRISM REARED MONKEYS
240
N~
~ 200
:Iw
0
~
180
-Hi 'AI
120
80
1 2
MAG. PARVO.
LGN LAYERS
Figure 1. The average size of cells of the LGN of seven

monkeys which had optical dissociation by viewing the
world through prisms for 30 days, beginning at 30 days of
age. The cell sizes are compared to those of normal
monkeys (Crawford, unpublished observations).
from the right and left eyes. We found no differences and

concluded that optical dissociation had no effect upon
LGN cell size.
However, when we later compared cell sizes in these
monkeys to a population of normal monkeys (Fig. 1) , we
found that all of these animals had clear and significant
reduction in cell size in all LGN layers. As a group,
cells of the magnocellular layers were -12%, and those of
the parvocellular layers -18% smaller than normal
(p<.OOl). Animals reared like this have no amblyopia, and
characteristically have very few binocular cells, and all
turn out to be stereoblind.
These results are consonant with the idea suggested in
this paper that functional integrity of the binocular
system is necessary for normal LGN morphology.
Bagolini: I would like to pose a question to Dr Headen.

Since the foveal function is the one that suffers most
both in deprivation and in strabismic amblyopia, I would
expect a shrinkage of the cells or a damage of the cells
in the geniculate body mainly in those cells that receive
input from the foveal area. Do you have any evidence that
there is selectivity in the damage.
Headen: The shrinkage seems to be less in the macular

representation in the monkey.
Singer: It is worth emphasizing that activity-dependent

pruning of connections does not only depend on
competitive but also on synergistic interactions between
converging afferents. There are numerous examples from
other systems which indicate that a critical amount of
cooperativity is needed for the stabilization of
converging inputs. Since at the time when binocular
cooperation prevents shrinkage of LGN cells layer IV is
no longer plastic I wonder whether synergistic
interactions occur in layer VI. Is there evidence for a
prolonged critical period in layer VI?
von Noorden: Would anyone like to comment on that.
Levi: While we certainly don't understand the functional

significance of cell sizes changes, it is not fair to say
that there is no significance because amblyopia is a
macular disorder. Certainly that maybe true for
strabismic amblyopia but form deprivation has much more
wide-spread effects and anisometropic amblyopia and form
deprivation amblyopia probably influence peripheral as
well as central visual function. Therefore I'm not sure
we should discount these cell size changes because they
are marked in the periphery.
Sloper: This macular difference was only present for

parvocellular cells. The difference between deprived and
undeprived magnocellar cells was as big at the macula as
more peripherally but the difference for parvocellular
cells was only between a half and two-thirds as much.
von Noorden: In our monkey experiments we found

consistently a sensitivity gradient that goes as follows:
parvocellular layer magnocellular layer monocular
segment. Your findings seem to be along the same line. I
wonder whether you have any theories on that?
Sloper: The hypertrophy in general is more marked for

parvocellular cells, but affects both types to some
extent. The late shrinkage is specific for parvocellular
cells in all the conditions we have looked at. Dr Headen
has described here the results in monocular closure, but
the same is true after early enucleation and in the long
term after reverse suture. It appears that anything which
upsets binocular cooperation selectively causes late
parvocellular shrinkage without affecting the
magnocellular cells. I would like to ask Dr Crawford
whether he also finds this with his prism dissociated
animals.
Crawford: We found with the prism-reared animals, that

the percentage of cell shrinkage was a bit more for the
parvocellular layers. But, we are talking about only 5%
difference, which I think that, given the sample size, is
a trivial difference. So, we get deprivation effects in
both divisions of the lateral geniculate nucleus.
von Noorden: We go on to the paper by Dr Smith et al.

presented by Dr Harwerth. These results are similar to
those in a group of patients with monocular congenital
cataracts that underwent axial length measurements and
showed almost consistently a significant difference
between the deprived and the non-deprived eye*. The
deprived eyes were significant longer than the non-
deprived eye except for two patients who showed the
opposite: the deprived eye was shorter than the non-
deprived eye. Thus, humans also show a tendency towards
myopia after visual deprivation in infancy but exceptions
do occur and some of these eyes actually become
hypermetropic.
Campos: I wonder whether changes of refraction after lid

closure are simply attributable to form deprivation or
whether a pressure effect exerted by the lid may play a
*von Noorden, G.K. and Lewis, R.A. (1967): Ocular axial
length in unilateral congenital cataracts and
blepharoptosis, Inv. Ophthal. Vis. Sc. ~. 150.
role. Monolateral cataract is in fact often associated
with microphthalmia, whereas monolateral ptosis patients
are not rarely myopic.
von Noorden: There is evidence that opacification of the

cornea alone suffices to produce these changes.
Harwerth: In addition, the dark rearing experiments of

Raviola and Wiesel* showed that light stimulation was
required for the development of experimental myopia from
lid suture.
Eggers: I just like to add a related fact that perhaps

muddies this water a little bit. In Atlanta we have a
project studying infant aphakia in monkeys. We find that
the aphakic eye becomes substantially shorter in axial
length compared to the normal eye. The lens plays a role
in the growth or the eyeball.
Bagolini: Dr von Noorden, I would like to know what type

of cataracts are the ones presented in your slide. A
posterior polar cataract is usually due to an imperfect
reabsorption of the hyaloid artery. There may have been a
difficulty in the circulation of the eye so there may
have been some other causes besides deprivation to lead
to an abnormal growth of the eye.
von Noorden: The patients that were reported all had

complete congenital cataracts.
Bicas: If the amblyopic eye tends to be more myopic, why,

then, it remains (or becomes) more hyperopic in bilateral
hypermetropias?
Haase: I cannot answer directly this question but in my
patients who I observed for a period of 10 years or more
the leading eye grew and needed less hypermetropic
corrections that the amblyopic eye did, the result was
new anisometropia.
Harcourt: Many patients with treated bilateral congenital
cataracts also have in later life abnormally long
posterior segments, and the common long term complication
of retinal detachment may then be related to their myopia
as well as to aphakia.
von Noorden: With regards to Dr Crawford's paper: What is
*Raviola, E. and Wiesel, T.N. (1978) Effect of dark

rearing on experimental myopia in monkeys. Invest.
Ophthalmol. Vis. Sci. 17, 485-488.
the significance in behavioural terms of a loss of so-

called binocular cells? There seems to be a
misundertstanding between clinicians and basic scientists
about the term binocularity. For clinicians it means
sensory fusion, the blending of two identical images
coming from the two eyes into a single mental percept.
Neurophysiologists talk about binocular cells when they
describe cells that respond successively equally well to
stimulation from either eye. What is the significance of
these binocular cells?
Crawford: I am not sure that I understand your question.

There are several things that the loss of cortical
binocular cells means in behavioral terms. For one thing
it means that unless you have them you don't have the
perception of stereopsis. The general physiological
description of what constitutes a binocular cell,
obviously, is a cell that has input form the two eyes,
and they come in a variety of forms. Most important are
those that are sensitive to horizontal image disparities
and which I consider in the presentation tomorrow on the
physiological basis for stereopsis.
von Noorden: What is the electrophysiological equivalent

of sensory fusion? Are there cells or centers in the
visual cortex that are responsible for this specific
activity, separate from stereopsis? It is certainly
possible to have sensory fusion and be stereoblind.
Blakemore: For stereopsis I presume that there must be a

system of binocular neurons, each selectively sensitive
to a particular horizon tal disparity with variation in
optimal disparity amongst the population of cells (see
Barlow, Blakemore and Pettigrew, 1967) *. Fusion without
steropsis might conceivably occur in one of three states:
1) binocular neurons all tuned to exactly the same
horizontal disparity;
2) failure of the mechanism for interoretinq the signals
from disparity-selective cells;
3) breakdown of binocularity but maintenance of
independent, monocular systems of egocentric visual
direction, such that single objects in space give rise to
the same sense of direction through the two eyes.
Singer: As I mentioned before, amblyopic patients can
compute the trajectories of objects moving in depth by
evaluating binocular signals. This ability is preserved
in the whole visual field except in the fovea. Thus,
*Barlow, H.B., Blakemore, C. and Pettigrew, J.D. (1967)

The neural mechanism of binocular depth discrimination.
J. Physiol. 193, 327-342.
there is preservation of a binocular function that is

most likely realized outside area 17. I believe we are
also forced to postulate an involvement of non-striate
areas in the context of anomalous correspondence. In
prestriate areas receptive fields tend to be large and
retinotopy is no longer very precise. In cases where the
squint angles are so large that anomalous correspondence
can by no means be established in area 17, binocular
integration must occur in non-striate areas. In these
cases there should of course be no stereopsis, since this
function must reside in areas with small receptive fields
and precise retinotopy.
Held: What is the opposite of fusion?
Campos: Rivalry is a pnenomenon which causes the lack of

sensory fusion.
Bicas: I have another question. Many adult patients with

deep amblyopias show an improved visual activity of the
good eye, vision is usually better than that found in
normal subjects. Does this support the view that the
visual system may have some compensation to the
unilateral amblyopia?
Blakemore: Freeman and Bradley (1980)* reported that the

non-amblyopic eyes of amblyopes tended to have slightly
higher vernier acuity (but not better resolution acuity)
than the eyes of normal observers. However, Johnson et
al. (1982) ** pointed out that amblyopes are likely to be
well-practised psychophysical observers (because of years
of visual testing) and therefore differences in training
might explain this result. They found no such difference
in vernier acuity between the non-amblyopic eye of one
amblyopic patient and the normal eye of that patient's
well-trained identical twin.
Campos: I would like to comment on Dr Crawford's findings

concerning the loss of binocular cells after bilateral
deprivation. In humans, an early induced bilateral low
visual acuity causes nystagmus and loss of binocularity
because of a bilateral visual acuity which is too low to
be compatible with binocularity. Therefore, even if a
*Freeman, R.D. and Bradley, A. (1980) Monocularly
deprived humans: non-deprived eye has supernormal vernier
acuity. J. Neurophysiol. 43, 1645-1653.
**Johnson, C.A., Post, R.B., Chalupa, L.M. and Lee, T.J.
(1982) Monocular deprivation in humans: a study of
identical twins. Investig. Ophthalmol. Visual Sci. 23,
135-138.
sensory-defect nystagmus is not present in monkeys, the

loss of binocular cells should not be surprising.
Crawford: I don't think so. Nystagmus is not a notable

characteristics in these animals.
von Noorden: We now come to dr Blakemore's paper that

showed, among other things, that there appears to be no
relationship between visual acuity and cortical
dominance.
Harwerth: I was interested in the early portion of the

sensitive period. In the cat, you presented data showing
that there is a short period of life before the sensitive
period starts. Do you think that a similar period exist
in the monkey?
Blakemore: I don't think that there is unequivocal

evidence on this point in man but in monkey there does
not seem to be an initial post-natal period of
insensitivity to the effects of monocular deprivation. If
you will forgive a teleological argument, there is a good
reason for such a period in kittens because they are born
with closed eyes and the two eyes do not always open
synchronously (Blakemore and Cummings, 1975)*. The delay
in the start of the sensitive period ensures that the
kit ten is not affected by such natural monocular
deprivation.
Sloper: We have seen changes in cell sizes after 10 days

in a monkey with monocular closure started at birth.
von Noorden: What type of neuronal changes might underlie

the degradation of spatial performance in the cortex of
deprived monkeys?
Blakemore: In normal monkeys, the cells with the highest
resolution in both LGN and cortex, have spatial acuity
corresponding roughly to that expected of a single foveal
cone. To achieve such performance at the cortex requires
that there should be very little convergence from
incoming geniculate axons. I suspect that visual
stimulation is required for certain cortical neurons to
"select" input from a very small number of LGN axons with
spatially coherent receptive fields. The degradation in
performance after deprivation could, then, be due to
increased spatial convergence on to cortical neurons.
Singer: I would like to comment on the important
*Blakemore, C. and Cummings, R.M. (1975) Eye-opening in

kittens. Vision Res. 15, 1417-1418.
observation that normal ocularity can go along with poor

resolution and vice versa. This may be accounted for by
the evidence that not only the afferent connections from
the two eyes but also intracortical connections require
for their normal development that visual experience is
available. This holds true for both cortico-cortical and
for intraareal tangential connections. These prominent
and very sophisticated networks develop post-natally and
are pruned by visual experience. Consequently, some of
the disturbances that we see after deprivation may
actually result from connectivity errors at this level of
intracortical organization. The organizaiton of these
intercolumnar connections suggests that they mediate
long-range interactions. They could thus be involved in
functions requiring spatial integration such as e.g.
hyperacuity.
Crawford: Dr Blakemore, you emphasized a couple of times

in your presentation the use of data from the best cells
encountered: and using the best cell concept, you in fact
concluded that there were no differences in the response
between cells in the deprived and normal layers in LGN.
What about the average cell? Could you comment about the
rationale for using the best cell. How is that related to
the particular visual function?
Blakemore: In the LGN there is only a small range of

spatial resolution for the receptive fields centred at
any one point in the visual field and the results are
much the same whether one compares the cells with the
best resolution or the mean resolution for all cells
between normal and deprived layers. However, in the
cortex there is enormous variation in spatial performance
from cell to cell: the mean resolution of cortical cells
in normal animals is much lower than that of LGN cells,
but the acuity of the best cells is very similar in
cortex and LGN. Another reason for paying special
attention to the best cells found is that, in normal
monkeys, the performance of the best cells (in LGN and
cortex) is remarkably close to human psychophysical
performance at the corresponding point in the visual
field. If one believes that the visual acuity and
contrast sensitivity of the whole animal might depend on
the properties of the best-performing neurons in its
visual pathway, it is obviously important to consider
their properties specifically.
Crawford: Yes, I agree, especially in threshold

considerations that would be the case.
von Noorden: Dr Blakemore's data are somehow at odds with

our data which have shown a pretty good correlation
between visual acuity and the number of neurons that

could be driven through each eye. We also have the
clinical example of the child who can switch amblyopia
from one eye to the other under the influence of patching
or of the monkey in whom by reverse lid closure we can
switch the cortical dominance from one side to the other
in a matter of days. In the past this cortical dominance
shift matched up well with visual performance. Do we have
to throw all this overboard?
Blakemore: I'm sorry if I have given the wrong

impression. In most cases there is very good agreement
between the relative spatial performance of cortical
neurons through the two eyes and the overall ocular
dominance. But there are exceptions. Very late monocular
deprivation can shift ocular dominance without altering
resolution. Equally very brief early deprivation may
change contrast sensitivity without obviously shifting
ocular dominance.
von Noorden: We will now discuss the paper by Dr

Harwerth.
Siostrand: Your findings of different sensitive periods

are extremely interesting. For a clinician it would be
very important to know if you have any information on
whether you have any difference in recovery or
reversibility of the different psychophysical functions.
Have you performed reverse suturing of the eye lid in
these animals?
Harwerth: We did not perform reverse suture in the

animals used in the studies of sensitive periods reported
here, because these were long-term deprivation
experiments. Crawford (this volume) has presented some of
our data on reverse deprivation in which we used shorter
periods of deprivation. In those studies, the spatial
modulation sensitivity functions usually did not show
full recovery while the temporal modulation sensitivity
and increment threshold spectral sensitivity functions
demonstrated complete recovery. Therefore, there is also
a difference in the reversibility or recovery of
different psychophysical functions.
Campos: Can one differentiate in monkeys strabismic from
deprivation amblyopia? A second question is related to
the use of a bilateral patch: is it effective and safe to
propose?
von Noorden: Our lab has shown that bilateral visual

deprivation in monkeys for period as long as 6 weeks has
no effect on a variety of visual functions tested.
Binocular patching for longer periods caused irreversible

changes. Can we make some conclusions from critical
periods determined in monkeys to the critical period
determined in man? Is there some conversion factor that
would allow us to patch both eyes of a baby with
monocular cataracts safely to prevent the deleterious
effect of competition in the visual system? Do you have
information that you think we can use?
Harwerth: I can not give you a specific conversion

factor, but the data from our experiments would seem to
indicate that a short period of bilateral patching would
be safe.
von Noorden: A 1:4 ratio has been mentioned. Using this

ratio, your data would indicate that humans remain
sensitive to abnormal visual input up to about the age of
8 years. This is, in fact, in accordance with our
clinical experience. You don't get amblyopia in a child
older than 8 or 9 years. There seems to be a good
correlation. Is there anything you can add to that.
Harwerth: The data for the 1:4 ratio for the relative
rates of development of monkeys and humans comes from
comparisons of the relative rates of development of
visual acuity in the two species*. The generalizability
of the ratio would be strengthened if other functions
that were compared for monkeys and humans resulted in the
same ratio. One other function that has been compared and
found to develop at about the same 1:4 rate is
accommodation**· In addition, studies of the development
of the ocular components of cynomolgus monkeys have shown
an approximate ratio of 1:3 for monkey to human
development***· Therefore, the results of several studies
seem to converge and I do not think that the 1:4 ratio is
very far off.
Helveston: Visual results after surgery for uni.lateral

cataract under the best of circumstances are poor.
Therefore, in my opinion bilateral patching a patient for
* Boothe, R.G., Dobson, V. and Teller, D.Y. (1985)

Postnatal development of vision in human and nonhuman
primates. Ann. Rev. Neurosci. ~, 495-545.
**Howland, H.C., Boothe, R.G. and Kiorpes, L. (1982)
Accommodative defocus does not limit development of
acuity in infant Macacca nemestrina monkeys. Science 215,
1409-1411.
***Kiely, P.M., Crewther, S.G., Nathan, J., Brennan,
N.A., Nathan, E. and Madigan, M. (1987) A comparison of
ocular development of the cynomolgus monkey and man.
Clin. Vision Sci. 1, 269-280.
any period of time before surgery puts the good eye at

risk. On the other hand visual results after treatment of
bilateral cataracts is better.
Harcourt: Could Dr von Noorden please tell us whether he

advocates bilateral occlusion in young infants undergoing
surgery for bilateral cataracts, during the short
interval of perhaps one week between the operations on
the two eyes. A temporary but profound inequality in
visual input exists during that period.
von Noorden: On theroretical grounds this seems advisable

but I have no clinical proof that it works.
Singer: I suggest one should investigate the effects of

synchronous binocular stimulation before proceeding with
binocular patching.
Crawford: We have tested these animals up to periods of 3

years and they never recover binocular neurons. They
become irreversably stereoblind.
Campos: I have followed Dr von Noorden' s advice since

1981 to patch bilaterally in bilateral cataracts, waiting
between the surgery of the first and the second eye. I
cannot say that I was able to preserve reliably binocular
vision. Probably there was an inbalance before surgery or
other unknown elements are involved.
von Noorden: I agree on that, but let us talk about

another situation, a baby not visually deprived from
birth that develops a corneal injury or another condition
requiring temporary patching. Should one patch the good
eye also?
Bagolini: I think that a monkey model is not always a

good one. I learned today that monkeys don't develop
nystagmus with bilateral deprivation, while a child
develops nystagmus during the first 2 years of life. My
practical suggestion, would be to patch the patient
bilaterally but not long enough to see nystagmus appear.
When nystagmus appears in my experience the visual acuity
starts to approach from two to four tenths of visual
acuity.
von Noorden: Let us now discuss the paper of Dr Held, who

pointed out that there is a certain time between the 3rd
and 4th month of life when everything begins to happen at
once. Stereopsis, disparity, sensitivity for vergence
movements and rivalry occur at about the same time.
Harcourt: I am rather anxious that not too much should be

deduced from behavioural testing in 2-3 months old

subjects. Dr Held suggested that the lack of difference
in visual acuity between the two eyes in strabismic
children with a marked fixation preference at that age,
and who later developed amblyopia, might indicate a
"grace" period before the sensitive period. Might it not
just be that the testing procedure at that age was
inadequate to indicate small differences between the
visual function of the two eyes. Does he know whether
there were any such differences indicated by visual
evoked potentials {VEP) in the same children.
Held: In answer to your first question, we tested 4

infants with early esotropia and a marked fixation
preference with uniform results. That is a small sample.
However, the result has been confirmed with seven other
infants from another group tested by Eileen Birch {Birch
& Stager, 1985) *. Of course I would prefer to have a
sample of one hundred such patients and I agree that our
results are subject to qualification as a result of the
small sample. VEP measurements show much more rapid
development such that people claim that infants are up to
adult acuity levels by the end of the first year. Dr
Maffei's group did some work on the problem and, if I
understand correctly, he says that the time course of
development of amblyopia was very similar to the
behavioural work which is presumably an answer to Dr
Harcourt's second question.
Blakemore: I share some of Dr Harcourt's worries about

behavioural performance of very young infants. For
monkeys it is clear that during the first few weeks of
life the visual acuity determined by preferential looking
is considerably worse than the performance of cortical
cells. So one has to ask whether there is some post-
cortical spatial filter that limits behavioural
performance or whether very young animals do not espress
their full behavioural potential in the preferential
looking method because of lack of motivation or some
other factor.
Held: I think the only answer I can give to that is the

curious fact that, as Dr Aslin pointed out in his talk,
VEP measurements of stereopsis in infants give exactly
the same age of onset as behavioural methods. That
suggests that there are some real differences which we do
not understand between the rates of development of
grating acuity measured by the different methods.
*Birch, E. and Stager, D. {1985) Monocular acuity and

stereopsis in infantile esotropia. Invest. Ophthal.
Visual Sci., 26, 1624-30.
DISCUSSION: NORMAL AND ABNOR,\iAL VISUAL DEVELOPMENT 301
von Noorden: Dr Held reported that strabismus can be
present for several weeks or even months before amblyopia
sets in. This observation is only meaningful if
correlated with the fixation behaviour. Many strabismic
children alternate before preferring one for fixation and
becoming amblyopic.
Held: I agree that it is a serious problem. I have no

real answer to that other than the fact that it appeared
in our sample that the infants had strong fixation
preferences all along.
Aslin: I have two comments. The first has to do with the

question of VEP acuity versus preferential looking (FPL)
acuity. Norcia and Tyler (1984)* have shown that under
approximately 3 months of age estimates of acuity based
on VEP are superior to estimate based on FPL. In
contrast, Sokol has reported very similar acuity
estimates using exactly the same kinds of flashed
displays with both VEP and FPL procedures even in young
infants. There is something of a "red herring" here in
that the criteria one uses to define threshold on these
tasks is quite different. In the case of VEP, threshold
is typical extrapolated from an amplitude of zero,
whereas in the case of FPL, threshold is based on
behaviour that exceeds 50% correct. Because of these
different criteria, the important point is not the
absolute estimates of acuity but the relative changes in
acuity over age. When one compares these relative changes
across the two techniques, they are more or less the
same. The second comment has to do with the development
of amblyopia and strabismus. The very poor sensory
capabilities that we see in normal infants up until 4
months of age, including contrast sensitivity deficits,
acuity deficits, and absence of fusion and stereopsis,
may be analogous to the period of eyelid closure in cat
mentioned earlier by Dr Blakemore. The normal visual
system seemed to be "buffered" from the effect of
anomalous visual experience by having sensory mechanisms
that are largely unaffected by all but the most extreme
forms of visual deprivation.
Held: I wish that it were entirely true. But the idea
that they are buffered against all the possible anomalies
is belied by the effects of early astigmatism, which has
a high incidence in infants ranging from 20-50%. One can
detect the visual effects of early astigmatism years
later even though emmetropization has completely
*Norcia, A.M. and Tyler, C.W. (1984) The growth of

spatial and temporal resolutions in human infants.
Invest. Ophthalmol. Vis. Sci., 25, 161. ARVO suppl.
302 G.K. von NOORDDI
eliminated the presence of astigmatism as early as a year

of age. So we are not entirely protected.
Haase: A question to Dr Held regarding development of

amblyopia weeks or months after the manifestation of
squint. We observed exactly the same course: first
manifestation of squint, weeks or months later amblyopia
occurred. Did you see any correlation between the
manifestation of laterality of the brain, and this
manifestation of the leading or dominant eye?
Held: No, we have no way of assessing brain dominance as

far as vision goes in infants with our technique. We
cannot present lateralized stimuli.
Campos: I would like Dr Held to elaborate a little bit

more on the differences in response between males and
females that he found.
Held: In vernier acuity and in a measure of rivalry-

fusion that I have not discussed, we have found that
females develop significantly more rapidly than males. My
tentative conclusion is that one of the major constraints
on grating acuity is the retina whereas the major
constraint on processing of vernier acuity and
binocularity is cortical. Only cortical functions appear
to be differentiated sexually. The question of what
causes this difference is rather interesting and I just
happen to have a couple of slides. One that I borrowed
from Dr Garey shows that the most rapid rate of
synaptogenesis occurs between 2 and 8 months of life, the
busiest time let us say in the developing nervous system.
The question then is what could modulate a difference
between male and females during that very rapid period of
synaptogenesis and my hunch has been, if I may have the
next slide please, that it is testosterone. Male infants
have high levels of testosterone in their blood during
the early months of life. Females have none. There is
evidence that testosterone can modulate neural growth. My
suspicion is subject to verification which we are trying
to produce.
Helveston: Dr Held referred to acquisition of stereopsis

as a marker of change. We noticed the same marker when we
measured stereopsis at 9 months using the dynamic random
dot sterogram in normal infants. No strabismic infants
demonstrated stereopsis. In examining more than 2.500
infants between 3 hours and 3 days of life we have not
been able to find a single case who fit the criteria for
diagnosis or infantile esotropia. However, we have
collected several patients who were straight in the first
few months but at 4 months developed the characteristics
of infantile esotropia. I would suggest that we recognize

a loose relationship between the eyes during the first 4
months. At 5 months it becomes a "fish or cut bait
situation". The eyes are straight or the patient is
strabismic. We have observed 6 patients with congenital
unilateral or bilateral VI nerve palsy who recovered
completely with normal stereopsis.
von Noorden: I am having problems understanding how one

can equate Snellen acuity with grating acuity. In my view
these are not quite the same and there are descrepencies,
especially in amblyopes. I have also heard several
speakers talking about stereopsis and vergences in the
same breath. We know that it takes disparate retinal
stimulation outside Panums space to induce a vergence
which is a motor response. Stimulation of disparate
elements within Panums space elicits the sensation of
stereopsis which is a sensory response. Are there
relationships?
Held: Well, I completely agree with you about acuity. The

Snellen correlated poorly with grating acuity but better
with vernier acuity. The use of the Snellen number is
simply an arithmetical conversion. In regard to the
second question, there is obviously a circular
relationship. In order to have stereopsis one has to have
a certain accuracy of vergence unless one has very
special kinds of stereograms which are redundant. So you
need vergence to get stereo but you also need a measure
of disparity in order to get open loop changes of
vergence. Namely, if you want to use a depth step by
placing a prisma in front of the eyes which requires you
to change vergence by a jump, the only way that the
system can calculate that step without hunting is to have
a disparity measure.
von Noorden: I would like for Dr Schor to comment on

that.
Schor: It is important to note that stereopsis is really

a discriminating task. It involves sensing differences
between multiple dis pari ties in the visual field and it
requires further processing than simply correlating
features and obtaining an absolute disparity. The
vergence system simply responds to absolute disparity.
Possibly absolute disparity processing develops prior to
relative disparity processing. Hence fusion and motor
vergence could develop prior to stereopsis.
Held: That may be true but the important question is: are
there disparity mechanisms other than cortical?
Schor: There is the OKN syndrom where you have an

asymmetric horizontal OKN which is related to both
cortical and subcortical development of binocular vision
and that maturation of OKN occurs at about the same time
as the onset of steropsis.
Held: Does that buy you vergence control do you think?
Schor: No, that is simply another measure of

binocularity.
von Noorden: Are there any comments on Dr Singer's paper?
Lindstrom: You showed that the plasticity with monocular

deprivation disappears after APV infusion in the cortex.
I find this very difficult to conciliate with the rest of
your findings, mainly because APV interferes with the
NMDA receptors which are excitatory receptors. In the
case where it is known that these receptors are involved
in plasticity, that is in the hippocampus, the NMDA
receptors underly long term potentiation, which is an
intensification of the activated synapses and not a
depression of the non-stimulated ones.
Singer: This is true. We are lacking a good in vitro

model for activity-dependent depression of inactive
pathways. The fact that blockade of NMDA-receptors
abolishes repression of the deprived afferents is
compatible with the evidence that repression requires
concomitant activation by afferents from the open eye.
This activation is reduced by blockade of NMDA receptors.
Eggers: One of your diagrams implied that the muscle
afferents might play a role in the gating function that
supervises plasticity. Can you laborate on that?
Singer: In 4-week-old dark-reared kittens proprioceptive

input from extraocular eye muscles was disrupted by
bilateral section of the ophthalmic branch of the
trigeminal nerve. At the same time kittens had one eye
closed and then were raised in normal visual environment.
Behavioral tests revealed that these kittens failed to
develop normal visuo-motor reflexes. They bumped into
objects, showed abnormal landing responses and behaved
very much like binocularly deprived animals.
Correspondingly, even though these animals had
undisturbed monocular vision for altogether 4 weeks, the
cortex was in a functional state very similar to that
found after binocular lid suture.
Aslin: I would like to ask you a question about the

feasability of your neural correlation model for
enhancement of synapses by synchronous visual input. It

seems to me that this model could work in principal. For
example, if you have a single point in a dark room it
would seem quite easy for the vision system to correlate
these two retinal inputs. However, in the normal
environment, which is terribly cluttered, there are many
elements that could potentially match in the two retinas.
I wonder, given that you know the time constant
associated with firing rates and you know in general how
much visual experience an organism has per day, whether
it might be possible to create a computer simulation to
determine whether this in principal mechanisms is in fact
feasible in the lifetime in the organism.
Sinaer: Experience-dependent pruning of connections is

actually implemented in most connectionists models, but I
don • t think that we know enough about the parameters to
arrive at a realistic estimate of time courses.
Certainly, we are not dealing with a one-trial learning
procedure where a single mismatch leads to disconnection
of pathways. The selection mechanism must be able to
average and to extract the statistical probability of
coherent activation over weeks before decisions become
irreversible at the end of the critical period.
von Noorden: Dr Singer, I am still puzzled how the

decision is being made in the visual cortex which one of
two equally appropriate excitations from the two eyes to
accept and which one to reject. Let us take the example
of a strabismic patient who looks at an object in visual
space. There happens to be another equidistant object
that stimulates the fovea of the deviating eye. One image
is accepted, the other is rejected. How does this
selection take place?
Singer: Even if the levels of activation from both eyes

are absolutely symmetrical, temporal incongruency will
cause competition: slight assymmetries in initial
coupling strength of afferents converging from the two
eyes will amplify. The effect is that at any particular
neuron only afferents from either the right or the left
eye will consolidate. The disruption of binocularity as
it occurs with alternating squint is thus well accounted
for. It is more difficult to explain the suppression of
one eye when inputs from both eyes are equally acceptable
in terms of image quality. In that case a central bias
has to be created. This could be achieved either by
reducing the amplitude of the response of one eye by
inhibition or by facilitating the response of the other
by positive feedback. Both mechanisms have been shown to
exist. We ignore, however, how and where the decision is
made of which eye ought to be favored. Single cell data
from experiments in awake monkeys might suggest a

mechanism for input selection. When the monkey shifts its
attention within the visual field this leads to an
enhancement of neural responses to stimuli presented in
attended part of the field. Such attention-dependent
enhancement has so far been observed only in subcortical
and prestriatal visual areas, but it appears entirely
conceivable that it can be back-propagated to striate
cortex or even the geniculate. The necessary connections
are there. A slight enhancement of responses from one eye
would suffice to promote selective consolidation of the
facilitated subset of afferents. This is particularly so
if, as I have proposed in my talk, the pruning process
has a threshold which is reached only if there is
sufficient cooperativity between retinal and non-retinal
input signals.
Lindstrom: I think there are good reasons to believe that

the cortico-geniculate system is part of an amplifying
neuronal machinery used in visual attention. At least in
the cat, the neurons of this system are heavily biased
towards one or the other eye. So here is clearly a
neuronal mechanisms which could enhance the input from
one eye with respect to the other.
Bagolini: I have a question to pose, particularly to Dr

Singer, which goes a little bit further on speculation
about suppression. I think that it is logical to believe
that suppression, which is a clinical term, is supported
by a synaptic inhibition somewhere in the visual cortex.
Clinically we measure the intensity of suppression
creating a graded artificial imbalance between the two
eyes. The simple way is to have optical filters of
progressive density. The density of the filter placed in
front of one eye which is capable of provoking diplopia
provides a good measure of the intensity of suppression.
Usually we measure the intensity of suppression in some
clinical cases in which we fear diplopia may develop.
What does intensity of suppresion mean from a
neurophysiological point of view? That synaptic
inhibition of the cells can be more or less strong or
that the suppression is related to a larger or smaller
number of inhibited cells? Will somebody be prepared to
speculate on this problem?
Singer: Both mechanisms are conceivable. Neurons could

either be silenced completely, thus reducing the number
of active neurons or their level of activation could be
reduced in a graded way. In both cases inhibition could
prevent experience-dependent developmental processes.
There is good evidence from in vitro studies that
GABAergic inhibition interferes very effectively with the
NMDA receptor mechanism and thereby can block plasticity.
Blakemore: I should like to ask the clinicians for their

view of the relationship between conventional rivalry,
suppression and amblyopia. Is it possible, for instance,
to have suppression (reduced visibility) without a change
of acuity?
Bagolini: Between suppression and retinal rivalry there

is a strict relationship. If one eye is strongly dominant
the rivalry will always be at the expense of the non-
dominant eye ending up with suppression. Concerning
visual acuity, suppression is a condition that is only
present in bionocular vision as Dr von Noorden pointed
out in his paper. Suppression disappears in the deviated
eye when we cover the fixing eye, and its visual acuity
remain normal. If, however, the patient is not
alternating, in the long run the eye usually becomes
amblyopic.
Maffei: Dichoptic presentation of patterns similar in

shape but of very different contrast results in the
perception of only the high contrast pattern (binocular
suppression) . When recording from binocular neurons of
the eat's visual cortex we have found an effect which is
strinkingly similar to this perceptual phenomenon*. If a
high and a low contrast grating are presented
simultaneously, one to each eye, the cell's response to
the low contrast stimulus is suppressed. Thus the
neurophysiological basis of suppression could be in a
mechanism of non-linear selective amplification at the
cellular level.
*Berardi et al. (1986) Exp. Brain Res.§]_, 581-584.

Session III
PSYCHOPHYSICS RELATED TO
24
VISUAL PERCEPTION IN STRABISMUS
EMILIO C. CAMPOS
Strabismic patients develop mechanisms attempting

to allow a visual perception as close as possible to
that one of normals in spite of the deviation. These
mer.hanisms however are usually effective only in
childhood, i.e., in the plastic age of the visual
system.
DIPLOPIA, CONFUSION, AND ADAPTIVE MECHANISMS
At the onset of any type of strabismus, diplopia

and confusion are present. Shortly thereafter both are
eliminated by means of suppression of the image of the
deviated eye. This is an active inhibition process
taking place at the level of the striated cortex.
Elimination of diplopia and confusion through

suppression is achieved at considerable cost, because
all binocular functions are lost. Later on, particularly
in congenital strabismus, another anti-diplopic
mechanism develops, namely anomalous retinal
correspondence (ARC) . ARC is present both in large- and
small-angle strabismus (for small-angle strabismus we
mean a deviation not exceeding 10 degrees). In small-
angle strabismus ARC provides to a certain extent a
binocular cooperation. ARC can be defined as a
functional rearrangement of the spatial value of the
retinal elements. According to classical theory, for
each retinal element in one eye, there is normally only
one retinal element in the fellow eye that corresponds
to the same position or direction in space. These are
corresponding elements. The normal spatial
correspondence of retinal images is disrupted in
311
312 E.C. CAMPOS
strabismus. Because of ocular deviation, corresponding
points in space are projected onto non-corresponding
retinal elements. In the case of ARC some patients
acquire binocular integration, as though the spatial
"map" of the deviated eye has shifted so that
anatomically disparate retinal positions gain functional
correspondence with the spatial coordinates of the
fellow eye. Thus, the patient may be capable of
binocular visual localization and in certain cases even
of coarse stereopsis: (Mehdorn, 1982; Schor et al,
1983). This adaptive modification of binocular
correspondence is referred to as anomalous binocular
vision (Bagolini, 1967). In this situation, there is
usually no suppression of the image of the deviated eye.
ANOMALOUS BINOCULAR VISION
Bagolini (1967) showed that in patients with small-

angle esotropia and ARC an area of binocular single
vision can be found, which is similar to to the Panum's
area of normals and was therefore defined as pseudo-
Panum's area. However, the pseudo-Panum's area has
certain characteristics that clearly differ from the
Panum's area of normals. First, the pseudo-Panum's area
is much wider than the normal Panum's area, and it is
variable in size in the same subject. Whereas the
Panum's area of normals supports stereoscopic
perception, the latter is often absent or coarse, if
present, in strabismus. The pseudo-Panum's area can be
explained in terms of retinal correspondence (Bagolini,
1967). Unlike normals, in patients with ARC one point of
the fixing eye corresponds to a more extended retinal
area of the deviated eye.
Anomalous binocular vision has been demonstrated

also with techniques of binocular campimetry (Campos,
1982). In fact, an anomalous superimposition of the two
visual fields was found in more than fifty patients with
small-angle esotropia and ARC. Inside the examined area
which involved roughly 30 degrees of the central field,
there was a binocular single perception sustained by ARC
with little or no suppression. It has to be noted,
however, that the above mentioned results were obtained
by using as test-targets £usable stimuli (i.e. equal for
the two eyes) and monocularly presented items in order
to differentiate a single perception due to suppression
from one due to binocular cooperation. If dissimilar
stimuli were used, retinal rivalry would have been
provoked. In patients with strabismus, rivaly inevitably
VISUAL PERCEPTION IN STRABISMUS 313
causes suppression of one image. It is therefore

mandatory, to use fusable stimuli which are considered
to be "non dissociating" (Bagolini, 1967), when
qualitative assessments are made on the binocular
sensory status of strabismic patients. If information is
searched on the depth of the sensory status, normal or
anomalous, then "dissociating" (i.e. departing from the
every-day seeing condition of the patient) stimuli can
be used. With my non-dissociating technique, it was
possible to find in large-angle esotropia wide
suppression scotomas in the central field of the
deviated eye. In the periphery there is ARC. Small-angle
exotropia patients behave as those with small-angle
esotropia. Patients with large-angle exotropia show
suppression scotomas which override the mid-line in
the center and ARC in the periphery. The hemianopic
suppression scotomas described by Jampolsky (1952) in
exotropia are simply an experimental artifact.
Recent electrophysiological evidence indicates that

these psychophysical findings are an expression of
binocular cortical integration in strabismus. Visual
evoked responses were used to assess this cooperation
(Campos and Chiesi, 1983).
As said, suppression and ARC usually coexist in

large-angle strabismus, although in small-angle
strabismus there may be only ARC. This has been known
since Bielschowsky (1900) and Tschermak - Seysenegg
(1952). Sometimes suppression and ARC coexist also in
small-angle deviations. In the latter group, however,
after an initial stage of pure suppression, followed by
the coexistence of suppression and ARC, the only anti-
diplopic mechanism left may be ARC. Travers (1938)
brought forward the concept that suppression is a
prerequisite for the establishment of ARC. Harms (1938)
even considered the existence of regional suppression
scotomas as a proof of ARC. Using non-fusable test-
targets Hallden (1982) and Herzau (1980) showed
suppression scotomas in the fovea and in that part of
the peripheral retina of the deviated eye receiving the
same image as the fovea (fixation point scotoma).
However these results are in my opinion an experimental
artifact, as dissociating testing techniques eliciting
retinal rivalry have been used. This is why my results
(Campos, 1982) contradict those described above.
Various entities differently defined have

superimposable sensorial characteristics as far as
314 E.C. CAMPOS
binocular cooperation is concerned with small-angle

strabismus. They include microstrabismus, microtropia,
monofixation syndrome. One could conceive of a
spectrum that ranges from normal binocular vision with
bifixation at one end by means of Ogle's fixation
disparity, the various manifestations of microtropia, to
small-angle esotropia at the other end (von Noorden,
1985). Subnormal binocular vision ( de Decker and Haase,
1976) should be mentioned here. It is present in cured
esotropia in which ortotropia, i.e., parallelism of the
two eyes, is achieved , but stereopsis is lacking and
normal correspondence with occasional suppression may be
found.
Anomalous binocular vision is kept stable by

anomalous vergence movements which have been discussed
in another session of this Symposium by Bagolini and
Schor.
Probably proprioceptive information from the

extraocular muscles may play a role in the stabilization
of the eye position as well. The role of proprioception
will be discussed later on in this session.
Apart from the presence of a binocular single

perception, and of a motor fusion mechanism, patients
with strabismus and ARC have other functional
advantages. Campos (1983) reported patients with ocular
torticollis and a manifest deviation even with the
compensatory head position. Such patients have ARC in
the position of torticollis and suppression of the
deviated eye in other gaze positions. Clearly, they
prefer ARC over suppression and will keep their head
turned or tilted to take advantage of this sensorial
adaptation. In a recent study Campos et al. (1987)
showed also that esotropic patients with ARC are able to
judge distances better than those with suppression and
statistically not differently from normal subjects.
Binocular sensory adaptations in strabismus

disappear instantaneously when one eye is closed. They
are therefore purely binocular phenomena. The visual
perception of strabismics has been analysed also under
monocular conditions. Here, alterations are found only
in the presence of amblyopia, which is the consequence
not only of suppression as usually accepted, but also
of ARC (Campos, 1980).
AMBLYOPIA
There is no convincing evidence that strabismic

amblyopia affects the visual perception of strabismic
patients. In fact, the everyday seeing condition of a
patient is with both eyes open. Therefore, the fovea of
one of the two eyes is definitionwise not used or mis-
used. A perceptual problem which may be encountered in
the presence of amblyopia under binocular conditions, is
the change of localization which takes plane with change
of fixation. This can happen spontaneously or may be
induced forcedly by compelling the habitually deviated
eye to fixate. In this case there may be either a
persistence of anomalous localization (ARC) or, if ARC
is not deeply rooted, the patient will revert to normal
correspondence and may or may not experience diplopia,
according to the characteristics of the suppression
scotoma which has to shift to the fellow eye (von
Noorden, 1985).
The perceptual characteristics of amblyopia will be

discussed in another group of presentations (Levi,
Haase). Here, I would simply outline the importance of~
contrast sensitivity in assessing the monocular
function. In this way differences between strabismic and
anisometropic amblyopia have been suggested. ( Levi and
Harwerth 1977, 1978; Hess et al. 1980; Lundh and
Lennerstrand, 1983) Suprathreshold testing of contrast
sensitivity appears a promising method for assessing
objectively visual function by means of visual Evoked
Responses (Lennerstrand and Franzen, 1986).
Localization and partition anomalies seem also to be
related with strabismic amblyopia (Bedell and Flam,
1981).
I would like to stress that the analysis of

characteristics of amblyopia should include testing
under binocular conditions as well, if informations have
to be collected on the essence of the visual act of
strabismic patients. Monocular testing provides only
with partial information.
ADULT ONSET STRABISMUS
A short comment on visual perception of patients

with adult onset strabismus is appropriate. In these
patients suppression and ARC do not develop. However,
some of them are not disturbed by diplopia and/or
confusion. This situation, although well known to
316 E.C. CAMPOS
clinicians is not clear. It may be defined as
unawareness of diplopia and could be due to a higher
association level mechanism.
THE CONCEPT OF DISSOCIATION IN TESTING
A final section of this review should be devoted to

the diagnosis of suppression and ARC respectively. The
concept of "dissociation" as applied to testing of the
binocular sensory status has been already introduced.
Any finding is therefore stronghly related to the
testing procedure used. It is still a matter of debate
whether "non-dissociating" or stronghly dissociating
tests should be used in order to assess qualitatively
the sensory status of the strabismic patient. I believe
that non-dissociating tests should be prefered because
they match the casual seeing condition of the patient
and inform also on weakly established adaptive phenomena
(Campos, 1978). There is not a general agreement though
on this point. Hopefully, during this symposium this
controversy will be solved. Being as it may, it is
anyway mandatory when referring to suppression and/or
ARC, to specify the testing procedure used for assessing
the sensory status. This is most important also for
basic scientists because they may be unwillingly using
an heterogeneous group of subjects for their
experiments.
REFERENCES
Bagolini, B. (1967). Anomalous correspondence: defini-

tion and diagnostic methods. Doc. Ophthalmol., ~. 346-
386, 1967.
Bedell, H.E. and Flom, M.C. (1981). Monocular spatial

distortion in strabismic amblyopia. Invest. Ophthalmol.
Visual Sci., lQ, 263-268.
Bielschowsky, A. (1900). Untersuchungen ueber das Sehen

der Schielenden. v. Graefe's Arch.Ophthalmol., ~, 406-
509.
Campos, E.C. (1978). On the reliability of some tests

for the binocular sensory status in strabismic pa-
tients. J. Ped. Ophthalmol. & Strabismus, li• 8-14.
Campos, E.C. (1980). Some functional abnomalities in
amblyopia. Trans. Ophthal. Soc. U.K.,~, 413-418.
Campos, E.C. (1982). Binocularity in comitant stra-

bismus: binocular visual field studies. Doc. Ophthal-
mol., ~' 249-281.
Campos, E.C. (1983). Ocular torticollis. Int. Ophthal-

mol., _§_, 49-53.
Campos, E.C., Aldrovandi, E. and Bolzani, R. (1987).

Distance perception in strabismic patients. Submitted
for publication to Doc. Ophthalmol.
Campos, E.C. and Chiesi, C. (1983). Binocularity in

comitant strabismus: II. Objective evaluation with vis-
ual evoked responses. Doc. Ophthalmol., ~' 277-293.
de Decker, W. and Haase, W. (1976). Subnormales Binok-

ularsehen - Versuch einer Einteilung des Mikrostra-
bismus. Klin. Mbl. Augenkeilk. 169, 182-194. Hallden, u.
(1982). Suppression scotoma in concomitant strabismus
with harmonious anomalous correspondence. Acta Ophthal-
mol., _§_Q, 828-834.
Harms, H. (1938). Ort und Wesen der Bildhemmung bei

Schielenden. v. Graefe's Arch. Ophthalmol., 138, 149-
210.
Herzau, V. (1980). Untersuchungen ueber das binokulare

Gesichstsfeld Schielender. Doc. Ophthalmol., i1' 221-
284.
Herzau, V. (1981). Anomales beidaugiges Einfachsehen bei

grossen Schielwinkel. Untersuchungen mit Schweiftest
Glaesern. Klin. Mbl. Augenheilk., 178, 81-84.
Hess, R.F., Campbell, F.W. and Zimmern, R. (1980). Dif-

ferences in the neural basis of human amblyopia: the
effect of meanluminances. Vision Res.,~' 295-305.
Hess, R.F. and Howell, E.R. (1978). The luminance-depen-

dent nature of the visual abnormality in strabismic
amblyopia. Vision Res.,~' 931-936.
Jampolsky, A. (1955). Characteristics of suppression in
strabismus. Arch. Ophthalmol., 2!, 683-696.
Lennerstrand, G. and Franzen, 0. (1986). Contrast sensi-

tivity testing with Suprathreshold stimuli: a comparison
of psychophysical and electrophysiological results. In
International Strabismological Association Proceedings
318 E.C. CAMPOS
of the Fifth Meeting (E.C. Campos ed.), ETA, Modena,

!:):""93-1 0-1-.-
Levi, D.M. and Harwerth, R.S. (1977). Spatia-temporal

interaction in anisometropic and strabismic amblyopia.
Invest. Ophthalmol. Visual Sci. ~, 90-95.
Levi, D.M. and Harwerth, R.S. (1978). Contrast evoked

potentials in strabismic and anisometropic amblyopia.
Invest. Ophthalmol. Visual Sci. l2• 571- 574.
Lundh, B.L. and Lennerstrand, G. (1983). Effects of

amblyopia therapy on contrast sensitivity as reflected
in the visuogram. Acta Ophthalmol., £l, 431-446.
Mehdorn, E. (1982). On the nature of suppression scoto-

mas in strabismus. In Strabismus, International Sympo-
sium on Strabismus (M.C. Boschi and R. Frosini £ds.).
Florence, p. 195-198.
Noorden, G.K. von (1985). Burian-von Noorden's Binocular

Vision and Ocular Motility. Theory and Management of
Strabismus. c.v. Mosby, St. Louis.
Ogle, K.N., Mussey F. and Prangen, A. de H. (1949).

Fixation disparity and fusional processes in binocular
single vision. Am. J. Ophthalmol., 12, 1069-1087.
Schor, C.M., Bridgeman, B. and Tyler, C.W. (1983).

Spatial characteristics of static and dynamic stereoac-
uity in strabismus. Invest. Ophthalmol. & Visual Sci.,
~, 1572-1579.
Travers, T. a B. (1938). Suppression of vision in squint

and its association with retinal correspondence and
amblyopia. Br. J. Ophthalmol.,~, 577-604.
Tschermak-Seysenegg, A. von (1952). Introduction to

Physiological Optics. (Translated by P. Boeder). ~c.
Thomas, Springfield. Ill.
25
ROLE OF OCULOMOTOR PROPRIOCEPTION IN THE
VISUAL SYSTEM OF THE CAT
L. MAFFEI and A. FIORENTINI
INTRODUCTION
The oculomotor muscles of various mammals contain

proprioceptive receptors. In the cat these have been described as
spiral endings (Bach-y-Rita and Ito, 1966) and more recently as
palisade receptors (Alvarado Mallart and Pincon-Raymond, 1979).
Primates, including man have oculomotor proprioceptors in the form
of muscle spindles as well as myotendinous receptors (Ruskell,
1978; Richmond et al., 1984). The majority of proprioceptive
fibers run extraorbitally in the ophthalmic branch (OB) of the Vth
cranial nerve (Batini et al., 1975). The central connection of
proprioceptive afferents is not well known anatomically. However
recent evidence indicates that the first order neurons are
confined to the semilunar ganglion both in cat and monkey (Porter
and Spencer, 1982; Porter et al., 1983). In the monkey the
brainstem terminations of these neurons are at the level of the
spinal trigeminal and cuneate nuclei (Porter, 1986).
In spite of this scarse anatomical knowledge, there is

abundant electrophysio1ogical evidence indicating that
proprioceptive afferents project to many visual centers, i.e.
reach the reticular formation (Fi1lenz, 1955), the superior
colliculus (Rose and Abrahams, 1975; Donaldson and Long, 1980),
the cerebellum (Fuchs and Kornhuber, 1969) the lateral geniculate
body (Donaldson and Dixon, 1980) and the visual cortex (Buisseret
and Maffei, 1977).
There have been only a few experiments in the past pointing

to a role of oculomotor proprioception in visual functions. We
have approached this problem in the cat in a series of
electrophysiological and behavioural experiments. Our approach has
been to investigate the effects of the section of the OB of the
Vth nerve electrophysiologically and behaviourally.
319
320 L. MAFFEI and A. FIORENTINI
ELECTROPHYSIOLOGICAL EXPERIMENTS
Recording of~ Movements in Adult Cats with Unilateral Section

of the OB of the Vth Nerve
Electrooculographi c recordings of eye movements in alert

adult cats with chronically implanted electrodes have shown that
after unilateral proprioceptive deafferentation, the eye
ipsilateral to the section can display pendular movements in
complete darkness. This suggests a possible role of proprioception
in maintaining a stable position of the eye in the absence of
visual cues. Visually triggered saccades and pursuit eye movements
are not grossly affected (Fiorentini and Maffei, 1977).
Recording from Single Neurons in Cortical Area 17 of Kittens

Raised with Unilateral Section of the OB of the Vth Nerve
The OB of the Vth nerve was cut in kittens 5-6 weeks old and
the animals were then raised in a normal visual environment until
the 4th month of age, when recording from the striate visual
cortex was performed. The proportion of cortical neurones that
could be activated by a stimulus in either eye was found to be
much lower in these animals than in normal controls. Sixty percent
of the neurons were found to respond to one eye only (Maffei,
1979). These findings have been confirmed and extended by Trotter
et al., (1983).
Evidence both from experiments in animals with one eye

surgically immobilized and kept in darkness (Maffei and
Fiorentini, 1976) and from preliminary experiments in animals with
unilateral proprioceptive deafferentation, indicate that even if
these manipulations are performed in adult animals, they produce a
decrease in the proportion of binocular cortical cells.
BEHAVIOURAL EXPERIMENTS
Orienting Behaviour
The role of oculomotor proprioception in the orienting

behaviour was investigated in adult cats trained to jump from a
start box to a luminous target at various possible locations in
the right or left visual hemifield (Fiorentini et al., 1982). The
experiments have shown that the jumping performance undergoes
substantial and long lasting modifications after unilateral
section of the OB of the Vth nerve: a systematic bias is
introduced in the jump direction, that is ipsiverse to the side of
the section and present in jumps guided binocularly or monocularly
with either eye. Control experiments permit to interpret these
findings as due to the imbalance introduced into the
proprioceptive system by the asymmetric proprioceptive
deafferentation.
OCULOMOTOR PROPRIOCEPTION IN THE CAT 321
Depth Discrimination and Stereoacuity
The role of oculomotor proprioception in a visuo-motor task

implying depth discrimination was investigated in adult cats with
a behavioural jumping stand technique (Mitchell et al., 1979). In
cats with unilateral section of the OB of the Vth nerve the
threshold for binocular depth discrimination was found to be much
increased with respect to the presurgical threshold (Fiorentini
et al., 1985).Control experiments showed that postsurgical visual
acuity was normal in either eye. No obvious sign of strabisms was
observed in the deafferented cats. The postsurgical deterioration
in depth discrimination therefore should be accounted for by the
unilateral disruption of proprioceptive information. This implies
that oculomotor proprioception plays a role either in the control
of motor behaviour in the animal required to perform the
discrimination task, or in the sensory processes that lead to
depth perception, or both. In order to distinguish between these
possibilities, we carried out a second behavioural experiment in
which the stereoacuity of cats was measured with an operant
conditioning technique that does not require a motor orienting
response from the animal.
In cats with unilateral proprioceptive deafferentation,

stereoaocuity was largely impaired and did not exceed postsurgical
monocular depth sensitivity (Fiorentini, Cenni and Maffei, 1986).
Control experiments indicate that the loss in stereoacuity is not
likely to be caused by an impaired ability of the cat to perform
correct disjunctive eye movements. Our finding therefore points to
an involvement of proprioceptive signals in the processes
subserving depth perception, and is supported by the
electrophysiological results reported above that show a reduction
in binocularly activated neurons in animals following the section
of the ophthalmic branch of the Vth nerve.
DEVELOPMENT OF PROPERTIES OF VISUAL CORTICAL CELLS AND OCULOMOTOR

PROPRIOCEPTION
Trotter, Gary-Bobo and Buisseret (1981) and Gary-Bobo,

Milleret and Buisseret (1986) have investigated the role of
oculomotor proprioception in the development of functional
properties of visual cortical cells in the cat.
The paradigm of their experiments was based on previous

observations of the effects of visual deprivation (dark rearing)
and of subsequent brief exposure to a normal visual environment on
the orientation selectivity of cortical cells in area 17. Indeed,
while in normally reared kittens most cortical cells are
orientation selective by the age of 6 weeks, in dark-reared
kittens, most cells lack a clear orientation selectivity. However,
6 hours of normal visual experience at this age are sufficient to
restore normal cortical properties.
Th.is restoration does not occur, i f the dark reared kittens

are bilaterally deafferented from oculomotor proprioception before
the 6 hours of visual exposure.
Another instance of a role of oculomotor proprioception in

gating developmental processes in young animals is provided by an
experiment by Buisseret and Singer (1983) on monocularly deprived
kittens with bilateral deafferentation of oculomotor
proprioception. In these animals, the deafferentation prevents the
loss of binocularity of visual cortical units that is usually
found in intact monocularly deprived kittens.
Recently Hein and Diamond (1983) reported evidence on the

role of oculomotor proprioception on the development of visually
guided behaviour. Bilateral deafferentation of proprioceptive
fibers prevents dark reared kittens from developing normal visuo-
motor behaviour in spite of subsequent exposure to normal visual
experience: these kittens fail to give correct orienting visuo-
motor responses.
COMMENTS
Investigation of the role of oculomotor proprioception in

v1s1on has been largely neglected for a long time, following the
classical controversy between supporters of outflow and inflow
theories of visual localization. The increasing knowledge of the
many anatomical projections of proprioceptive signals to various
sites in the brain and in particular to high visual structures as
the superior colliculus and the visual cortex, requires to

reexamine the role of oculumotor proprioception in vision. Indeed
the presence of proprioceptive signals in high visual centers is
not likely to be devoided of any physiological significance.
The results in the cat reported here and those of other

laboratories start to spread some light on some possible
functional roles of proprioception in vision. Of particular
relevance is the finding on the deterioration of depth perception
in the cat.
In man very scarce attention has been devoted to this

problem, after the ingenious experiments in which Skavenski
(1972), by applying a load to the eye, provided evidence for the
role of oculomotor proprioception in the control of eye position
in the dark. Recently however, evidence for the role of oculomotor
proprioception in processing information about eye-position was
obtained by Steinbach and Smith (1981). They tested open-loop
pointing in strabismic patients who had undergone surgery for the
first time to re-align their squinting eye. They found that the
pointing responses were very accurate, although the patients had
not had any post-operative visual experience until the moment of
testing. This proved that they had information about the new
position of their eye, that is likely to be derived from

oculomotor afferent.
Other evidence for a role of proprioception in human visuo-
motor responses has been recently obtained by Campos et al (1986)
in experiments on human patients with unilateral Herpes Zoster
Ophthalmicus, a viral affection of the trigeminal nerve. These
patients have reduced or absent corneal sensitivity, indicating a
serious involvement of the fibers of the ophthalmic branch of the
fifth nerve. When tested with the affected eye for open-loop
pointing to a bright source in the dark, these patients showed
systematic pointing errors, while pointing was correct when
viewing with the normal eye.
These findings indicating that in man oculomotor

proprioception plays a role in providing information about eye-
position are in line with the results obtained in the animal,
showing that proprioceptive unilateral deafferentation produces a
bias in visuomotor orienting responses (Fiorentini, et al., 1982).
Campos et al. (1986) also tested their patients for global

stereopsis and did not found any impairment (TNO test). This is at
variance with the results obtained in cats with proprioceptive
deafferentation: in these animals stereoacuity is considerably
impaired. This may simply reflect a difference between the two
species. However the discrepancy is not really conclusive in that
depth perception was tested in different ways in the human and
animal experiments: the former was a test for global stereopsis,
while the second was a test of stereoacuity with single bar
elements. These are indeed two different aspects of depth
perception which seem to have at least partly different
neurophysiological substrates (Poggio et al., 1985; Julesz, 1971).
In addition, experiments in monkeys (Cowey, 1983) show that
ablation of the portion of cortical areas V1 or V2 covering the
central 5 deg of visual field cause a considerable impairment of
local stereopsis (stereoacuity measured with single line elements)
but do not affect global stereopsis.
References
Alvarado-Mallart, R.M. and Pincon-Raymond, M. (1979). The

palisade endings of cat extraocular muscles: a light and
electron microscope study. Tissue and cells, 11, 567-584.
Bach-y-Rita, P. and Ito, F. ( 1966) . Properties of stretch

receptors in cat extraocular muscles. J. Physiol. , ( Lond.)
186, 663-688.
Batini, c., Buisseret, P. and Buisseret-Delmas, C. (1975).

Trigeminal pathway of the extrinsic eye muscle afferents in
cat. Brain Res., ~ 74-78.
Buisseret, P. and Maffei, L. (1977). Extraocular proprioceptive

projections to the visual cortex. Exp. Brain Res., ~ 421-
425.
Buisseret, P. and Singer, W. (1983). Proprioceptive signals from
extraocular muscles gate experience-dependent modifications of
receptive fields. Exp. Brain Res., 51, 443-450.
Campos, E.C., Chiesi, C. and Bolzani, R. (1986). Abnormal spatial

localization in patients with Herpes Zoster Ophthalmicus.
Arch. Ophthalmol., 104, 1176-1177.
Cowey, A. (1985). Disturbances of stereopsis by brain damage. In

Brain Mechanisms and Spatial Vision. (eds. D.J. Ingle, M.
Jeannerod and D.N. Lee). Martinus Nijhoff Publ. Dordrecht,
Holland.
Donaldson, I.M.L. and Dixon, R.A. (1980). Excitation of units in

the Lateral Geniculate and Contiguous nuclei of the cat by
stretch of extrinsic ocular muscles. Exp. Brain Res., 38, 245-
255.
Donaldson, I.M.L. and Long, A.C. (1980). Interaction between

extraocular proprioceptive and visual signals in the superior
colliculus of the cat. J. Physiol., (Lond.) 298, 85-110.
Fillenz, M. (1955). Responses in the brainstem of the cat to

stretch of extrinsic extraocular muscles. J. Physiol., (Lond.)
128, 182-189.
Fiorentini, A., Berardi, N. and Maffei, L. (1982). Role of

extraocular proprioception in the orienting behaviour of cats.
Exp. Brain Res., 48, 113-120.
Fiorentini, A., Cenni, M.C. Maffei, L. (1986). Impairment of

stereoacuity in cats with oculomotor proprioceptive
deafferentation. Exp. Brain Res.,~ 364-368.
Fiorentini, A., Maffei, L., Cenni, M.C. and Tacchi, A. (1985)

Deafferentation of oculomotor proprioception affects depth
discrimination in adult cats. Exp. Brain Res., 59, 296-301.
Fiorentini, A., Maffei, L. (1977). Instability of the eye in the

dark and proprioception. Nature, 269, 330-331.
Fuchs, A. and Kornhuber, H.H (1969). Extraocular muscle afferents

to the cerebellum of the cat. J. Physiol., (Lond.) 200, 713-
722.
Gary-Bobo, E. , Milleret, C. and Buisseret, P. (1986). Role of eye

movements in developmental processes of orientation
selectivity in the kitten striate cortex. Vision Res., 26,
557-567.
Hein, A. and Diamond, R. (1983), Contribution of eye movement to

the representation of space. In Spatially oriented behaviour.
(eds. A. Hein and M. Jeannerod) Springer Verlag New York.
Julesz, B. (1971). Foundations of Cyclopean Perception. Univ.
Chicago Press.
Maffei, L. (1979). Possible role of oculomotor proprioception in

the cat. Trans. Ophthal. Soc., UK 99, 375-376.
Maffei, L. and Fiorentini, A. (1976). Asymmetry of motility of the

eyes and change of binocular properties of cortical cells in
adult cats. Brain Res., 105, 73-78.
Mitchell, D.E., Kay, M. and Timney, B. (1979). Assessment of depth

perception in cats. Perception ~. 389-396.
Poggio, G.F., Motter, B.C., Squatrito, M.S. and Trotter, Y.

(1985). Responses of neurons in visual cortex (Vl and V2) of
the alert macaque to dynamic random-dot stereograms. Vision
Res., 25, 397-406.
Porter, J.D. and Spencer, R.F. (1982). Localization and morphology

of cat extraocular muscle afferent neurons identified by
retrograde transport of Horseradish Peroxidase. J. comp.
Neural., 204, 56-64.
Porter, J.D. (1986). Brainstem termination of extraocular muscle

primary afferent neurons in the monkey. J. comp. Neurol.,247
133-143.
Porter, J.D., Guthrie, B.L. and Sparks, D.L. (1983). Innervation

of monkey extraocular muscles: localization of sensory and
motor neurons by retrograde transport of Horseradish
Peroxidase. J. camp. Neural., 218, 208-219.
Richmond, F.J.R., Johnston, W.S.W., Baker, R.S. and Steinbach,

M.J. (1984). Palisade endings in human extraocular muscles.
Invest. Ophthalm. Vis. Sci., 25, 471-476.
Rose, P.K. and Abrahams, V.C. (1975). The effect of passive eye
movement on unit discharge in the superior colliculus of the
cat. Brain Res., 97, 95-106.
Rus~ell, G.L. (1978). The fine structure of innervated

myotendinous cylinders in extraocular muscles of rhesus
monkeys. J. Neurocytol., Z 693-708.
Skavenski, A.A (1972). Inflow as a source of extraretinal eye

position information. Vision Res., 12, 221-229.
Steinbach, M.J. and Smith, D.R. (1981). Spatial localization after

strabismus surgery: evidence for inflow. Science, ~. 1407-
1408.
Trotter, Y., Gary-Bobo, E. and Buisseret, P. (1981). Recovery of

orientation selectivity in kitten primary visual cortex is
slowed down by bilateral section of ophthalmic trigeminal
afferents. Develop. Brain Res., !• 450-454.
Trotter, Y. Fregnac, Y. and Buisseret, P. (1983). Sinergie de la

vision et de la proprioception extraoculaire dans les
mechanismes de plasticite' fonctionelle du cortex visuel
primaire du chaton. CR Acad. Sci., 296, 665-668.
26
EXTRAOCULAR MUSCLE PROPRIOCEPTION AND VISUAL
FUNCTION: PSYCHOPHYSICAL ASPECTS
MARTIN J. STEINBACH, MARIA A. MUSARELLA

and BRENDA L. GALLIE
ABSTRACT: We have found evidence for oculomotor

proprioception by measuring spatial localization (using open-loop pointing
responses) in patients before and after they have had surgery on their eye
muscles. Proprioceptors located at the musculotendinous junction, the
palisade endings, appear to be playing an important but undefined role in
supplying eye position information. For both enucleation and strabismus
surgical patients, the post-operative localization changes are manifested
in the constant, but not the variable, pointing errors. These changes in
accuracy rather than precision indicate that the underlying inflow
calibration signal is a slowly-acting one.
INTRODUCTION: The evidence that proprioception provides

knowledge of eye position is accumulating and becoming very convincing
(e.g., Maffei, this Symposium; Steinbach, 1987). In particular, studies of
people undergoing surgery to correct strabismus or having some
abnormality of the afferent pathway for proprioception are especially
instructive (e.g., Mitsui, 1986; Campos et al., 1986, and this Symposium).
We will briefly review our evidence from strabismus surgery patients
indicating the presence of an inflow signal which we believe depends on
the palisade endings located at the musculotendinous junction. We will
then present new data and analyses from enucleated and other patients
which indicate that the proprioceptive signal is a slowly-acting one.
EVIDENCE FOR INFLOW AND THE ROLE OF

THE PALISADE ENDINGS
FIRST-OPERATION VS MULTIPLE OPERATIONS FOR

CORRECTION OF STRABISMUS: Our original evidence of a
proprioceptive (inflow) eye position signal was obtained by testing
strabismus patients before and after surgery (Steinbach and Smith, 1981 ).
327
328 M.J. STEINBACH, M.A. MUSARELLA and B.L. GALLIE
Using an "open-loop" pointing task, that is, one in which there is no

feedback about pointing accuracy, we found that patients having surgery
for the first time could point to targets with surprisingly little error using the
operated eye the instant it was unbandaged. If there were no recalibration
for the changed position of the eye in the orbit, there should have been an
error that matched the amount of rotation of the eye. In the absence of
vision, the only source for information about the changed position of the
eye had to be proprioception. We also tested another group of patients
having the same muscles operated on for the second or more time. These
patients mislocalized targets after surgery by the amount of the rotation of
the eye, indicating the absence of a proprioceptive signal.
Testing the non-operated eye of these two groups of patients also

revealed information about the use or non-use of a proprioceptive signal.
On average, the newly-operated patients showed large shifts in localizing
responses using this untouched eye, while the re-operated patients
showed hardly any at all. It is known that we use the positions of both
eyes when localizing targets, even when only one eye is seeing (Ono and
Weber, 1981 ). The altered position of the operated eye was therefore
being monitored in the newly-operated patients, but not in the multiply-
operated patients.
This startling difference between the two groups of patients led to

our anatomical examination of the region of the surgery and the
subsequent uncovering of the palisade endings at this location (Richmond
et al., 1984). We presume that these receptors are important in some not
yet understood way for supplying the position information that was present
in the newly-operated patients and missing (due to scarring or repeated
resections) in the multiply-operated patients. Our new studies described
below also implicate the musculotendinous region as important for inflow.
Bock and Kommerell (1986) failed to replicate the Steinbach and

Smith (1981) findings. A significant difference in anesthetic procedures
used during surgery may account for this. Kommerell (personal
communciation) used retrobulbar injections of topical anesthetic whereas
our patients were under general anesthesia. The retrobulbar block,
although limited in anesthetic duration, could have had a longer-term
effect on proprioceptive afferents.
MYOTOMY VS RECESSION EFFECTS: Steinbach et al. (1987)

measured localizing responses in two groups of children having different
(first-time) muscle weakening procedures to correct esotropia. In the
standard recession procedure the muscle is cut very close to the insertion
while in the marginal myotomy procedure the musculotendinous region is
first crushed, and this is followed by half-muscle-width cuts taken from
opposite sides (von Noorden, 1985). Any proprioceptors in this region
should be disrupted more by the myotomy than by the recession
procedure.
PSYCHOPHYSICS OF PROPRIOCEPTION 329
Using a computer game to measure pointing responses in children

{Steinbach et al., 1986), we have found that when testing the operated eye
the recession causes less of a localizing shift than does the myotomy.
When testing the non-operated eye, the shift is larger in the recession
patient than in the myotomy patient.
These results indicate that following the recession there is relatively

accurate information about the surgically-rotated eye's new position, thus
accounting for the small shift when testing the operated eye and the large
shift when using the non-operated eye. The myotomized children do not
have this inflow information, and this is reflected in the larger shift found
when using the operated eye, and the small shift when using the non-
operated eye. We assume that the greater disruption of the palisade
endings by the myotomy procedure has led to the less-useful signal about
changed eye position. These results parallel those found in the newly-
operated and multiply-operated patients described earlier.
It is not clear how the palisade endings could be supplying position

information. They appear to be registering the tension of individually-
contracting muscle fibers {Richmond, et al., 1984). By itself, the tension at
a single insertion point is ambiguous as to absolute eye position
{Steinbach, 1987). Future studies will have to address this question.
EVIDENCE FOR THE SLOW TIME COURSE OF

PROPRIOCEPTIVE RECALIBRATION
EFFECTS OF ENUCLEATION: If we use proprioceptive information

about the positions of both eyes in judging egocentric visual direction,
what will happen to a person who has had an eye removed? The trauma
to orbital fascia and muscles, as well as the markedly altered lengths and
tensions in the extraocular muscles, could produce a very "noisy" signal
from any proprioceptors found in these tissues. One might expect,
therefore, that such surgery would lead to a change in the precision of
localizing responses made using the remaining eye, i.e., an increase in
response variability with no change in the constant error1. It was our
expectation, therefore, that patients undergoing enucleation surgery
would, on average, point to targets in the same direction that they did
before surgery, but that the scatter of their pointing responses would be
greater. What we found was just the opposite: no change in variability {or
precision) in all of the patients and some indication of a change in
constant error {or accuracy) in many of the patients.
i1n a localization task, or in any repeated test of human performance,

there are two basic parameters: accuracy and precision. Accuracy refers
to the averaged localization tendency and is also described as the
constant error. Precision refers to the dispersion of these localization
responses around their mean value, and is also referred to as the variable
error {see Howard, 1982, pp. 3-4).
Patients were required to point to a visual target with the unseen

hand (they were all right-handed). The apparatus and procedures were
the same as used by Steinbach and Smith (1981 ). Measurements were
taken at least once the day before surgery and then after the enucleation
as soon as the patient was comfortable enough to do the test, which was
usually 48, but sometimes as early as 24 hours after the operation.
Subsequent testing depended on the patient's schedule, but usually two,
and for half the patients, three, post-op measures were obtained within the
first week.
Thirteen surgical patients were assessed (see Table 1). Twelve

had enucleations (eleven were for malignant melanomas, and one was for
removal of a painful blind eye). An evisceration was done on the
thirteenth patient for a traumatically-blind eye. Six normal subjects were
tested in a control procedure; they were all binocularly normal young
adults. The controls were tested for three consecutive days both prior to
and immediately following 48 hours of monocular occlusion.
Results: The first question to be answered concerned the nature of
the change in localizing responses produced by the surgery. We were
expecting changes in variable error (or precision) in the reponses. This
would be indicated by an increase in the standard deviation of the pointing
responses over the days of testing after the surgery, without any
concomitant change in accuracy (or constant error). Table 2 shows the
averaged standard deviations for the patients on the pre-op and and
subsequent post-op sessions. These values are within the range of one to
two degress found for normal subjects (Howard, 1982). Clearly, there was
no change in the precision of their pointing responses.
TABLE 1
CHARACTERISTICS OF PATIENTS TESTED
AGE AT VISION
PATIENT SURGERY DIAGNOSIS LE RE
1 70 ANTERIOR CHOROIDAL MELANOMA, RE 20125 HM

2 81 MELANOMARE HM 20/25
3 39 MELANOMA OF CHOROID LE HM 20/25
4 65 MELANOMALE 20/20 HM
6 57 MELANOMARE 20/400 20/20
7 50 MELANOMARE CF 20/20
8 54 MELANOMARE 20/200 20/20
9 16 IRIS MELANOMA LE 20/15 20/15
10 15 PHTHISICAL EYE, LE TRAUMA 20/20 NLP
11 72 MELANOMARE 20/100 20/20
13 09 RE EVISCERATION LP 20/20
•Abbreviations for vision: HM =hand motion; CF =counting fingers; LP = light perception;

NLP- no light perception
TABLE 2
Pre- and Post-Operative Measures of Precision for Enucleated and Strabismic Surgical
Patients (Averaged Standard Deviations of the Open-Loop Pointing Responses Given in
Degrees)
Study Pre-Op Post-Qp Testing Session
2 2 3 4
Enucleates 1.3 (12)* 1.8 (12) 1.8 (12) 1.5 (11) 1.0 (5)
Strabismics: 1.7 (9) 1.5 (9) 1.6 (3)

Steinbach & Smith (1981)
Strabismics: 3.1 (10) 3.2 (10)
Steinbach et al. (1987)
*Numbers in parentheses are the numbers of patients tested in that session.
The patients' pointing accuracy (as measured by their averaged

constant errors in localizing the targets), however, did show changes that
were systematic ones for more than half of the patients, and these can be
seen in Figure 1. For patients 1, 2, 3, 4, 6, and 9 there is a change in
localizing responses which peaks at the second or third post-operative
day and then moves towards recovery to the pre-op levels. The other
enucleated patients do not show this pattern as clearly, and one patient,
number 12, shows no change compared to the pre-op pointing responses.
(The days of testing were primarily determined by the patient's schedule.
Out-of-town patients could be tested only when they returned to Toronto
for follow-up and this accounts for the large gaps seen in between test-
days for some of the patients.)
Could these shifts come about because vision was suddenly

deprived to one eye? Most patients had only counting finger vision or light
perception in the eye to be enucleated (see Table 1) so this is probably
not too important a consideration. But patient #9 had an iris melanoma
and had perfect vision in the eye removed. One could argue that suddenly
rendering a binocularly normal human monocular could cause a shift in
the apparent direction of targets. We tested for this by asking binocularly
normal subjects to wear a patch that totally occluded vision (and most of
the light) for two days. They were tested using the same apparatus and
task for three days both before and again immediately after the occlusion.
Fig. 2 displays the results which show that 48 hours of monocular
occlusion has no or only a trivial effect on the localizing responses for
most of the subjects. Subject 2 is the exception and we have no
explanation for her anomalous data. These control data also show the
relative consistency of the localizing responses over time. With only minor
peturbations, four subjects are very consistent over the six days of testing
(CSs 3-6), while a fifth (CS-1) shows a slight leftward drift.
6
4
2
0
-2
-4
Pre P1 P2 P7
visible target
Pre1 Pre2 P1 P2 P3 P14

of the 13 surgical patients in pointing to
Pre P1 P2 P3 p 12
Fig. 1. Average constant
8
6
4
2
0'-r-~-..--,---,,..---,---,.---
Pre 1 Pre 2 P 1 P 2 P 3 P 6 P 13 Pre P1 P3
TEST SESSION
Fig. 1. Average constant error of the 13 surgical patients in pointing to a visible target
without sight of hand. All were enucleated, except for patient 13, who had an evisceration.
Could the stress of surgery, including the general anesthetic and

the severing of the optic nerve, alter localizing responses? We tested a
nine-year old girl who was undergoing an evisceration procedure on an
eye blinded by blunt trauma six months earlier. In this rarely-performed
procedure, done under a general anesthetic, the contents of the globe are
removed, and this includes scraping off the retina and the long posterior
ciliary nerves. The muscles and orbital fascia are untouched, however.
Her results are in Fig 1 (Patient No. 13). Virtually no change in constant
or variable errors resulted.
J~_./ /
CS1 CS2
~
sa: /--
0
CI:I-
~
=iJ Pre 1
I I I I I I
-4
CI:I Pre 2 Pre 3 P1 P2 P3
I I I I I I
=n--- v-
UJC) Pre 1 Pre 2 Pre 3 P1 P2 P3
I--
~~
..____ .___...
CS3
=!~
CS4
1-1-
(f)(/)
z-:-
0 ·-
ut
wUJ
I I I I I I I I I I I I
C)....J Pre 1 Pre 2 Pre 3 P1 P2 P3 Pre 1 Pre 2Pre 3 p1 P2 P3
~~
:il ..___....
css CS6
Jl~
UJ(f)
------
>-
<( +
~
I I I I I I -4 I I I I I I
Pre 1 Pre 2 Pre 3 p 1 P2 P3 Pre 1 Pre 2 Pre 3 P 1 P2 P3
TEST SESSION TEST SESSION
Fig. 2. Average constant error for six normal control subjects who wore a patch for 48 hours
over one eye. The localizing responses made using the unpatched eye are shown for the
three days preceding the occlusion and the three days following the occlusion.
We believe the changes in localization shown by most patients after

removal of an eye represent the visual system's use of proprioceptively-
derived information from the enucleated eye's orbital tissues and muscles.
It further appears that this information takes two or three days in most
cases to have its maximal effect. Evidence for the slow temporal course of
proprioception comes from those patients who could be tested on the first
post-op day (n = 6). The shift measured on that day was not as big as it
would become on the second or third post-op day, suggesting a
recalibration time constant spanning days. We also believe that the
marked intersubject variability is a consequence of the differing surgical
procedures used by the seven surgeons of this study in cutting and
suturing the eye muscles during the enucleation procedure.
NATURE OF THE LOCALIZATION SHIFT: ACCURACY CHANGES,

NOT PRECISION. In all of our studies to date, whether of patients having
strabismus or enucleation surgery, the change in localizing responses that

we find is always reflected in the accuracy (or constant error) rather than
the precision (or variable error). This was especially surprising when we
tested the enucleated patients because we believed the gross dissection
of muscles and orbital tissues in that procedure would produce a very
noisy proprioceptive signal. Table 2 shows that in the strabismus patients
as well, the pre-op and post-op measures also displayed similar amounts
of variability. For the adults tested in the Steinbach and Smith (1981)
study, the ranges are the same as those found in the enucleates. The
Steinbach et al. (1987) data were collected using a different apparatus
and testing younger children. These slightly larger than adult values are
similar to those found testing other children with the same apparatus
(Steinbach et al., 1986), and most importantly, they do not change pre-
and post-operatively.
Consistent with these findings in surgical patients are those

obtained from patients with disruption to the proprioceptive afferent path.
Campos et al. (1986) reported that standard deviations of open-loop
pointing responses did not exceed 2.5 deg in patients with active herpes
zoster ophthalmicus (a virus which affects the ophthalmic division of the
trigeminal nerve) while there were changes in their pointing accuracy. In
addition, Steinbach (1986b) reported on a single deafferented patient
whose ophthalmic nerve had been sectioned because of trigeminal
neuralgia. This patient showed shifts in his constant, but not variable,
errors of localization following disruption of his binocular vision.
We interpret the finding of no change in variable error as additional

evidence for the stable and long-term nature of the proprioceptive signal.
It is as if inflow provides the underlying signal of eye position from which
the short-term signal about eye position is read out and provided to
consciousness. This short-term signal is, of course, outflow (also called
corollary discharge or efference copy; see Matin, 1986 or Carpenter,
1977). We believe this temporal dichotomy of function, first suggested by
Ludvigh (1952), reconciles the large body of data which supports outflow
(see e.g., Stark and Bridgeman, 1983; Steinbach and Skarf, 1985) with
the vast amount of information about the anatomy and neurophysiology of
eye muscle proprioceptors (Steinbach, 1986b; 1987). The eye muscles
not only have their obvious motoric functions, but they should be thought
of as sense organs as well (Steinbach, 1986a).
ACKNOWLEDGEMENTS: We thank Drs. J.D. Morin, R. Buncic, J.

Cardarelli, W. Macrae, and W. Thompson for allowing us to test their
patients, Eric Steinbach for preparing the figures, Diane Lawless for
preparing the ms, and Maureen Reed and Sharon Melrose for assistance.
We benefitted from the comments of Professors I. P. Howard and H. Ono
on an earlier draft of this paper. We gratefully acknowledge the courage
and generosity of our patients. Supported by NSERC of Canada grant
A7664, NIH EY05960, The Hospital for Sick Children Foundation, the
Department of Ophthalmology, University of Toronto, and Dean R.

Bordessa of Atkinson College.
REFERENCES
Bock 0. and Kommerell G. (1986) Visual localization after strabismus

surgery is compatible with the "outflow" theory. Vision Res. 26, 1825-
1829.
Campos E. C., Chiesi C. and Bolzani, R. (1986) Abnormal spatial

localization in patients with herpes zoster ophthalmicus. Archs
Ophthalmol. 104, 1176-1177.
Carpenter R. H. S. (1977) Movements of the Eyes Pion, London.
Howard I. P. (1982) Human Visual Orientation. Wiley, Toronto.
Ludvigh E. (1952) Control of ocular movements and visual interpretation

of environment. Archs Ophthalmol. 48, 442-448.
Matin, L. (1986) Visual localization and eye movements. In Handbook

of Perception and Human Performance, Volume 1: Sensory
Processes and Perception (Edited by Boff K. R., Kaufman L. and
Thomas. J. P.), pp. 20-1 - 20-45. Wiley, New York.
Mitsui, Y. (1986) Strabismus and the Sensorimotor Reflex.

Excerpta Medica, Tokyo.
Ono H. and Weber E. (1981) Nonveridical visual direction produced by

monocular viewing. J. Exp. Psycho!., Hum. Perc. Perform. 7,
937-947.
Richmond F. J. R., Johnston W. S. W., Baker R. S. and Steinbach M. J.

(1984) Palisade endings in human extraocular muscles. Invest.
Ophthalmol. & Vis. Sci. 25, 471-476.
Stark L. and Bridgeman B. (1983) Role of corollary discharge in space

constancy. Perc. & Psychophys. 34, 371-380.
Steinbach M. J. (1986a) Muscles as sense organs. Arcl:ts Ophthalmol.

104, 1148-1149.
Steinbach M. J. (1986b) Inflow as a long-term calibrator of eye position in

humans. Acta Psychologies 63, 297-306.
Steinbach, M. J. (1987) Proprioceptive knowledge of eye position.

Vision Res, in press.
Steinbach M. J. and Skarf B. (1985) Why illusions of movement from a

passively moved eye depend on the direction of gaze. Binoc. Vision
1, 23-26.
Steinbach M. J. and Smith D. R. (1981) Spatial localization after

strabismus surgery: Evidence for inflow. Science 213, 1407-1409.
Steinbach, M. J., Kirshner, E. L. and Arstikaitis, M. J. (1987) Recession

vs. marginal myotomy surgery for strabismus: Effects on spatial
localization. Invest. Ophthalmol & Vis Sci, in press.
Steinbach, M. J., Kirshner, E. L. and Ono, K. (1986) Interactive computer-

based apparatus for testing eye-hand coordination in children. Binoc.
Vision 1, 203-208.
von Noorden, G. K. (1985) Burian-von Noorden's Binocular Vision

and Ocular Motility (3rd Ed.), St. Louis, Mosby.
27
EXTRAOCULAR MUSCLE PROPRIOCEPTION AND VISUAL
FUNCTIONS: CLINICAL ASPECTS
EMILIO C. CAMPOS
INTRODUCTION
Orientation in space involves many complicated

factors. Most of them are defined on relatively awkward
basis and have been accepted as such for a long time.
The classical assumption has been that no evidence
exists to suggest that proprioception from the eye
muscles is active in spatial localization, stereopsis,
or the interpretation of motion on the retina (Irvine
and Ludvigh, 1936; von Noorden, 1985). Moreover some
data were interpreted as evidence that the eye lacks
position sense (Ludvigh, 1952). Nevertheless, a model
has been proposed (Ludvigh, 1952 b) which combines
centrifugal with centripetal information. The visual
cortex is assumed to be a short term indicator for
visual orientation (outflow information) . Information
from the eye muscles is used as a calibration system for
the short-term mechanism and is employed in long-term
conditions (inflow information).
The existence of proprioceptive organs in the human

extra-ocular muscles is anatomically well established
(Daniel, 1946; Cooper and Daniel, 1949). Palisade
endings have been identified in the scleral distal
portion of these muscles (Richmond et al., 1984).
Evidence exists from animal studies that proprioceptive
information from the extra-ocular muscles travels in the
ophthalmic branch of the trigeminal nerve (Batini and
Buisseret, 1974; Manni and Bortolami, 1982). The
perikaria of first-order neurons are in the semilunar
ganglion. The peripheral nerve process innervates the
muscle spindle, whereas the central nerve processes
337
338 E.C. CAMPOS
terminate in the ipsilateral portion of the spinal

trigeminal nucleus and in the main sensory trigeminal
nucleus. Second-order neurons have been identified in
these nuclei and project on the cerebellum and the
mesodiencephalic areas.
Proprioception seems to be involved in the control

of saccadic eye movements through the cerebellum (Fuchs
and Kornhuber, 1969). It has also been demonstrated
that the oculomotor system in humans takes advantage of
inflow information (Skavenski, 1972) because appropriate
corrective eye movements compensate for passively
applied loads in the dark. Outflow prevails over inflow,
however, in the case of conflict (Skavenski et al.,
1972). Other animal experiments have shown excitation of
cells of the LGN by streching the extra-ocular muscles
(Donaldson and Dixon, 1980).
More recently, a definite role of oculomotor

proprioception in space perception in cats was
demonstrated (Maffei and Fiorentini, 1984). Details on
this work has been provided by Dr. Maffei in this
Session.
Steinbach and Smith (1981) pointed out the role of

inflow information in humans, by examining patients
with strabismus before and after surgery. Dr. Steinbach
provided insite on the psychophysical aspects of space
perception earlier in this Session.
I shall discuss briefly the clinical role of

proprioception on space perception. Informations on this
aspect is quite limited, however.
HERPES ZOSTER OPHTHALMICUS EXPERIMENTS
An inflammatory lesion of the ophthalmic branch of

the trigeminal nerve, without ocular motility disorders,
has been shown to cause anomalies in space localization.
Patients were asked to point to targets without sight of
their hand (open-loop experiment) . Pointing errors
tended to disappear with remission of the disease
(Campos and Chiesi, 1985; Campos et al., 1986). This
series of experiments has been instrumental for showing
that:
a. proprioceptive information travels also in humans
along the trigeminal nerve.
b. spatial orientation depends also from peripheral
information.
PROPRIOCEPTION AND VISUAL FUNCTIONS 339
SURGERY FOR STRABISMUS
Two aspects are involved, i.e., monocular and

binocular space perception.
Monocular post-operative pointing errors.
Steinbach and Smith (1981) and I (Campos et al.,

1986) showed that after strabismus surgery there are
monocular pointing errors attributable to peripheral
information. If visual experience would have been the
only factor involved in spatial localization, the post-
operative pointing error had to correspond to the amount
of angle change induced by surgery. This was not the
case. Steinbach and Smith did not find pointing errors
in reoperated patients and assumed that this could be
attributed to the fact that a muscle resection operation
eliminated the distal proprioceptive organs.
I found pointing errors in recession operations,

recession plus posterior fixation suture (Cuppers
operation) and resection operations. It was anyway
impossible to correlate pointing errors with type of
procedure, amount of correction and tendency towards
relapse of the strabismus. Kommerell (1986) was unable
to replicate post-operative pointing errors in
strabismic patients. It has to be noted, however, that
he never occluded the patients binocularly after
surgery. I have always been careful in eliminating any
visual experience before testing the patients. My
results grossly coincide with those of Steinbach and
Smith.
Binocular condition and ocular proprioception.
It has been postulated (Cristini and Schiavi, 1963;

Schiavi et al., 1969) that the immediate subjective
adaptation of the patient to the change of the angle of
strabismus induced by surgery is of proprioceptive
origin. These authors based their reasoning on the fact
that a patient was able to perceive a cross with
Bagolini's striated glasses after a surgically induced
change in the angle of strabismus, without visual
experience, as he did before surgery when the angle of
deviation was different. Therefore, anomalous retinal
correspondence (ARC) of which the perception of a cross
with the striated glasses is the expression in the
presence of a deviation, is influenced by the periphery
(inflow).
340 E.C. CAMPOS
Bagolini (1976) denied this proprioceptive origin

of simultaneous changes in ARC. He explained this
phenomenon with the concept of the point-to-area
correspondence. A change in the angle of strabismus kept
anyway the image inside the area of the deviated eye
corresponding anomalously to the fovea of the fixing
eye. Thus, no change in subjective perception was
present. Bagolini's explanation, although logical is not
applicable to all cases; particularly to those in which
the change in the angle of strabismus greatly exceeds
the area of anomalous correspondence.
The body of information on the role of

proprioception on space perception is clearly limited.
Some remarks can nevertheless be made. In animal
experiments in which muscle proprioception is studied,
either passive loads or stretch are applied in the
muscles. This is quite an artificial condition and may
be transferred with difficulty to humans. I have been
able to show changes in spatial orientation in
strabismus patients in which I stretched a muscle one or
two days after an adjustable recession surgery.
In the muscles there are anyway many proprioceptive
organs. Their presence is not justified with known
functions. The human experimentation described above
shows that proprioception is involved at least in
orientation. It could so be postulated that
proprioception plays a role also in the perception of
absolute distances. In a sense the concept of "indirect
sensory function of the ocular muscles" expressed by
Tschermak-Seysenneg (1952) could be interpreted in
terms of proprioception.
Clinical situations exist in strabismus which are

not yet well understood. Only two will be mentioned
here. First, the pathogenesis of the so-called excentric
fixation is still the matter of debate. It takes place
in certain cases of amblyopia. The patient in a
monocular situation, does not use the central fovea but
an excentric retinal area for fixation. Here, the
principal visual direction normally held by the fovea,
is aquired by an extrafoveal area. A relationship
between excentric fixation and ARC has been postulated
by Cuppers and his followers. von Noorden (1985)
questioned this assumption assuming that egocenter and
relative visual directional localization were confused.
He also stated that no evidence exists of possible
changes in the egocenter.
However, I have shown (Campos and Chiesi, 1985;

Campos et al., 1986) that localization anomalies in
herpes zoster ophthalmicus patients manifest themselves
as errors in egocentric localization found in recent
onset paralitic strabismus. Therefore, new insite in the
problem of egocentric localization as well as ARC could
be looked for in proprioception.
The second problem is the result of surgical

procedures. It is well known that strabismus operations
results are often poorly predictable. Apart from
mechanical and innervational factors (as anomalous
vergence movements discussed earlier in this Symposium),
also proprioceptive information from the operated muscle
may play a role (Steinbach, 1986). The capability of
modulating these factors appears as an ideal goal.
REFERENCES
Bach-y-Rita, P. (1975). Structural-functional correla-

tions in eye muscle fibers. In: Basic mechanisms of oc-
ular motility and their clinical-rmplications. (eds.~
Lennerstrand and P. Bach-y-Rita). Pergamon Press, New
York, p. 91-111.
Bagolini, B. (1976). Part one. Sensorial anomalies in

strabismus (suppression, anomalous correspondence,
amblyopia). Doc. Ophthalmol., !!,1-22.
Batini, C. and Buisseret, P. (1974). Sensory peripheral

pathway from extrinsic eye muscles. Acta Ital. Biol.,
112, 18-32.
Campos, E.C. and Chiesi, C. (1985). Possible role of the

trigeminal nerve in egocentric localization. Lausanne,
1985. Acta Strabologica, Paris, C.E.R.E.S., p. 131-137.
Campos, E.C., Chiesi, C. and Bolzani, R. (1986). Abnor-

mal spatial localization in patients with herpes zoster
ophthalmicus: evidence for the presence of propriocep-
tive information. Arch. Ophthalmol., 104,1176-1177.
Campos, E.C., Chiesi, C. and Maccaferri, M. (1986).

Possible role of ocular proprioception in space
perception: analysis of the effect of surgery on
strabismic patients. Bordeaux 1986, Acta Strabologica,
Paris, C.E.R.E.S, In press.
342 E.C. CAMPOS
Collins, C.C. (1975). The human oculomotor control

system. In Basic motility and their clinical applica-
tions. (eds. Lennerstrand ~and Bach-y-Rita P.).
Pergamon Press, New York, p. 145-180.
Cooper, S. and Daniel, P.M. (1949). Muscle spindles in

human extrinsic eye muscles. Brain, 22, 1-24.
Cristini, G. and Schiavi, L. (1963). I disordini visivi

della strabico. Riv. di Oto-Neuro-Oftalmol., ~, 93-221.
Daniel, P. (1946). Spinal nerve endings in the extrinsic

eye muscle of man. J. Anat., ~, 189-193.
Donaldson, I.M. and Dixon, R.A. (1980). Excitation of

units in the lateral geniculate and contiguous nuclei of
the cat by stretch of extrinsic ocular muscles. Exp.
Brain Res.,~, 245-255.
Fuchs, A.F. and Kornhuber, H.H. (1969). Extraocular

muscle afferents to the cerebellum of the cat. J.
Physiol., 200, 713-722.
Gelhorn, E. (1949). Proprioception and the motor cortex.

Brain, 22, 35-62.
Kommerell, G. (1986). Personal communication.
Ludvigh, E. (1952). Possible role of proprioception in

the extraocular muscles. Arch. Ophthalmol., ~, 436-
441.
Ludvigh, E. (1952). Control of ocular movements

and visual interpretation of environment. Arch.
Ophthalmol., ~, 442-448.
Maffei, L. and Fiorentini, A. (1984). Electrophysiologi-

cal and behavioural evidence for the role of the oculo-
motor proprioception on visual functions of the cat.
Doc. Ophthalmol., ~, 97-100.
Manni, E. and Bortolami, R. (1982). Proprioception in

eye muscles. InFunctional Basis of ocular Motility
Disorders (eds. G. Lennerstrand,-o.s. Zee and E.L.
Keller). Wenner-Gren Symposium Series, 11, 55-64.
Noorden, G.K. von (1985) .Burian-von Noorden's Binocular

Vision and Ocular Motility. Theory and Management of
Strabismus. Mosby, St. Louis.
Strabismus. Mosby, St. Louis.
Richmond, F.J.R., Johnson, W.S.W., B-aker, R.S. and

Steinbach, M.J. (1984). Palisade Endings in Human Extrao-
cular Muscles. Invest. Ophthalmol. & Visual Sci.,~,
471-476.
Schiavi, L., Meduri, R. and Puddu, P. (1969). Comporta-

mento della corrispondenza retinica anomala negli stra-
bici sottoposti a correzione chirurgica: le varianti
indotte dalla mutata propriocettivita. Ann. Ottal. e Cl.
Ocul., 22• 471-490.
Skavenski, A.A. (1971). Extraretinal correction and

memory for target position. Vision Res., !!• 743-746.
Skavenski, A.A. (1972). Inflow as a source of extrareti-

nal eye position information. Vision Res., ll• 221-229.
Skavenski, A.A., Haddad, G. and Steiman, R.M. (1972).

The extraretinal signal for the visual perception of
direction. Percept. Psychophysic., !!• 287-293.
Steinbach, M.J. (1986). Muscles as Sense Organs. Arch.

Ophthalmol., 104, 1148-1149.
Steinbach, M.J. and Smith, D.R. (1981). Spatial locali-

zation after strabismus surgery: Evidence for inflow.
Science, 213, 1407-1409.
Tschermak-Seysenegg, A. von (1952). Introduction to

physiological optics. (Paul Boeder trans) Charles-c.
Thomas, Springfield, Ill., p. 219.
28
ON THE PHYSIOLOGICAL BASIS FOR STEREOPSIS*
M.L.J. CRAWFORD
The primary stimulus for contemporary ideas on the

physiological substrate for stereopsis can be dated from
a 1959 publication by Hubel and Wiesel describing the
neuronal response properties of neurons in the striate
cortex of cats. However, the recognition of
appreciation of depth through binocular viewing and
speculations on such functions go back to antiquity.
According to the scholarly account by Polyak (1957),
early thought on binocular visual function is reflected
in Galen's interpretation of the function of the optic
chiasm, which had been identified earlier as an
anatomical structure by Aristotle, and was subsequently
well known to the Alexandrian Greeks of the 2nd century
B.C. Though Galen thought of animal (human) behavior as
being determined by the ebb, flow, and mixture of
various 'spirits' or pneuma through hollow tube-like
nerves, he brought to prominence the role of the brain
as the principal organ of sensation and perception.
This was in contrast to the ideas of Aristotle and his
contemporaries, who attributed various animal behaviors
to assorted other organs, the heart being principal
among them. Among others, one function of the optic
chiasm assigned by Galen was to permit harmonious
mixture of the visual spirits returning to the brain
from each eye, thereby eliminating diplopia. How this
was accomplished was unclear. Ibn al-Haitham (Alhazen)
in the early part of the 11th century speculated that it
was only after the confluence of the two streams of
spatial information in the chiasm (a la Galen) and their
superimposition and fusion, could the 'final arbiter'
located in the brain combine them into a single image.
The location of the final arbiter has moved about
* Supported by NIH grant EY01120
345
346 M.L.J. CRAWFORD
in the head throughout history. Although Galen

recognized the connections of the optic nerves with the
forebrain, his idea of binocular vision, fusion, and the
absence of diplopia were a function of the chiasm.
Alhazen emphasized the additional stage of binocular
processing by the final arbiter of the cerebrum. Des
Cartes (1596-1650) believed that there was no mixing of
the 'spirits visual' in the chiasm and (as Ibn al
Haitham had suggested 600 years earlier) that there was
a spatial projection of the retinal surface upon the
ventricles, and thence united in the pineal gland~ the
final arbiter. Newton (1642-1727) extended the idea of
partial decussation of projections from the eyes at the
chiasm to the left or right halves of the brain, not the
pineal body. It was then the brain which united the two
cerebral impressions into a single mental impression.
John Taylor (1703-1772) elaborated upon the spatial
representation of the decussation at the chiasm, as did
Johannes Muller (1837), but it was a slow process in
getting information from the two eyes put together,
especially in the cortex.
Hermann Wilbrand (1851-1935) before the turn of

this century integrated the ideas of Galen, Ibn al
Haitham, Des Cartes, Newton, Taylor and Johannes Muller
into an anatomical concept for binocular vision.
Compatible with ideas today, Wilbrand concluded that
the two homonymous retinal halves were represented in
the same overlapping area of the cortex. His hypothesis
required that two sets of optic nerve fibers arising
from identical conjugate points of the homonymous halves
of both retinae come to lie very close together in the
optic radiations, and each terminate in adjacent minute
territories in the visual cortex. To Wilbrand, the
cortical visual center was a mosaic of such double
units, each representing a point in the ipsilateral and
a point in the contralateral homonymous halves of the
two retinae. In binocular vision, these pairs of points
worked together 'syndamically' to provide the basis for
the perception of depth. Polyak, after a lifetime of
scholarly research and experimentation, summarized the
anatomical substrate for binocular vision and stereopsis
in precisely the same terms in the 1950's, acknowledging
that his description was an elaboration of the
description by Galen some eighteen centuries earlier
(Polyak, 1957, p.789)1
During the rise of physiology and psychophysics in

the last century, there was considerable study and
advancement of ideas concerning the optics, geometry,
and motor systems of the binocular system, but the
neural substrate for such remained obscure. Binocular
oculomotor movements were functions of the 'will' as was
ON THE PHYSIOLOGICAL BASIS FOR STEREOPSIS 347
awareness and perception (Hering, 1887).
The first half of this century saw the emergence of

electrophysiology of the nervous system concentrate upon
receptors and nerve physiology, while investigation upon
the cortex was restricted to the study of gross
potentials, most notably the EEG.
It was the period of the 1950s and early 1960s with

the work of Kuffler (1953) and of Mountcastle (1957),
and the mapping of receptor surfaces upon the cortex,
combined with concurrent improvement in electrorecording
techniques and microelectrodes, that led to the most
significant observations for the physiological basis for
binocular vision and stereopsis. In 1959 Hubel and
Wiesel reported that cells in the striate cortex of cat
responded to stimulation of either eye~ had similar
receptive field stimulus requirements for stimulation of
either~ and varied in the degree of control (eye
dominance) that each eye exercised over the cell. More
than anything it was the publication of that paper which
triggered the flurry of research which continues today.
In short, the history of the electrophysiology of
binocular vision and stereopsis can be dated from this
time. It is to be remembered that the anatomical
pathways for binocular vision had, in a general way,
already been defined, and the functional necessity of
binocular convergence somewhere within the cortex was
recognized, even though there was a fundamental lack of
specific neural substrate. The binocular cell of the
striate cortex of the cat was immediately seen as the
potential neural substrate for stereopsis.
Hubel and Wiesel (1959) concluded that the

monocular receptive fields of binocular cortical cells
were the same and that they were located in "roughly
homologous parts of the two retinas". Moreover, when
the centers of the two receptive fields were stimulated
simultaneously, they observed response summation, or
antagonism, depending upon the alignment of the stimulus
upon the receptive fields, concluding that summation
occurred when corresponding parts of the receptive
fields were stimulated (as would be the case for objects
which lay in the plane of fixation) and any deviation
from superimposition would result in a diminished
response. They did not mention stereopsis and the
possible role of the binocular cell in stereopsis.
Later, in their 1962 paper (Hubel and Wiesel, 1962),
they described the additional binocular cell type which
responded when, and only when, corresponding retinal
points were stimulated simultaneously, and suggested
that even though they had no functional role to propose
for these cells, the difference in eye-dominance might
348 M.L.J. CRAWFORD
have import for judging the location of a target in

space (p.l49). Clearly, at that time the potential role
of binocular cells as the substrate for stereopsis was
not appreciated by these investigators. The reasons for
this probably had much to do with the precision of their
measurements and mapping of binocular field locations
relative to a retinal landmark, e.g. the area centralia.
Many of the binocular receptive fields which they judged
to be superimposed on corresponding points, could well
have been on horizontally disparate points on the two
retinae. At the same time, since they observed
differences in eye-dominance, they were apparently
trying to fit that phenomenon into service for binocular
vision.
As concisely presented in a recent review by Bishop

and Pettigrew (1986) the discovery of the disparity-
sensitive binocular cell of the striate cortex was made
possible by the technical advances in eye stabilization
during paralysis, allowing precise receptive field
mapping. In a paper read to the 1966 meeting of the
American Academy of Optometry, and published the
following year (Barlow, Blakemore and Pettigrew, 1967)
Blakemore described the horizontal disparities of
binocular neurons in the striate cortex of cat. Most
importantly, the authors proposed explicitly that it was
the sensitivity of these cells to horizontal disparities
which held potential for mediating the sense of depth or
stereopsis:
"To summarize briefly, three things are required of

a system to perform stereopsis. It must pick out
features of the retinal images that belong to each
other in the sense that they are images of the same
object. These cortical cells certainly do that.
Secondly, the depth detectors must respond
selectively to specific depths. The binocularly-
driven cells are indeed sensitively tuned to their
optimal depth. Thirdly, this optimal depth must
vary from one detector to another and I hope that
we have proved that this is true for these cells.
I put it to you that the very first relay station
in the visual cortex of the brain in the cat, and
probably also in man, contains the basis of a
physiological mechanism for binocular depth
perception."
An important discovery during this period of the
elucidation of the cortical physiology capable of
mediating stereopsis was made by Julesz (1960,1963,1964)
who introduced the now famous random dot stereogram
(RDS), and the use of which showed for the first time
that horizontal disparity alone was sufficient to induce
the perception of form as well as the percept of depth,

i.e. stereopsis.
A paper in 1970 by Hubel and Wiesel described

disparity sensitive binocular cells in Areas 17 and 18
of the monkey cortex, and Bough (1970) in the same issue
of Nature showed that monkeys could be trained to detect
figures embedded in static RDSs, giving evidence that
the monkey could utilize horizontal disparities, and
therefore probably had stereopsis.
Professor Peter Bishop and a long list of his

collaborators have led the field of electrophysiologists
in the development and utilization of techniques to
precisely measure the receptive field properties of
cortical binocular cells in the anesthetized and
paralyzed cat. These investigators have detailed the
receptive field properties of binocular cortical cells,
showing: (1) that most binocular cells have identical
receptive field properties in each eye (Joshua and
Bishop, 1970): (2) that the receptive fields are made up
of several spatially distinct sub-field components
responsive to dark or light bars and that these response
sub-field components may be different for direction of
stimulus motion (Maske et al., 1985): and, (3) that the
individual binocular receptive field in either eye is
most often made up of several separate binocularly-
identical sub-fields of ON and OFF response to an
increase or decrease in luminance contrast (Maske et
al., 1984, fig. 4).
Most important among these observations is (3), the

detailed ON and OFF multiple subfield structure of the
binocular fields. Figure 1 is an idealized sketch of
the binocular receptive fields in cat cortex for two
neurons. Under the normal condition of fusion, a visual
stimulus which evokes an ON response in the right eye
(1) does the same to the companion left eye field. In
the second example (B), when a stimulus evokes an OFF
response in the right eye (2) it stimulates the matching
OFF field in the left eye. During bifoveal fixation,
the binocular receptive fields are always in proper
position so that any visual stimulus falls upon
comparable parts of the subfields, therefore meeting the
requirement of simultaneous and congruent stimulation of
binocular fields: the conditions prerequsite to
stereopsis (see Maske et al., 1984). There is no reason
to doubt that the monkey has similar binocular receptive
field organization. Collectively, the precisely mapped
receptive field and receptive sub-field properties from
the cat show sensitivities to visual field disparities
to match the psychophysical discrimination in monkey and
human observers.
350 M.L.J. CRAWFORD
A B
] 80 " '
RE
LE ] 80 " ' ] 80 " '
D D
Figure 1. An illustrative example of the ON and OFF

sub-field configurations for two cortical binocular
neurons (A,B). The ON sub-field is indicated by the
histogram of response to the passage of a light bar
(L) in the direction indicated by the arrows. The
OFF sub-field is indicated in the same manner by the
heavy stroke histogram, in response to the passage
of a dark bar (D). The right eye (RE) and left eye
fields (LE) are identical in this illustration, but
may differ in the amplitude of the response, while
the number and spatial order of the sub-fields are
essentially the same. In example A, a stimulus
which evokes an ON response at (1) would evoke a
simultaneous ON response in the opposite eye field.
In B, the same would apply for an evoked OFF
response (2). This figure was fabricated after the
results of Maske et al., (1984).
Doctor Gian Poggio of the Johns Hopkins Medical

School embarked on a program of research into the
physiological substrate for binocular vis~on and
stereopsis in the early 1970's, choosing to study the
response properties of cortical visual neurons in the
alert monkey. The results obtained from this laboratory
leave no doubt that under conditions favorable for
stereopsis, neurons capable of the resolution and
fidelity necessary for stereopsis are abundantly present
and functional in the visual cortex of the alert monkey.
The basic paradigm employed by these investigators has

been to have the restrained monkey work on a basic
fixation task as they advance a microelectrode through
the visual cortex to isolate single cells. As the
animal fixates a target, the binocular visual stimuli
are separated and presented to the visual field through
the use of polarizers. The monkey does not have to
report whether or not he sees the binocular stimulus, so
there is no behavioral correlate of the neuronal
response. Never the less, this approach obviates the
use of confounding anesthesia and uncertainties
associated with eye alignment under paralysis.
Poggio and Fischer (1977) found most cells in

monkey foveal striate, as well as prestriate Area 18
cortex, to have binocular input and they described and
classified some 213 of these cells into two categories,
Balanced and Unbalanced cells. Balanced Cells (40%) were
those that gave similar responses to stimulation of
either right or left eyes, and would be comparable in
eye dominance to those found in categories 3-5 of the
scheme devised by Hubel and Wiesel (1962). Unbalanced
Cells (60%) were those which gave qualitatively
different responses to stimulation of the two eyes.
Most all (84%) of the cells that they encountered were
differentially responsive to the location of the
stimulus in depth. They described four types of cells
sensitive to stimuli having horizontal disparities
designed to induce (in the human observer) the
perception of depth: Near and Far neurons, Tuned
Excitatory, and Tuned Inhibitory neurons. Near and Far
Neurons (N/F) were cells which usually--gave an
unbalanced response to monocular stimulation i.e. showed
strong monocular eye dominance, gave an excitatory
binocular response when the stimulus was located beyond
the point of fixation (F-neuron), and showed an
inhibitory response when the binocular stimulus was
moved closer than the plane of fixation. Again, the
complementary neuron (Neuron) had just the opposite
response~ vigorous response to a binocular stimulus
closer than the plane of fixation, and inhibition of
response for presentations beyond the point of fixation
(Poggio and Fischer, 1977, fig. 6 & 7). Tuned
excitatory (Te) neurons were more common and showed a
narrow sensitivity to depth in the fronto-parallel plane
of the point of fixation. Stimuli located more of less
than about lo in front of or behind the fixation plane
failed to elicit a response (Poggio and Fischer, 1977,
fig. 4). Therefore, neurons of this type were
considered candidates for mediating stereopsis within
the range of Panum's fusional area. Since these cells
were seen to have sharp steroacuity to less than 3' of
arc (likely a gross overestimation), they were thought
352 M.L.J. CRAWFORD
f TIOled Excitatory
j
i
TIOled - o r y
<1------~~~~------i----~~~==~----
Dopth
!Re«awn alter Pogglo et aL, 1985)
Figure 2. Response profiles for the four basic

types of disparity-sensitive neurons found in visual
cortex of the alert monkey. The tuned excitatory
and inhibitory neurons have narrow response profiles
(+-0.1 deg.), most often in or near the plane of
fixation (*). The other reciprocal neurons excite
or inhibit their response for targets nearer (Near)
of farther away (Far) than the plane of fixation.
The transition over the plane of fixation from peak
excitation to peak inhibition is about , +-0.4 deg.
(Redrawn from Poggio et al., 1985, fig. 2).
to meet the sensitivity requirements necessary to

mediate the psychophysical needs of local stereopsis.
Interesting, the complementary neuron type, the Tuned
inhibitory (Ti) neuron ceased responding when both eyes
were stimulated simultaneously. Populations of these
two types of neurons were suggested to play a role in
maintaining fusion once a target had been bifoveally
fixated. Collectively, these cell types were designated
as 'depth cells' or 'disparity detectors'.
Poggio and his associates next employed dynamic

random dot stereograms (Julesz, 1960, 1963, 1964) as a
binocular stimulus to seek out the neurons responsible
for global stereopsis. They (Poggio & Talbot, 1981:
Poggio, 1984: Poggio et al., 1985) have presented a most
convincing case that indeed neurons of the monkey visual
cortex are very sensitive to horizontal image
disparities embedded in random dot stereograms. Of some

244 neurons recorded from Area 17 and 18 of four
monkeys, virtually all neurons responded to random dot
stereograms. One class of these 'cyclopean' neurons,
presumed to serve stereopsis, consisted of two
reciprocal types, Te and Ti, which had very narrow
disparity-tuning ranges around the plane of fixation.
The second class, N/F, like those responsive to solid-
figure stereograms, were neurons which responded with
excitation if the disparities were crossed (nearer than
the fixation point) and with inhibition if the
disparities were uncrossed (farther than the fixation
point). The response profiles for these neurons are
illustrated in Figure 2. Additional features of these
'cyclopean' neurons were that most had balanced eye
dominance, complex receptive fields, were found more
frequently in the supragranular layers of cortex, and
were found somewhat more frequently in Area 18 than in
Area 17.
In summary, Poggio and his associates have shown in

the monkey, and under conditions normal for stereopsis,
single neurons of Areas 17 and 18 which have the
response variety and sensitivity of horizontal disparity
to serve the full scope of stereopsis, both 'local and
global'.
The next step in tying the cortical binocular cell

function to stereopsis is to show that these cells are
necessary for stereopsis. This is done below where it
is shown that monkeys having few cortical binocular
neurons fail at stereoscopic tasks.
A few years ago we decided to study stereopsis in

the infant monkey, and more specifically, to look at
cortical binocular cells as the possible neural
substrate. Hubel and Wiesel (1974) had shown that
binocular neurons were present in the cortex of the
newborn, visually inexperienced monkey, and in a similar
manner, we had mapped cortical receptive fields in
infant monkeys (Crawford et al., 1975). We looked for a
way to manipulate the numbers of binocular neurons that
a young monkey had in his cortex, and the technique of
optical dissociation of the binocular visual world by
prisms during infancy has proven to be quite effective.
Creating a condition of chronic diplopia by having the
infant monkey view the world through prisms during the
first months of life (we've done this only on compliant
monkeys younger than 90 days of age) rapidly decreases
the population of binocular cells in the visual cortex:
the cells becoming monocular neurons responsive
exclusively to stimulation of one eye or the other
(Crawford and Von Noorden, 1980a). The time course of
354 M.L.J. CRAWFORD
binocular cell loss from the cortex is quite rapid,

being reduced to less than half (25%) the normal
percentage (>85%) with 30 days of prism viewing; within
60 days they are virtually all gone. These monkeys,
which have virtually no cortical binocular cells,
manifest no oculomotor or vergence movement defects as
discernible on clinical examination. There is no
measureable strabismus in these animals, and finally,
given extensive subsequent binocular viewing experience,
binocular cells do not recover in those monkeys
(Crawford et al., 1984, fig. 1).
We tested these monkeys and found them to have

normal contrast sensitivity and acuity in each eye, but
they failed to show evidence of normal binocular
summation under binocular viewing conditions (Crawford
et al., 1983). These results assured us that the prism-
rearing technique had produced deficits in binocular
functions without significant confounding effects.
The final behavioral test on these monkeys was that

for stereopsis. We used dynamic random dot stereograms,
and required the monkey to detect the appearance of a
form (a grating) hidden in the stereogram. Moreover, we
compared these with normal monkeys, and as well with
children who had comparable histories of eye
misalignment (i.e., strabismus which had been
corrected). The results were most compatible in that
none of the prism-reared monkeys, or the children with
histories of strabismus, could see the random dot
stereograms. On the other hand, the control monkeys and
normal children could easily see the stereograms, i.e.,
they had stereopsis (Crawford et al., 1984, fig. 2).
In summary, the case for the role of the cortical

binocular neuron as being the neural substrate for
stereopsis is extremely strong. From these results,
there is no doubt that they are necessary for
stereopsis, the detection and perception of depth from
horizontal binocular image diparities. However, the
results of studies on these monkeys shown to have few
cortical binocular cells, to be stereoblind, but to have
apparently normal binocular foveation, eye alignment and
vergence capacity, raises questions concerning the
extent of the role of cortical binocular cells in such
functions as bifoveal fixation and vergence movements.
A more detailed examination of the vergence capacities
of these monkeys is now is order to determine the
contribution of cortical disparity information to the
oculomotor system (see Mays, 1984; Cumming and Judge,
1986; and Judge and Cumming, 1986).
R E F E R E N C E S
Barlow, H. B., Blakemore, c., & Pettigrew, J. D. (1967).

The neural mechanism of binocular depth discrimination.
J. Physiology, 193, 327-342.
Bishop, P. o., & Pettigrew, J.D. (1986). Neural

mechanisms of binocular vision. Vision Res., 25, 1587-
1600.
Bough, E. w. (1970). Stereoscopic vision in the macaque

monkey: a behavioural demonstration. Nature, 225, 42-
44.
Crawford, M.L.J., Blake, R., Cool, S.J., & Von Noorden,

G. K. (1975). Physiological consequences of unilateral
and bilateral eye closure in macaque monkeys: some
further observations. Brain Res.,84, 150-154.
Crawford, M. L. J., Smith, E. L., Harwerth, R. s., & Von

Noorden, G. K. (1984). Stereoblind monkeys have few
binocular cortical neurons. Invest. Ophthal. and Vis.
Sci., 25, 779-781.
Crawford, M. L. J., & Von Noorden, G. K. (1980).

Optically induced commitant strabismus in monkeys.
Invest. Ophthal. and Vis. Sci., 19, 1105-1109.
Crawford, M. L. J., Von Noorden, G. K., Meharg, L. S.,

Rhodes, J.W., Harwerth, R. s., Smith, E. L., & Miller,
D. D. (1983). Binocular neurons and binocular function
in monkeys and children. Invest. Ophthal. and Vis. Sci.,
24, 491-495.
Cumming, B. G., & Judge, s. J. (1986). Disparity-induced

and blur-induced convergence eye movement and
accommodation in the monkey. J. Neurophysiology, 55,
896-914. --
Hering, E. (1968). The theory of binocular vision. In B.

Bridgeman & L. Stark (Eds.), Ewald Hering: The theory of
binocular vision. New York: Plenum Press
Hubel, D. H., & Weisel, T. N. (1974). Sequence

regularity and geometry of orientation columns in the
monkey striate cortex. J. Comp. Neurol., 158, 267-293.
Hubel, D. H., & Wiesel, T. N. (1959). Receptive fields

of single neurons in the eat's striate cortex. J.
Physiol., 148, 574-591.
Hubel, D. H., & Wiesel, T. N. (1962). Receptive fields,

binocular interaction and functional architecture in the
356 M.L.J. CRAWFORD
eat's visual cortex. J. Physiol., 16~, 1~6-154.
Hubel, D. H., & Wiesel, T. N. (197~a). The period of

susceptibility to the physiological effects of
unilateral eye closure in kittens. J. Physiol., 2~6,
419-436.
Hubel, D. H., & Wiesel, T. N. (197~b). Cells sensitive

to binocular depth in area 18 of the macaque monkey
cortex. Nature, 225, 41-42.
Joshua, D. E., & Bishop, P. o. (197~). Binocular single
vision and depth discrimination. Receptive field
disparities for central and peripheral vision and
binocular interaction on peripheral single units in cat
striate cortex. Exp. Brain Res., 1~, 389-416.
Judge, s. J., & Cumming, B. G. (1986). Neurons in the

monkey midbrain with activity related to vergence eye
movement and accommodation. J. Neurophysiology, 55,
915-93~.
Julesz, B. (196~). Binocular depth perception of

computer generated patterns. Bell System Tech. J., 39,
1125-1162.
Julesz, B. (1963). Stereopsis and binocular rivalry of
contours. J. Opt. Soc. Am., ~. 994-999.
Julesz, B. (1978). Global stereopsis: cooperative
phenomena in stereoscopic depth perception. In R. Held,
H. w. Leibowitz & H.-L. Teuber (Eds.), Handbook of
Sensory Physiology: Perception. Berlin: Springer.
Maske, R., Yamane, s., & Bishop, P. o. (1984). Binocular
simple cells for stereopsis: comparison of receptive
field organizations for the two eyes. Vision Research,
24, 1921-1929.
Maske, R., Yamane, s., & Bishop, P. o. (1985). Simple
and B-Cells in cat striate cortex. Complementarity of
responses to moving light and dark bars. J.
Neurophysiol., 53, 67~-685.
Mays, L. E. (1983). Neurophysiological correlates of
vergence eye movements. In c. M. Schor & K. J.
Ciuffreda (Eds.), Vergence eye movem~nts: Basic and
clinical aspects. Boston: Butterworth.
Mountcastle, v. B. (1957). Modality and topographic

properties of single neurons of eat's somatic sensory
cortex. J. Neurophysiol., 2~, 4~8-434.
Poggio, G. F. (1984). Processing of stereoscopic

information in primate visual cortex. In: G. M.
Edelman, w. E. Gall & w. M. Cowan (Eds.), Dynamic
asEects of neocortical function. New York: John Wiley &
Son.
Poggio, G. F., & Fischer, B. (1977). Binocular
interaction and depth sensitivity in striate and
prestriate cortex of behaving monkey. J. Neurophysiol.,
40, 1392-1405.
Poggio, G.F., Motter, B . . , Squatrito, s., & Trotter, Y.

(1985). Responses of neurons in visual cortex (Vl and
V2) of the alert macaque to dynamic random-dot
stereograms. Vision Res., 25, 397-406.
Poggio, G. F., & Talbot, w. H. (1981). Mechanisms of

static and dynamic stereopsis in foveal cortex of the
rhesus monkey. J. Physiol., 315, 469-492.
Polyak, s. (1957). The vertebrate visual system.

University of Chicago Press: Chicago.
Powell, T. P. s., & Mountcastle, v. B. (1959). Some

aspects of the functional organization of the cortex of
the postcentral gyrus of the monkey: A correlation of
findings obtained in a single unit analysis with
cytoarchitecture. Bull. Johns Hopkins Hasp. 105, 133.
Wiesel, T. N., & Hubel, D. H. (1974). Ordered

arrangement of orientation columns in monkeys lacking
visual experience. J. Camp Neural., 158, 307-318.
29
STEREOPSIS AND STRABISMUS
EUGENE M. HELVESTON
STEREOPSIS AND STRABISMUS
Stereopsis testing in the patient with strabismus or

suspected of having strabismus is designed to evaluate input
from the two eyes contemporaneously and depends upon
effective central processing to produce a subjective
awareness of depth. Three principle areas of clinical
strabismus management are aided by testing for stereopsis;
these are: 1) screening for strabismus, 2) prediction of
strabismus treatment outcome, and 3) assessment of
strabismus treatment results.
Screening
Historically, laboratory stereopsis testing has been

carried out by measuring "real" depth utilizing special
instrumentation. An example of such a device is the Verhoff
apparatus. This method of testing is complicated, time
consuming, requires precise judgment and cooperation on the
part of the person being tested and utilizes a cumbersome
apparatus. Because of these constraints, the Verhoff and
similar instruments and other similar devices such as the
Howard Dohlman, are not clinically applicable tools.
Current clinical stereopsis tests fall into two main
categories. One type of stereopsis test is more or less
self contained because no anaglyph or polaroid spectacles
are required. The Frisbee test is such a test made up of a
solid plastic plate with laterally displaced targets
embedded at different depths by a lamination process of
clear plastic sheets. An individual capable of recognizing
stereopsis will see the embedded objects stereoscopically or
359
360 E.M. HELVESTON
with a sense of depth. Depth perceived is related to the

lateral displacement of the embedded objects, and it is
measured as a function of seconds of arc displacement.
These in turn are a factor of the viewing distance. The
second type of stereopsis test, the Lang test, uses the
panography technique combining the lenticular screen of Hess
and the random dot patterns of Julesz. The commercially
available Lang test measures stereo acuity using three
targets; a star, a cat and a car. When held at a 40 em
distance from the subject, the objects are displaced at 600
and 1200 seconds of arc disparity. The test plate must be
held by the examiner in a fronto-parallel manner. Otherwise
the patient may receive monocular clues by moving the stereo
test in a flip-flop manner. This test has been recommended
by Lang for screening in kindergarten and also for use by
pediatricians. The stereo disparity is greater than other
recommended screening tests. However, even with other
stereo screening tests which measure disparity to as low as
15 arc seconds, the precise clinical cutoff for passing or
failing the test is not well established. That is, how much
strabismus and/or amblyopia is compatible with passing the
test at 600 or 1200 seconds of arc disparity, etc.
Arbitrarily, 168 to 240 arc seconds of disparity has been
the range as a referral cutoff, but 3,000 plus seconds of
arc disparity (Titmus fly) has been used as a screeneing
cutoff.
Currently, the most widely used clinical test uses

polaroid glasses worn by the subject while viewing symbol"s
which are displaced with selected lateral displacement.
This produces the illusion of depth for the person able to
appreciate stereopsis. The Titmus measures stereo acuity
from 3000 seconds to 40 seconds of arc disparity. The
Randot test, which uses random dots and the polaroid glasses
principle, has a gross screening plate at 600 seconds of arc
disparity and has targets which are reduced to 20 seconds of
arc disparity.
A third stereo test, the Random dot E, combines stereo

acuity with the ability to recognize the direction of a
tumbling E. This test uses a single target which at various
distances from the subject produces different disparity
values. These vary from 504 seconds of arc disparity at 1/2
meter to 50 seconds of arc disparity at 5 meters. The
Random dot E uses the Julesz Random dot technique with
polaroid glasses to produce the stereo target.
The TNO also utilizes the Julesz principle but with red
and blue-green random dots. This test requires red and
blue-green anaglyph spectacles to produce the awareness of a
red and green embedded target which is seen in depth. The
STEREOPSIS AND STRABISMUS 361
TNO test has disparity values ranging from 1,980 to 15

seconds of arc second disparity.
Traditional screening tests for VlSlon primarily deal

with the measurement of visual acuity as the benchmark of
ocular function. These vision screening tests require the
recognition of Snellen optotypes, (alphabet or illiterate
E), pictures, Landolt rings or the matching of objects. Such
testing may be carried out at the distance, at near, or
both. This type of screening will detect most significant
uncorrected refractive errors which are either bilateral or
unilateral as well as functional amblyopia. Visual acuity
testing also will detect organic defects which are either
binocular or monocular and which are related specifically to
testing done under monocular conditions. An exception to
this may be the latent nystagmus which may have satisfactory
visual acuity only under binocular conditions. Defects
which specifically effect binocular vision but not monocular
acuity (strabismus without amblyopia or stereo blindness)
are not detected by visual acuity testing. In recent years,
stereo acuity testing has been done by many vision screeners
as an adjunct to visual acuity testing. It also has been
done alone with the assumption that stereo acuity
determination also provides sufficient information regarding
visual acuity.
Walraven introduced the TNO test for screening in

Europe. This testing on patients 2 to 7 years old indicated
that failure to recognize stereopsis at 240 seconds of arc
disparity yielded what were described as excellent screening
criteria when validation experiments were carried out on
children with known ocular health. Walraven reported that
the TNO test yielded 60% less over-referrals than the
Titmus.
Simons and Reinecke, using the Random dot E (RDE) test

with a failure cutoff at 250 arc seconds of disparity
concluded that "a clinical version of the Julesz random dot
stereogram, the Random dot E (RDE), appears to be a reliable
screening test (i.e. no under referrals) for amblyopia as
well as for a variety of potential amblyopia-related visual
dysfunctions including constant large angle and
microtropias, intermittent and accommodative esotropia, and
anisometropia greater than 2 diopters." They add that
passing the RDE appears to be a fairly fail safe indicator
that none of these conditions are present.
On the other hand, when Muma, et al, compared the RDE,

Randot, Titmus and TNO, they found an unacceptable
under-referral rate which cast doubt on the value of our
present stereo acuity tests as valid screeners for subtle
362 E.M. HEL VESTON
defects in binocular vision in the young child.
Hammond and Schmitt compared the RDE with the

orthorater, sight screener, telebinocular, Massachusetts
vision kit, illiterate E, and modified clinical technique.
Their data show that the RDE is more effective than 5 out of
6 other techniques used to identify children who need
further visual examination. Additionally, the modified
clinical test, according to their studies, determined a
referral rate for children with non visual problems such as
excess lacrimation, nevi, and abnormal cup/disc ratio along
with minimal astigmatism.
Commenting on the RDE as the sole criterion for visual

screening, both von Noorden and Campos support the value of
visual acuity testing along with stereo testing. They
strongly disagree with any scheme which eliminates visual
acuity testing and advocates stereo acuity testing alone.
Ruttum, et al, using 600 seconds of arc disparity as a

stereo acuity cutoff for referral, screened 6,000 children
with both the Lang and the RDE stereo test and concluded
that there was an effect on reducing the rate of visual
acuity failure overreferrals. These tests also found 10
children with abnormal binocular vision. All of the
children were also tested for visual acuity using Snellen
optotypes.
Helveston, et al, using 140 seconds of arc disparity as

a pass level, found that 95% of 1,912 six to nine year olds
had "normal" stereo acu.ity and 94% of the same group had
visual acuity of 20/30 or better in the better eye.
It appears that stereo acuity testing provides some

information regarding visual acuity but it will not replace
visual acuity testing as a screening method. Instead it is
adjunctive.
One of the difficulties with eye muscle function

screening has been that it requires considerable expertise
to detect small or even moderate amounts of strabismus. On
the other hand, stereo acuity testing enables even an
untrained observer to obtain results which uncover
strabismus or other binocularity defects. Visual acuity
testing which has been carried out for many years is well
established in the screeners armamentarium. Stereo acuity
testing will undoubtedly be added to this routine in the
future.
Prediction
Prediction of strabismus treatment outcome utilizing

stereo acuity testing as a routine part of motility
evaluation is done in most clinics. Since this testing can
be completed by technicians it can serve as a type of office
screening. Moreover, the use of sensitive stereo acuity
testing by experienced personnel may uncover other uses for
the results of stereo acuity testing.
Counseling parents regarding the functional state of

their child's eyes for example, is a practical use for
stereo acuity testing. This is carried out in our clinic on
a routine basis. An individual with straight eyes or with
small angle esotropia is usually presented with a stereo
test as the first clinical measurement. This is done before
cover testing or before either eye is occluded for vision
testing in order to avoid dissociation and to preserve the
patient's casual seeing state. If during the initial
evaluation stereo acuity is noted to be at a high level, it
is very unlikely that anything is seriously wrong with the
patient's eyes. For example, if an individual has 40
seconds of arc stereo acuity (that is one who sees 9/9
stereo circles) there is an overwhelming chance that the
exam will be normal. In such a case, no manifest strabismu&
will be present and near visual acuity will be 20/20 to
20/30 with no more than one line difference. However, it is
not uncommon for individuals with large angle manifest
strabismus to report apparently positive findings on stereo
acuity testing at levels of up to 200 seconds of arc
disparity on the Titmus test. More direct questioning by an
experienced observer will determine that the patient
selected the circle not on the basis of seeing it in depth
but on the basis of noting that one of the circles was
displaced laterally or rapidly alternated producing
"movement" of the displaced target. This testing artifact
can explain otherwise confounding results on stereo testing
in certain clinical situations.
In our clinic, the Randot and TNO plates are used less
frequently. While the TNO produces fewer monocular clues,
it is significantly more difficult, in our experience, for a
younger observer to appreciate. The Randot is certainly a
satisfactory device and one which deserves attention. The
Lang, or Frisbee test for screening may be less useful in an
office setting where testing of higher grades of stereo
acuity is required. Projected stereo charts using the AO
projector at distance are used for detecting malingerers who
will report seeing a letter out of one eye or the other
without knowing it.
364 E.M. HELVESTON
Two circumstances occur quite commonly in the clinic.

They relate to stereo acuity levels in patients with
intermittent exotropia and constancy of stereo acuity level
in a variety of strabismic and nonstrabismic patients from
one visit to the next. Individuals with large angle
intermittent exotropia in many instances report completely
normal stereo acuity. That is, a person with 40 prism
diopters of intermittent exotropia and an eye that is
deviating up to SO% of the time will be very likely to
demonstrate stereo acuity to 40 seconds of arc disparity.
This is a true finding because monocular clues are not
ordinarily detectable at this level. These results are
somewhat counter to the experiences reported by some
screeners who note that latent deviations are associated in
their experience with decreased stereo acuity levels. It is
our clinical impression that the latent deviation is not the
cause of the decreased stereo acuity however decreased
stereo acuity may coexist with other binocular conditions
which may contribute to some intermittent deviations.
A clinical observation of ours is that over many years,

and repeated visits, patients will report the same stereo
acuity. For example, an individual may have stereo acuity to
140 seconds of arc disparity on the Titmus and maintain this
level on repeated examinations spanning a period of years.
The stereo acuity neither improves nor decreases over this
time provided that no significant change occurs in visual
acuity or alignment. This gives the impression that the
stereo response is more or less "hardwired" into the
patient's central nervous system.
Stereo acuity testing has been carried out in our

laboratory in infants and very young children. This testing
uses the dynamic random dot stereogram or "Fox Box" with
optokinetically displayed stereo targets coupled with EOG
recording. Using this technique we have been able to
confirm the findings of Held and Sokol. In the normal
straight-eyed infant, gross stereo acuity can be measured at
approximately 4 months of age. In a series of infants with
congenital-infantile esotropia, none were found to have a
positive stereo acuity response during the time they
demonstrated a manifest deviation. However, after the eyes
were surgically straightened, several children responded
with a stereo acuity response. This response was found in
about half of the small series tested. During the time when
these patients demonstrated a stereo acuity response their
eyes were noted to be straight on stringent cover testing.
In all of these patients the stereo acuity response was lost
over a period of weeks along with the development of a small
angle of strabismus. Similar findings were noted by Leguire
and Rogers. However in their series, a higher percentage of

children demonstrated stereo acuity and no follow-up was
recorded. They recorded positive stereo acuity on the basis
of the eyes moving in the same direction as the target two
thirds of the time. To obtain a more accurate response to
stereo testing we placed EOG electrodes on the infants.
This technique allowed both quantitative and qualitative
evaluation of the response. That is; to score a positive
stereo response, movement of the eyes must be in the same
direction as the movement of the stereo target and must also
move at the same speed and have an amplitude similar to the
movement of the target.
Results from these experiments suggest that stereo

acuity response occurs as a type of maturation process in a
child who has the capability of developing stereo acuity.
One of these requirements is to have straight eyes. In
infants who have neither straight eyes nor stereo acuity,
stereo acuity can be more or less reawakened if the eyes are
rendered straight by peripheral manipulation such as
surgery. However, some weakness in this feedback system
seems to cause the stereo acuity to be lost with the
reappearance of strabismus. It is unknown which comes first
the strabismus or loss of stereopsis (motor fusion
capability).
Some children who are otherwise normal are said to be

stereo blind. The presence of straight eyes, equal normal
vision, and normal fusional amplitudes in these patients
indicate the differences between motor fusion and
stereopsis.
Archer, et al, have recently demonstrated fusion

amplitudes in individuals who maintained target fixation on
the basis of stereoscopically seen random dot targets. A
series of experiments refute the disclaimers who state that
edge effects and other nonstereoscopic factors are
responsible for maintaining the fusion lock during the
measurement of fusional amplitudes. These tests were also
carried out with dynamic random dot stereogram generated
targets in the arms of a modified haploscope. These tests
were aimed at reducing the potential for induced optokinetic
response. Such studies show that global stereopsis can
provide a target with fixation stability to serve as anchor
for the generation of large angle fusional amplitudes.
Assessment of Results of Treatment
Stereo acuity testing has been used extensively as a

measure of results of strabismus surgery. von Noorden
reviewed the tests used for the assessment of strabismus
function extensively in the 1981 Scobee lecture. He pointed
366 E.M. HELVESTON
out that a complete understanding of the tests used was

necessary before conclusions could be drawn. Stereo acuity
testing has been used by Ing to demonstrate fusion results
after surgery for congenital esotropia. He has reported
that a high percentage of patients after early surgical
treatment for congential esotropia were able to demonstrate
stereo acuity by appreciating the stereo fly at 3,000
seconds of arc disparity. This type of stereo acuity may be
found in small angle residual esotropia and says not so much
about the sensory results of the procedure but about stereo
potential of the residual angle. Parks used a cutoff of 67
seconds of arc disparity or 6/9 Titmus circles as criteria
for diagnosis of monofixation syndrome. This is a small
angle esotropia of 8 prism diopters or less with a central
scotoma, free alternation or amblyopia and remaining
fusional amplitudes.
In a series of congenital esotropia patients,

Helveston, et al, demonstrated stereo acuity would be more
likely to occur in presumed congenital- infantile esotropia
patients operated on after one year and with small residual
deviations as compared to those operated on before one year
and with similar residual deviations. It was speculated that
these patients operated on after one year were seen later
and operated on later on the basis of a later appearing
esodeviation. This raised the possibility that these
patients may have had a longer interlude of straight eyes
before being seen initially.
As a clinical tool, stereo acuity seems to be most

valuable for screening purposes. Its use in the clinic for
evaluation and prediction undoubtedly has some value, but
does not significantly alter treatment programs in most
cases.
Further studies using measurement of stereo acuity are

ongoing in our laboratory. We believe that these tests will
show that testing for stereo acuity may find its greatest
use as a tool to help us understand how and perhaps why the
eyes either stay aligned or develop strabismus.
References
1) Archer SM, Helveston EM, Miller KK, Ellis FD

(1986): Stereopsis in normal infants and infants with
congenital esotropia. Amer. J. Ophthalmo. lQl,
591-596.
2) Avilla CW, von Noorden GK (1981): Limitation of the

TNO random dot stereo test for visual screening. Amer.
Orthoptic J. }!, 87-90.
3) Donzis PB, Rappazzo JA, Burde RM, Gordon M (1983):

Effect of binocular variations of snellen's visual
acuity on titmus stereoacuity. Arch. Ophthalmo. lQl,
930-932.
4) Ehrlich MI, Reinecke RD, Simons K (1983): Preschool

vision screening for amblyopia and strabismus: Programs,
methods, guidelines. Survey Ophthalmol. 28, 145-163.
5) Fox F, Lehmkuhle S, Leguire LE (1978): Stereoscopic

contours induce optokinetic nystagmus. Vision Res.
1189-1192.
6) Frisby JP, Mein J, Saye A, (1975): Use of random

dot stereograms in the clinical assessment of strabismic
patients. Br. J. Ophthalmo. 59, 545-552.
7) Hammand RS, Schmidt PD (1986): A random dot E

stereogram for vision screening of children. Arch.
Ophthalmo. 104, 54-60.
8) Julesz B (1971): Foundations of cyclopean

perception. Univ. of Chicago Press, Chicago, USA.
9) Kohler L, Stigmar G (1973): Vision screening of

four-year-old children. Acia Paediatr Scand. 62,
17-27.
10) Marsh WR, Rawlings SC, Mumma JV (1980): Evaluation

of clinical stereo acuity tests. Ophthalmology, 87.
1265-1272.
11) Okuda FC, Apt L, Wanter BS (1971): Evaluation of

the TNO random dot stereogram. Amer. Orthopt. Jour.
34, 124-131.
12) Reinecke RD, Simons K (1974): A new stereoscopic

test for amblyopia. Amer. J. Ophthalmo. 78, 714-721.
368 E.M. HELVESTON
13) Ruttum MS, Bence SM, Alcorn D (1986): Stereopsis

testing in a preschool vision screening program. Jour.
of Pediatric Ophthal and Strabismus. 23, 298-302.
14) Simons K (1981): Stereo acuity norms in young

children. Arch. Ophthalmol. 99, 439-445.
15) Simons K, Reinecke RD (1974) A reconsideration of

amblyopia screening and stereopsis. Amer. J. Ophthalmo,
~. 707-713.
16) Walraven J (1975): Amblyopia screening with random

dot stereograms. Amer. J. Ophthalmol. 80, 893-900.
30
PSYCHOPHYSICAL CONSEQUENCES OF IMAGE
DEGRADATION AND BINOCULAR MISREGISTRATION
ON THE DEVELOPING VISUAL NERVOUS SYSTEM
DENNIS M. LEVI and STANLEY A. KLEIN
INTRODUCTION
Amblyopia represents a developmental abnormality of the "spatial

sense" resulting from image degradation and/or binocular image
misregistration early in life. Clinicians generally consider spatial vision in
terms of Snellen acuity, with a limiting acuity of about 20/20 (i.e. critical
detail of about 1'); however, the visual system is capable of making much
finer spatial discriminations. For example, relative position, size and
orientation can be judged with an accuracy of 3 to 6 arc seconds or better
(Klein and Levi, 1985). These low spatial thresholds are 5 to 10 times finer
than either the cutoff spatial frequency or the intercone spacing. For this
reason, Westheimer (1975) has coined the term "hyperacuity" to describe a
variety of tasks that involve sensing the direction of spatial offset of a line
or point relative to a reference.
This performance is remarkable considering that even the smallest

foveal cones are separated by about 30", and the point spread function of
the eye spreads the image over several cones. Moreover, Westheimer and
McKee (1975) have shown that hyperacuity is robust to retinal image
motion. They demonstrated that a vernier target could be swept across
about 100 cones in a 200 msec epoch without degrading performance.
Hyperacuity or positional acuity has been the focus of much recent
research and modelling, both in normal and anomalous vision. There are
several reasons for this. Firstly, it is of basic interest to understand how
the visual system achieves this high degree of accuracy - i.e. how it solves
the technical problem of "sub-pixel" resolution. Secondly, several lines of
evidence suggest that performance on hyperacuity tasks reflects cortical
processing (Klein and Levi, 1985; Levi and McKee, 1986). Thirdly, it seems
likely that the mechanisms underlying hyperacuity have the more general
task of form and shape analysis (Marr, 1982; Watt and Morgan, 1984). In
short, hyperacuity represents a sensitive indicant of the spatial sense, and
may provide a window into the operation of the cortical mechanisms of
369
370 D.M. LEVI and S.A. KLEIN
1000
300
0w
"'a:
0
&
~ & '
...
> @
5
0 & &
< '•
z @
&
0
;::: 30 ,, &
iii
0 ' '/ ' /
a..
@'
&
~~
~~~
"ot:! • p
.s 10 20 so
SNELLEN ACUITY (MINUTES OF ARC)
Fig. 1. The relationship between Vernier acuity (in arc sec) and Snellen
acuity (in arc min). The data are from several studies in our
lab. Data of normal control observers are shown by the open
circles. Preferred eyes of strabismic amblyopes are shown by
the P's in boxes, those of anisometropic amblyopes shown by the
P's. Data of the amblyopic eyes are shown by A's, S's and B's for
anisometropic and strabismic amblyopes and those with both.
Also included are data from Rentschler and Hilz (1985). The
open triangle is the mean data of their normal observers, the
filled triangle, the mean of the preferred eye of their
amblyopes. Letter in triangles show the amblyopic eyes of all
but 1 of their observers.
form perception. It is not surprising then that in amblyopia, which is
considered to be an abnormality of the "form" sense (Wald and Burian,
1944), that hyperacuity is markedly degraded. What is both interesting, and
surprising, is that in hyperacuity tasks, several lines of evidence now
suggest that the neural losses of strabismic and anisometropic amblyopes
are fundamentally different. The purpose of this chapter is to review
recent studies of the spatial sense and to point to mechanisms which may
underly the psychophysical and clinical anomalies of the amblyopic visual
system.
Hyperacuity, Snellen Acuity And Grating Acuity
There is a close connection between the Snellen acuity and the

Vernier acuity of amblyopes. Figure 1 plots the Vernier acuity of a group
of amblyopes against their Snellen acuity. These data are collated from
several studies (Levi and Klein, 1982; 1985; Rentschler and Hilz, 1985). A
PSYCHOPHYSICS OF AMBLYOPIA 371
8 400
{j]
() Ill
••
A 300
@ w
200 Ul 200
(,)
> a:
I-
5 s /
0 100 _,
Q 100 /
I
< /
/ 0 /
a:
w~
/ ::t
60
ii'l /
/
- 50
z 0w ~/
Ul
w /
a: Ul
w @ a: /
> 0 0/ ::t 40 /
/
1-
o"' /o z
~·t•
//®
wS 20 0
"
z
< ~
j::
(,)
20 ~/
~/
...-'z w
,/
/ /
10 / Ul /
iii
:I
p _, 10
tl
5 4(
:I ®
j::
0.. 5
.5 5 10 0 .5 2 3 5 10
GRATING ACUITY GRATING ACUITY (MINUTES)
(MINUTES)
Fig. 2A Unflanked vernier acuity vs grating acuity for normal control

observers (open circles), preferred eyes of anisometropic
amblyopes (P's in circles), preferred eyes of strabismic
amblyopes (P's in squares) and amblyopic eyes of strabismics
(S's), anisometropes (A's) and both (B's). Data are also shown for
vernier acuity using repetitive vernier gratings (white letters).
The dashed line has a slope of 1. On these log-log coordinates
the line indicates vernier and grating acuity being linearly
related. The solid line is for the normal periphery at 2.S 0 , S0 and
10° in the inferior visual field. (From Levi and Klein, 198S).
Fig. 2B The optimal bisection threshold is plotted against each observer's
grating acuity for three normal control observers (N's), the
preferred eyes of anisometropic amblyopes (P's in circles) and of
strabismic amblyopes (P's in squares) and of amblyopic eyes of
amblyopes with strabismus (S's), anisometropia (A's), and both
(B's). Also shown are data of two normal observers at 2.S 0 , S0
and 10° in the lower visual field (solid circles and squares). The
solid line has the form B = 0.6SG - 0.3 (where B is bisection
acuity and G is grating acuity) and provides a reasonable fit to
the data of the periphery and of the strabismic amblyopes (S's
and B's). The dashed line has a slope of 1 - indicating a
proportional relationship between bisection and grating acuity.
This line provides a good fit to the data of the anisometropic
amblyopes (A's). (From Levi et al., 1987).
similar relationship exists between bisection acuity and Snellen acuity (Levi
and Klein, 1983; 1987). Specifically, the slope of the line drawn through
the data of Figure 1 is unity, suggesting that Snellen acuity and position
acuity are affected in the same way by the amblyopic process.
Figure 2 plots vernier acuity (Figure 2A) and acuity for another
position discrimination task, viz. bisection (Figure 2B), as a function of
grating acuity. In each of these figures, the data of nonstrabismic
anisometropic amblyopes are shown by A's while those of strabismic
amblyopes, and those with both constant strabismus and anisometropia are
shown by S's and B's respectively. For both tasks the A's fall along the
straight line with unity slope, showing that grating resolution and positional
acuity are similarly affected by the amblyopic process. However, the
strabismic amblyopes show grater losses in the accuracy of positional
acuity than for grating resolution.
This result, is similar to the decoupling of Snellen acuity and grating

acuity in strabismic amblyopes (Levi and Klein, 1982a and b) and points to
the close link between positional acuity and Snellen acuity. (This link will
be examined in greater detail below). Levi and Klein (1982) showed that
the reduced positional acuity of amblyopes could not be accounted for on
the basis of eccentric fixation or faulty eye movements and thus represents
a neural deficit(s) in the amblyopic visual system. The nature of these
deficits will be discussed below.
Spatial Interference with Hyperacuity
Interference effects are ubiquitous in spatial vision. Such effects

which are known to clinicians as "crowding" occur for orientation
discrimination (Westheimer et al., 1976), stereoacuity (Butler and
Westheimer, 1978), Vernier acuity (Westheimer and Hauske, 1975, Levi et
al., 1985) and Snellen acuity. Flom et al., (1963) measured the spatial
extent of interference with acuity by placing flanking bars at various
distances from a near threshold Landolt C. They found that the linear
extent of interference was far greater in the amblyopic than in the normal
eyes, but· that when scaled to the unflanked acuity, the extent of
interference was similar in normal and amblyopic eyes. Similar
observations were reported by Hess and Jacobs (1979).
Vernier acuity is also markedly degraded when the target is flanked

by a pair of optimally positional flanks (Westheimer and Hauske, 1975; Levi
et al., 1985). This spatial interference with Vernier acuity is strongest
when the flanks are 2-4' from the target in normal foveal vision. It occurs
even if the target is presented to one eye, and the flanks to the other,
suggesting that this effect is cortically mediated. Levi and Klein (1985)
measured the extent of spatial interference with Vernier acuity in a group
of amblyopic observers, and compared this to the results of the normal
fovea and periphery. Their data showed that for the normal fovea, the
periphery and for both eyes of amblyopes, the strongest spatial
interference occurs when the contours are separated by about 30 times the
unflanked threshold. Thus, if the unflanked threshold is 5", the strongest
interference would be at a distance of 2.5'. In strabismic amblyopes,
because Vernier acuity is often much more degraded than grating acuity,
easily resolved targets may be subject to strong spatial interference, since
the zone of interference is tied to the Vernier acuity rather than the
grating acuity. Thus "crowding" occurs at spatial frequencies well below
the cutoff in strabismic amblyopes. As will be discussed below, we think
that the interference effects which occur for Vernier acuity, are closely
linked to the crowding effect for Snellen acuity, which is well known to
clinicians.
Cortical Modules for Spatial Processing

It has been suggested (Westheimer, 1981; Westheimer and McKee,
1977; Levi et al., 1985) that position coding requires a processing zone of
several minutes of arc in the normal fovea and correspondingly larger (in
proportion to the threshold) in the periphery. When interfering contours
are present within this zone, positional coding is degraded. Similar effects
occur for orientation discrimination and letter acuity, suggesting that each
may share a common basis. Hubel and Wiesel (1974) first suggested that
the striate cortex consists of a large number of repeating modules, each of
which carries out a highly stereotyped analysis of the inputs from a small
region of the visual field. Each module, or hypercolumn, consists of a pair
of ocular dominance columns (representing all possible orientations).
Several new anatomical techniques have clarified and extended the notion
that the visual cortex is organized in a modular fashion, and it now seems
likely that there may be several types of modules (e.g. cytochrome oxidase
blobs; hypercolumns) which may carry different information. These
modules have also been identified in the human cortex and are spaced at
approximately 1mm intervals (Hitchcock and Hickey, 1980; Horton and
Hedley-White, 1984).
Barlow (1979) suggested the intriguing notion that the processing zone
required for optimal hyperacuity has its anatomical basis In the modular
organization of the visual cortex. The finding that the dimensions of this
zone grow in peripheral vision, in proportion to the cortical magnification
factor (Levi et al., 1985; Yap et al., 1987) lend credence to this notion, and
suggest that spatial interference may reveal information about these
modules in human vision. In the normal fovea, the psychophysical spatial
processing module is 4-5 minutes of arc. This is approximately the spatial
extent of a normal human ocular dominance column, which is about 1mm
(Hitchcock and Hickey, 1980). It is Interesting to note that this is also
about the overall dimension of a threshold level Snellen letter (20 /20 letter
= 5').
ECCENTRICITY (OEG)
0 5
25
i
::::E
- 20
IIJ
0
z
..:
.. 15
1/)
0
~ 10
0
;:::::
a:
0 •
0
Fig. 3. Schematic illustration of cortical magnification and its

relationship to strabismic amblyopia. Each heavily outlined box
represents a normal 1mm cortical processing module and the
scattering of the centers of the receptive fields within it. At
each eccentricitYt_ the spatial extent of the module is scaled in
proportion to M- • Each stippled box represents a hypothetical
processing unit for a strabismic amblyope. The zone of
scattering of receptive field centers has been increased by a
const&nt. This increases scatter a lot at 0°, but proportionally
less with increasing eccentricity. (From Levi and Klein, 1985).
Figure 3 shows schematically the arrangement of these processing
modules as a function of eccentricity. The heavily outlined boxes represent
the normal visual system. Here the vertical axis is in mm of cortex, each 1
mm module being about the size of a normal human ocular dominance
column or a spatial processing module. The horizontal axis is
eccentricity. The horizontal extent of each box represents the spatial
extent of each lmm module which was constructed to be approximately 4'
at 0° eccentricity, and to increase in proportion to the inverse of the
cortical magnification factor. The stippled boxes represent the enlarged
processing modules which are found in amblyopes. There are several
possible reasons for the enlarged modules of the amblyopic eye: (1) In
order to achieve a high signal to noise ratio, a large fixed number of
neurons may be required (Sakitt and Barlow, 1982). The amblyopic loss of
neurons in both strabismus and anisometropia would therefore necessitate a
larger processing module. (2) The lack of fusion between the two eyes
occuring in strabismus may cause an extra scattering of receptive fields in
the deviating eye (Petti5rew, 1974). Extending the zone of scattering by a
constant amount (eg •• 5 ) would have the effect of increasing the relative
size of the foveal processing module a lot, but would have a proportionally
smaller effect with increasing eccentricity. This notion of processing
modules may help to explain the inability of strabismic amblyopes to
perform other hyperacuity tasks (e.g. bisecting at small separations (Levi
et al., 1987).
How can the losses of positional information in amblyopic eyes be

understood? A review of models for positional acuity is beyond the scope
of this chapter; however, there are now several models which attempt to
account for the high precision of foveal positional acuity on the basis of the
spatial filtering and spatial sampling properties of the visual system (e.g.
Watt and Morgan, 1985; Klein and Levi, 1985; Wilson, 1986). In the
following section we consider several factors which could account for the
losses which occur in the amblyopic visual system, and their implications.
Factors Which May Contribute to the Amblyopic Loss of Positional

Information
A number of factors have been suggested which could potentially

limit spatial processing, these include:
1. The contrast response of the mechanisms (receptive fields).
2. The spatial sampling density.
3. The collection of spatial samples.
4. Disorder of the topographical mapping of visual space.
1. Contrast Response Function
A key feature of several recent models for position discrimination is

the contrast sensitivity and suprathreshold contrast response function of
the putative spatial filters or receptive fields (Watt and Morgan, 1985;
Klein and Levi, 1985; Wilson, 1986a). In amblyopes, there is ample
evidence for reduced contrast sensitivity, thus it is plausible that reduced
contrast sensitivity or signal-to-noise ratio of the filters may result in
reduced positional sensitivity. The clear implication of this notion is that
there should be a proportional reduction in position sensitivity - i.e. that a
single scale factor should account for the loss of contrast sensitivity,
resolution and position acuity. In anisometropic amblyopes this appears to
be the case (Bradley and Freeman, 1985; Levi and Klein, 1982; 1983; 1985);
however, in strabismic amblyopes there appears to be an additional loss of
positional acuity. The weight of the findings suggest that at least to a first
order approximation the position losses of anisometropic amblyopes for a
variety of conditions (vernier vs spatial frequency; bisection vs separation)
can be understood in terms of a reduced contrast sensitivity (signal-to-
noise ratio) of the filters of the amblyopic eye. This reduced signal-to-
noise hypothesis for anisometropic amblyopia is controversial! Watt and
Hess (1987), like us, find that anisometropic amblyopes demonstrate
marked positional uncertainty. However, we differ on the interpretation.
They argue against the signal-to-noise hypothesis, because making their
target lines brighter in the amblyopic eye, failed to equate the
performance of the two eyes. However, it is likely that a thin line (like
that used in their study) is detected by mechanisms near the peak of the
contrast sensitivity function, while the position cue is detected by smaller
mechanisms which are most susceptible to image degradation due to
anisometropia. Thus simply adjusting the contrast of the line according to
its detection threshold would not necessarily equate the position
judgements. By and large, the extant data on anisometropic amblyopia
favors the signal-to-noise hypothesis. This hypothesis receives
physiological support from the studies of Eggers and Blakemore (1978) who
showed that neurons with small receptive fields in the cortex of kittens
reared with experimental anisometropia have reduced contrast sensitivity.
The recent work of Hess and Pointer (1985), suggests that it is not simply
blur per se, but rather abnormal binocular interaction due to monocular
image degradation, which results in the loss of sensitivity in the
anisometropic amblyope.
2. Spatial sampling density
An important consideration for positional acuity is the sampling

density of the mechanisms (i.e. the degree of overlap of receptive fields of
a particular class). Little is known about how visual information is actually
sampled by the two-dimensional array of cortical receptive fields. There is
clearly a trade-off between the need for adequate representation of various
sizes, orientations etc, and the requirement for efficiency (since there are
a finite number of neurons).
A disproportionate loss or aberrant position labelling of the cortical

neurons would result in critical local cues for position discrimination being
missed. Thus a sparse cortical sampling grain would be expected to result
in a greater loss of position acuity than resolution, grating acuity or
contrast sensitivity as occurs in strabismic amblyopia. If the sampling
losses were nonuniform local spatial distortions may occur (Hess et al.,
1978; Bedell and Floro, 1983).
An alternative scheme is one suggested recently by Blasdel (personal

communication). In the normal visual cortex, the anatomical processing
modules contain information from each eye (i.e. a left plus right ocular
dominance column) and all orientations. If, as seems likely, binocular
competition in strabismus leads to shrinkage of the inputs from the
deviated eye, than in any given module particular orientations may not be
represented. This, notion makes several specific predictions. Firstly,
grating acuity or contrast sensitivity should be less affected than Snellen
acuity which requires local information at all the orientations represented
in the letter. Secondly, Vernier acuity should be markedly degraded
because it relies upon very localized relative orientation and position
information. Thirdly, positional acuity may be especially poor with short
lines, and should benefit from the addition of samples because the noise
between samples along the length of lines would be uncorrelated.
3. The Collection of Spatial Samples
In foveal vision, the image of a stimulus is spread by the optical blur
function of the eye over several cones, and is highly magnified in the
cortex. Thus, in foveal vision even a single dot will be sampled by many
overlapping cortical receptive fields so sparse sampling is never present,
and adding samples will have little effect upon the accuracy of positional
information. On the other hand, in a sparsely sampled visual system,
positional uncertainty should be high with a single stimulus sample, and
adding samples would be expected to reduce positional uncertainty in
proportion to the square root of the number of samples. Levi and Klein
(1986) measured bisection thresholds for stimuli comprised of discrete
samples (dots). For the normal fovea and for anisometropic amblyopes
adding samples (up to about 5) resulted in only a small improvement in
thresholds. However, for strabismic amblyopes and in normal peripheral

vision, thresholds were very high with 1 sample showing marked intrinsic
positional uncertainty in the absence of spatial and temporal averaging.
Adding samples, up to about 10 resulted in a marked improvement in
performance In proportion to the square root of the number of samples, as
would be expected if sparse sampling were the critical limiting factor.
4. Disorder of the Topographical Mapping of Visual Space
In the normal mammalian visual system visual space is
topographically mapped in a highly ordered manner at a number of levels
along the visual pathway. Disorder of the spatial metric would be expected
to exert a profound influence upon visual processing of positional
information. This notion has been suggested by Hess (1982). Scrambling of
the spatial metric would likely have a marked effect upon positional acuity
and Snellen acuity (where local spatial relationships are critical) but should
have less influence upon resolution or contrast sensitivity. It is quite likely
that disorder of the spatial metric may be a consequence of abnormal
spatial sampling. Recently, Wilson (1986a) has attempted to model the
hyperacuity of amblyopes. He found that the data of strabismic amblyopes
could be quantitatively modelled by assuming both spatial undersampling,
and positional uncertainty, while that of anisometropic amblyopes was
modelled by a reduced gain of the small cortical receptive fields.
Summary
Amblyopia is characterized by marked spatial uncertainty. The

inability of amblyopes to judge relative position is most marked under
conditions where normal observers perform best, i.e. when the stimulus
features are close together. In anisometropic amblyopia the loss of
positional information is commensurate with the reduced resolution and
contrast sensitivity of the amblyopic eye. In contrast, strabismic
amblyopes show an extra loss in positional acuity, often accompanied by
aberrations of space perception. The imprecise spatial metric of the
amblyopic visual system appears to be closely linked with the amblyopes'
Snellen acuity.
It is likely that the high precision of spatial judgements in normal
foveal vision is the outcome of mechanisms which have the more general
task of form/pattern discrimination - thus the studies described above
provide a psychophysical window into the mechanisms of form perception in
amblyopia.
In anisometropic amblyopes it is hypothesized that the reduced

positional acuity and resolution share a common basis - i.e. the reduced
contrast sensitivity of the spatial filters of the amblyopic eye as a
consequence of binocular competition resulting from monocular image
degradation. In strabismic amblyopia a more tenable hypothesis is that
sparse spatial sampling due to loss of neurons or scrambling of their signals
due to binocular image misregistration may account for the high degree of
intrinsic positional uncertainty.
The results emerging from recent psychophysical investigations of

amblyopes have important implications for both neurobiologists and
clinicians. First, they imply that different forms of abnormal early visual
experience result in different neural consequences. Moreover, they raise
the question of whether amblyopias arising from different causes should be
treated as separate clinical entities.
REFERENCES
Barlow, H.B. (1979). Reconstructing the visual image in space and time.
Nature, 279, 189-190.
Bedell, H.E. and Flom, !VI.C. (1983). Normal and abpormal space
perception. Am. J. Optom. Physiol. Opt. 60, 426-435.
Bradley, A. and Freeman, R.D. (1985). ls reduced vernier acuity in

amblyopia due to position, contrast or fixation deficits? Vision Res. 25, 55-
66.
Butler, T. and Westheimer, G. (1978). Interference with stereoscopic

acuity: spatial, temporal and disparity tuning. Vision Res. 18, 1387-1392.
Eggers, H.M. and Blakemore, C. (1978). Physiological basis of

anisometropic amblyopia. Science 201, 262-267.
Flom, M.C., Weymouth, F.W. and Kahnemann, D. (1963). Visual resolution

and Contour interaction. J. Opt. Soc. Am. 53, 1026-1032.
Hess, R.F. (1982). Development sensory impairment: amblyopia or

tarachopia? Human Neurobiol. 1, 17-29.
Hess, R.F. and Jacobs, R.J. (1979). A preliminary report of acuity and
contour interactions across the amblyopes visual field. Vision Res. 19,
1403-1408.
Hess, R.F., Campbell, F.W. and Greenhalgh, T. (1978). On the nature of the
neural abnormality in human amblyopia: Neural aberrations and neural
sensitivity loss. Pflugers Arch. Ges. Physiol. 377, 201-207.
Hitchcock, B. and Hickey, T. (1980). Ocular dominance columns: evidence

for their presence in humans. Brain Res. 182, 176-179.
Horton, J.C. and Hedley-Whyte, E.T. (1984). Mapping of cytochrome

oxidase patches and ocular dominance columns in human visual cortex.
Phil. Trans. R. Soc. Long. B 304, 255-272.
Hubel, D.H. and Wiesel, T.N. (1974). Uniformity of monkey striate

cortex: A parallel relationship between field size, scatter, and
magnification factor. J. Comp. Neurol. 158, 295-306.
Klein, S.A. and Levi, D.M. (1985). Hyperacuity thresholds of 1 sec:

theoretical predictions and empirical validation. J. Opt. Soc. Am. A 2,
1170-1190.
Levi, D.M. and Klein, S.A. (1982a). Hyperacuity and amblyopia. Nature,
298, 268-270.
Levi, D.M. and Klein, S.A. (1982b). Differences in Vernier acuity between
strabismic and anisometropic amblyopes. Invest. Ophthal. Visual Sci., 23,
398-407.
Levi, D.M. and Klein, S.A. (1983). Spatial localization in normal and
amblyopic vision. Vision Res. 23, 1005-1017.
Levi, D.M. and Klein, S.A. (1985). Vernier acuity, crowding and
amblyopia. Vision Res. 25, 979-991.
Levi, D.M. and Klein, S.A. (1986). Sampling in Spatial Vision. Nature, 320,
360-362.
Levi, D.M., Klein, S.A., and Aitsebaomo, P.A. (1985). Vision Res. 25, 963-
977.
Levi, D.M., Yap, Y.L. and Klein, S.A. (1987). Positional uncertainty in
peripheral and amblyopic vision. Vision Res. 27, 581-598.
Levi, D.M. and McKee, S.P. (1986) Dichoptic Hyperacuity: the precision of
Nonius alignment. J. Opt. Soc. Am. A. 13, 12.
Pettigrew, J.D. (1974). The effect of visual experience on the development

of stimulus specificity by kitten cortical neurons. J. Physiol. 237, 49-7 4.
Rentschler, I. and Hilz, R. (1985). Amblyopic processing of positional

information. Part 1: Vernier acuity. Expl. Brain Res. 60, 270-278.
Sakkit, B. and Barlow, H.B. (1982). A model for the economical encoding of
the visual image in the cerebral cortex. Biol. Cibernet. 43, 97-108.
Wald, G. and Burian, H.M. (1944). The dissociation of form vision and light
perception in strabismic amblyopia. Am. Ophthalmol. 27, 950-963.
Watt, R.J. and Hess, R.F. (1987). Spatial information and uncertainty in
anisometropic amblyopia. Vision Res. 27, 661-67 4.
Watt, R.J. and Morgan, M.J. (1985). A theory of the primitive spatial code
in human vision. Vision Res. 25, 1661-1674.
Westheimer, G. (1975). Visual acuity and hyperacuity. Invest. Ophthal. 14,

570-572.
Westheimer, G. (1981). Visual hyperacuity. Prog. Sensory Physiol. 1, 1-30.

Westheimer, G. and Hauske, G. (1975). Temporal and spatial interference

with vernier acuity. Visipn Res. 15, 1137-1141.
Westheimer, G. and McKee S.P. (1975). Visual acuity in the presence of
retinal image motion. J. Opt. Soc. Am. 847-850.
Westheimer, G., Shimamura, K. and McKee, S.P. (1976). Interference with

line-orientation sensitivity. J. Opt. Soc. Am. 66, 332-338.
Westheimer, G. and McKee, S.P. (1977). Integration regions for visual
hyperacuity. Vision Res.17, 89-93.
Wilson, H.R. (1986a). Responses of spatial mechanisms can explain
hyperacuity. Vision Res. 26, 453-470.
Wilson, H.R. (1986b). Model of peripheral and amblyopic hyperacuity.
Invest. Ophthal. Visual Sci. Supp1 27, 95.
Yap, Y.L., Levi, D.M. and Klein, S.A. (1987). Peripheral Hyperacuity: 3-
dot bisection scales to a single factor from 0 to 10 degrees. J. Opt. Soc.
Am. A. In Press.
31
AMBLYOPIA: CLINICAL ASPECTS
W. HAASE
1. Definition and classification of amblyopia

We differentiate 2 etiologic lines of amblyopia: 1st
stimulus deprivation, 2nd suppression (von Noorden) .The
majority of our patients suffer from a combination of
both: moderate stimulus deprivation (ametropia or aniso-
metropia) with strabism. It is easy to agree with the
general definition and classification of amblyopia. On
the other hand it seems to me very difficult to draw a
line between a normal eye and an amblyopic eye. In our
country we treat patients wroreach an acuity less than
0.8 Snellen as amblyops. Crowding-phenomenon reduces the
line acuity to a level lower than Snellen- or single-
symbol-acuity. Again what value of this kind of acuity
represents the border between normal and amblyopic
eyes?
2. Examination of the patient
During the first 3 years of age diagnosis of amblyopia
is based on a few qualitative examinations: in squin-
ting children the ability of fixation of each eye in
free space is evaluated. About from the 4th month of
life on we can examine fixation ophthalmoscopically on
the fundus.
The question is: how often do we underlie an error in
the evaluation of fixation in the free space? In a
group of 4 year old children it was possible to compare
the fixation in free space with an acuity test. We
found a correct assessment in 52 out of 67 children
(78%). The reliability of the ophthalmoscopic fixation
test reaches the same level:
Examination of fixation
Ophthalmoscopy n Haidinger brushes 6 to 10 y. old
fov. 32 fov. children
fov. 10 eccentric
eccentric 37 eccentric n = 83
eccentric 4 fov. Table 1
381
382 W. HAASE
This comparison gives an idea of the frequency of error

also in younger children.
A basic procedure of the examination even of babies is
the objective measurement of the refractive error. This
is carried out by a retinoscopic (skiascopic) method.It
is possible to measure the refraction under cycloplegia
with an accuracy of ±0.5 D, the value is below any
amblyogenic factor.
The measurement itself is not the problem clinicians
have rather more the paralysis of accommodation. We have
to admit not having tests which can help us assess the
funtional disturbance in cases who do not squint,fixate
foveolarly and show anisometropia or marked high ame-
tropia.
Examination of the visual acuity:
The attempts to design tests adequate for children bet-
ween 2 and 5 years old seem to proceed (Lithander's
Kolt-test). Her basic idea was to play with the children
"music box" (Lithander).
The use of gratings to test acuity in children includes
some risks (Gstalter and Green; L. Mayer and Coworkers).
We also compared grating-tests with symbol-tests and
found systematical differences between both.
Amblyops discriminate much better large field gratings
than small field gratings or optotypes (Fig. 1).
11 normals, 10 formerly ambly-

opic patients and 18 amblyops
have been examined. Method:
Rectangular black/white
II I
3' gratings of 3°, 1° field size
and a very small field size
0. ., G)).·, ~
1 ,
~ ,.
P
equivalent to the outer dia-
meter of a Landolt ring. The
L------~~--~=---~~--~~-----
LR 0LR ,. 3"
spatial f5equences 0 varied bet-
ween 3 c/ and 60/ . Landolt
rings were also used from
Fig. 1
AMBL YOPIA:CLINICAL ASPECTS 383
0.1 to 2.0 acuity. To avoid the identification at the

cut off of the bars, we surrounded the test-field with
random dots. This pattern inhibits perception of the
edges of the gratings. Results: Grating acuity is supe-
rior to Landolt ring asuity bn amblyops but only with
large test fields of 1 or 3 diameter. In some amblyops
the difference becomes extreme: An anisometropic girl
6 y. without squint showed 11 steps, a 19 y. old lady
with strabismic amblyopia increa~ed acuity from 0.05
Landolt rings to 0.8 gratings, 3 field size (Fig.2).
--
Vilual KUity min/arc
-----
2.0
1.6
1.25
Fig. 2
These results were confirmed by a sine wave grating test

~Retinometer/Rassow and Wolf). Also with sine waves
amblyops reached much better acuity with large test
fields, whereas patients who suffered from macular dege-
neration did not improve their visual acuity with in-
creasing size of the test fields. The phenomenon seems
to be specific to amblyops. Because of this "summation-"
effect a periodic pattern of a large field size seems
not to be an adequate test to screen for amblyopia or
to examine visual acuity of amblyopic patients.
Problems in acuity-testing appear also when we use the
traditional targets. It is very important how the test-
sympols, which have to be discriminated are arranged to
each other (v.Hofe, Weckert 1930; Irvine 1948). After
the basic quantitative work of Flom et al. no practi-
cable attempt has been undertaken to design a test for
a quantitative measurement of the crowding-phenomenon.
Together with A. Hohmann we based an acuity test on
Landolt rings with a defined separation between the
rings. After experimental examinations with other
384 W. HAASE
separations (4-5~, 8-9' of arc, Fig, 3). We decided to

use 3 sets for clinical purposes (so called C-Test):
Fig. 3
lst Landolt rings more than 30 min of arc separated

from each other single types. 2nd acuity cards with a
separation of 17.2 min of arc in order to receive infor-
mation of the extension of the crowding. 3rd a test with
2.6 min. separation because the maximum of crowding
occurs within an area of 2-3 min. of arc around a con-
tour. The actual acuity with these line types compared
with the acuity of isolated Landolt rings represents a
quantification of the crowding. We found crowding in
90% of our amblyopic patients. (Haase and Hohmann 1982;
1986) .
Patients who suffer from organically based lesions
hardly showed crowding (Haase et al. 1985).
3. How does the patient see with the amblyopic eye?
Subjective reports of patients have been published by
several authors (v.Hofe, Sireteanu and Fronius, Irvine,
Pugh, Hess): The patient needs more time to identify
a certain detail. Sometimes he is able to see an object,
sometimes it fades away. The accuracy of seeing changes
with time. For example visual acuity can change between
20/400 and 20/70. Sometimes parts of a letter disappear
or they become double. Neighboured contours overlap
each other (Crowding) • The localisation of a thing can
inverse: the opening of a "C" seems to be right but is
AMBLYOPIA:CLINICAL ASPECTS 385
left or up but is down. Hess reported about extreme

distorsion of periodic gratings.
We asked patients to copy naive pictures from an origi-
nal. They should paint first with their good eye seeing
and after that by using the amblyopic eye.
The results can be summarized as follows:
a. Loss of detail-perception/resolution
b. Fading of a fixated object or of parts of it.
c. Change of the size of objects - they appear enlarged.
d. A change of the relative localisation.
e. Overlapping of different objects.
f. Diplopia - in certain parts of the visual field
can appear.
g. Color often seems to be pale, patients need a higher
color-intensity.
h. Anisometropic amblyopia without squint leads to less
distorsion of localisation than squint-amblyopia.
i. All amblyops copied the pictures less accurately than
patients with organically based low vision of a
comparable acuity.
4. Different forms of human amblyopia, !.Deprivation:
The animal experiments which were carried out to find a
model for amblyopia in humans itself became a model to treat
early stimulus deprivation amblyopia in humans adequate-
ly.
Several authors have shown that severe congenital obstac-
les of the optical media of the eye habe to be removed
within the first 3-4 months (bilateral) or the first
2-3 weeks (unilateral) in order prevent an irreparable
stimulus deprivation amblyopia (Beller et al., Frey et
al., Gelbart et al., Pratt-Johnson and Tillson, Vaegan
and Taylor). Clinicians main problem is not surgery but
the adequate and accepted optical supply and compliance
of the conservative therapy. Varying from animals the
sensitive period in humans seems to never end for the
whole life. This is well estabilished for the binocular
system and might be true also for the monocular func-
tion, although a decline of sensitivity up to the lOth
year of life (Vaegan and Taylor) is well known. We also
observed acquired stimulus deprivation amblyopia up to
the age of ten years. A special aspect of uniocular
stimulus deprivation in humans is the loss of sensitivi-
ty in the nasal visual field. We examined 4 patients,
2 suffered from congenital cataract - aphakia; 1 was a
boy with a total ptosis acquired at the age of 2.5 years,
examined at the age of 11 years, the perception of his
left eye (ptosis) was reduced to hand-movements (Fig.4).
386 W. HAASE
Fig. 4 Fig. 5
The fourth patient acquired cataract due to eye injury

at the age of 9 years, removal of the cataract followed
at the age of 19 yea rs, now h e is 31 y. old, acuity is
reduced to hand-movements os (Fig. 5 ) . His right eye is
normal.
4.2 Ametrop ic I Anisometropic amblyop ia
In humans as well as in animals bilateral stimulus de-
privation l ead s to less severe amblyop ia than uniocular
deprivation. This is n o t only true for early depriva-
t ion amblyopia but also for the moderate fo rm o f ame -
trop ic amblyopi a . A sample of 50 cases with parallel
eyes and bilateral ametropia (5 3 hyperme trops) 5 D.
showed that the "cure"-rate measured by an acuity level
of 0.8 single or 0.5 line and change of eccentric to
foveal fixation depended on the age of applicatio n of
correct glasses. The same was observed in patients with
squint but with a highe r rate of failure of the treat-
ment . True eccentr i c fix a tio n was a ve ry ra r e e ven t in
bil ate ral ame tropic a mblyops (Ta bl e 2 ) with p a rallel
eye s . Ametropia 5 D
Table 2: Parallel eyes Squint
examination examination
first last first last
Fov . fixation 35 49 36 48
Fov. scatter/both e yes 11 1
unioc ul a r scatte r 1 1 5 2)(.
ecce ntric f ix ation
both e yes 2 8
eccentric fixation
uniocular 2 6 5x
n = 50 n = 54
( 1 pa tient:s catter in ( X only six
1 eye , ecce ntric in the patien ts)
other )
AMBL YOPIA:CLINICAL ASPECTS 387
The fixation pattern typically in non-squinting high

ametropic patients is a scatter within the foveal area
without any preference of direction. Strabismic patients
usually show preference of direction when tehy fixate
eccentrically.
Anisometropic amblyopia - is said to be of a very un-
favourable prognosis, especially hypermetropic aniso-
metropia. The prognosis depends very much on the age of
detection and correction. Correct glasses appliP.d within
the first 2-3 years of life lead to a high rate of
"cure". In our patients we found only a slight depen-
dance of improvement on the degree of anisometropia
(Table 3).
Anisowetropia, parallel eyes

(mean difference 2. 8 D)
n = 18 hypermetropic anisom., n = 11 myop1.c anisom.
(mean d. = 2.0 D) (rrean d. 4.1 D)
0.6 s LR 11 (61%) 6 (73\)
0.5 11ne 10 C55%) 6 (73\)
not l.mproved 3 (16.6\) 1 ( 9\)
Anisometropia + Squint
(mean difference = 2.4 D)
Hypermetropic anisorr. 'Myopic anisom.
(mean d. 1.9 D) (mean d.= 3.75 D)
49 n = 17
n = 26
0.6 s
L
1
LR
-
1.5 D anisom.
21 ( 75%)
n 2 L 1
2
- 1.5 D anisom.
0.5 line 16 (64%) 2

not improved 5 (16%) ¢
n 21 I
0.6 s
>
LR
1.5 D anisom.
15 (71%)
n = 15 L
7
> 1.5 D anisom.
(47%)
0.5 line 12 (57%) 6 (40%)
not improved 6 ( 26\) 3 (20\)
Table 3: s = single Landoltring (LR); line = 2.6 LR

(C-Test). \ o f patients who reached a visual acuity
of 0.8 s, 0.5 line or better
4.3 Strabismus-Amblyopia
According to the reports in the literature the rate of
improved cases ceases the more the older the patients
are at onset oftreatment (Sattler 1927, Oppel 1964,
Dayson 1968). We assume that the amblyopia inducing
factor (s) continue to detoriate the amblyopic eye even
above the lOth year of life, because the difference of
the 2 acuities between the dominant and the amblyopic
eye increases with the age of the patients (Fig.6).
~88 W. HAASE
IWplofeoultylllnel
.._._,"'""'
Acuity clomlnent .,.
mlnue .aulty ~;•.,.
Fig. 6
-1 2-3 4·1 Adultl .. 27y.old
A challenging problem is the existence of bilateral

(strabismic)amblyopia in patients with moderate ametro-
pia. We observed 17 cases which had bilateraleccentric
fixation and ~5 D. ametropia. 3 of them had parallel
eyes. 8 patients changed to foveolar fixation by treat-
ment in both eyes. 5 could be cured in 1 eye only, 2
patients with parallel eyes improved spontaneously to
a norma 1 leve 1 .
In a 10 year old girl we observed the occurence of
bilateral eccentric fixation. The girl was our patient
since her 1st year of life because of infantile eso-
tropia. She had reached single Landolt ring acuity of
0.8 in both eyes and microstrabism after surgery. At
the age of 10 y. her acuity decreased in both eyes and
eccentric fixation developed (Fig. 7). Treatment im-
proved acuity and changed fixation to the foveola in
both eyes. The explaination of that development remains
speculative but I suspect a possible mechanism of
suppression which occures in a mosaic like pattern of
the visual field in each eye. This would not be a
strange new phenomenon - we know of this rivalry in nor-
mals. In this context, it is suspected that the good
eye in strabismic patients is also involved in the
whole pathologic process even in its monocular func-
tions (Hermann and Priestley, Haase 1984). We saw that
the acuity of the dominant eye in a group of uniocular
amblyops did not reach the age-dependend mean level of
the normals. New investigations (Flom et al.) of the
spatial distortion in amblyops showed significantly the
involvement of the nondeviated eye in strabismic
amblyops. We repeated this procedure. The spatial
distortion improved with treatment (Fig. 8).
AMBL YOPIA:CLINICAL ASPECT S 389
So we cannot agree with the conclusion of the authors

that the spatial distortion is a primary etiologic
defect of the strabismic visual system .
. 11. \' • 1 .. 1'4
tccrr trn IIIUoU"'' 1 Lovt.h L')'•·~
.n tl .. 111. t.r
10 )'L'IIIIr• an • ~<•'''''
1.0=- -
"'ho "'•• urdrr centro! b<IC'ilv•r
tJt llqlllrt
1.211
..........
_IT.,
-·IT
-
Oo
~~ , todooc.IILI" ~--1
01 eoc.fb. r
I
I
I
Ool
Ool
e t ~ n 12 a_.,,..,.
Fig. 7
Fjg. 8 a
390 W . HAASE
SPATIAL. DISJOR!IO!r:f
I beiOte t~ment l
"'
Legend for Fig. 8 a, b, c
The triangles (Fig. 8 a)
were constantly presented
monocularly to the obser-
ver. The light bar has
been flashed for 130 msec
randomi zed 10 times at
each displacement 1 - 5
(= 6' until 30' of arc).
The number of wrong answers
were counted for each dis-
placement.
b = before, c = after
treatment.
"'
Fig. 8 b
SPATIAL DISTORTION
l•tt•tr~)
Str.oilmk~.n•'t!
"'
"'
Fig. 8 c
AMBL YOPIA:CLINICA L ASPECTS 391
References
Beller, R.; Hoyt, C.S.; Marg,E.and Odom, J.V. (1981).
Good Visual Function After Neonatal Surgery for Congeni-
tal Monocular Cataracts. Am.J.Ophthal.,~, 559-565.
Dayson, A. (1968). 5 Year Survey of the Use of Occlusion
in the TreatmeDt of Eccentric Fixation. Brit.orthop .J.,
.?2_,66-74.
Florn ,M.C.; Kahnemann, C.and Weymouth,F.W . (1963) .Visual
Resolution and Contour Interaction. J.Optical Soc.Amer.,
2}, 1026-1037.
Flom,M.C. and Bedell, H.E. (1985) .Identifying Amblyopia

Using Associated Conditions, Acuity and Nonacuity-F ea-
tures. Am.J.Optome try & Phys.Opt.,& l, 153-160.
Frey,Th.; Friendly,D.a nd Wyatt,D. (1973). Re-Evaluatio n

of Monocular Cataracts in Children. Am.J.Ophtha l.,76,
381-388. --
Gelbart.S.S .; Hoyt,C.S.; Jastreleski, G.and Marg,E. (1982)
Long-Term Results in Bilateral Congenital Cataracts.
Am.J.Ophth al.,2l, 615-621.
Gstalter, R.J. and Green,D.G. (1971). Laser-Inter fero-

metric Acuity in Amblyopia. J.Ped.Ophthalmol.,~,251-256
Haase,W.und Hohmann,A. (1982). Ein neuer Test (C-Test)zur

quantitative n Prufung der Trennschwie rigkeiten (Crow-
ding) - Ergebnisse bei Amblyopie und Ametropie. Klin.
Mbl.Augenh eilk.,l80,210 -215.
Haase, W. (1984). Amblyopie-D iagnostik in Theorie und

Praxis der modernen Schielbehand lung, Ed.Meyer-
Schwickerath und Ullerich. F.Enke, Stuttgart.
Haase, W.; Mashiah,M.; Hohmann,A. and Schulz,E. (1985):

Quantitative Measurement of the Crowding Phenomenon in
Amblyopia - A New Test in Int.Sympos.S trabismus and
Amblyopia, Edit. P. Nemet and J.B.Weiss,C .E.R.E.S.Pa ris,
103-112
Haase,W. und Hohmann,A. (1986). Development of Separa-
tion-Ability of Contours During Childhood - Quantifica-
tion of the Crowding Phenomenon in Amblyopia in Detec-
tion and Measurement of Visual Impairment in Pre-Verbal
Children, Ed.B.Jay,Doc umenta Ophthal.Pro c.Series ~,
138-145.
Hermann,J.S . and Priestley,B .S. (1965). Bifoveal Instabi-

lity.Am.J.O phthal.Ser.3 , 60,452-459
Hess,R.F. (1982) :Developmen tal Sensory Impairment:A mblyo-
pia or Tarachopia? Human Neurobiol.l , 17-29.
392 W. HAASE
v.Hofe and Weckert,S.M. (1930). Untersuchungen iiber das

Sehen in Fallen von Schielamblyopie.Ber.Dtsch.Ophthalmol.
Ges.Heidelberg, if, 41-45.
Irvine,S.R. (1948). Amblyopia ex Anopsia.Observations on
Retinal Inhibition, Scotoma, Projection,Light Difference
Discrimination and Visual Acuity. Trans.Amer.Ophthal.
Soc.XL VL, 527-575.
Lithander, J. (1984). "Kolt-Test". Priifung der Sehscharfe
bei 2jahrigen. Z.prakt.Augenheilk.,~, 258.
Mayer, D.L.; Fulton,A.B. and Rodier,D. (1984). Grating
and Recognition Acuities of Pediatric Patients.
Ophthalmology,2l, 947-953.
Mayer, D.L. (1986). Acuity of Amblyopic Children for

Small Field Gratings and Recognition Stimuli.Investig.
Ophthal.Visual Sci., £2, 1148-1153.
v.Noorden,G.K. (1985). In Burian - von Noorden "Binocular

Vision and Ocular Motility" 3rd Edit., p.210-211, The
C.V. Mosby, St.Louis-Toronto-Princeton.
Pratt-Johnson,J.A. and Tillson,G. (1981). Visual Results

After Removal of Congenital Cataracts Before the Age of
1 Year. Canad.J.Ophthalmol.~, 19-21.
Pugh, M. (1962). Amblyopia and the Retina.Brit.J.Ophthal .

.!§_, 193-211.
Rassow, B. und Wolf, D. (1973). Erfahrungen mit dem

Laserinterferenzstreifentest bei der Messung des retina-
len Auflosungsvermogens. v.Graefes Arch.klin.Exp.Oph-
thalmol.~ 61-66.
Sireteanu, R. und Fronius, M. (1986). Verzerrte :Raum-

wahrnehmung bei Amblyopen. Z. prakt.Augenheilk.J,
243-246.
Vaegan and Taylor, D. (1979). Critical Period for Depri-
vation Amblyopia in Children. Trans.ophthal.Soc.U.K.,2i,
432-439.
32
DISCUSSION: PSYCHOPHYSICS RELATED TO
Moderator: EMILIO C. CAMPOS
Campos: We dealt this morning with three different

subjects, namely proprioception, binocularity in
strabismus and finally amblyopia. I would like to orient
the discussion on these three different subjects. Let me
start with proprioception. Are there comments on this
subject?
von Noorden: What is the practical importance of inflow

in the use of the eyes? Blind people often have no idea
where their eyes are and cannot put them in the correct
position when you tell them to look to the right, left,
or straight ahead. Inflow, if it exists, must produce a
very weak signal indeed.
Campos: Dr. Steinbach, do you like to comment on this?

Steinbach: Inflow is a very weak signal and that is why
it takes so long to have any effect. I have reviewed the
literature elsewhere (Steinbach, 1987) 1 and concluded
that proprioception is involved in maintaining the
stability and conjugacy of gaze, giving information
about visual direction, as well as playing a role in the
development of various visual and sensorimotor
behaviors. It must have a very long time constant but
there are numerous examples of oculomotor plasticity
that requires days to take effect. Many examples of such
slow change are seen when there is sudden onset rectus
paresis, or when the vestibule-ocula r reflex is tampered
l)Steinbach, M.J. (1987) Proprioceptive knowledge of eye

position. Vision Research (in press).
393
394 E.C. CAMPOS
with (summarized in Leigh & Zee, 1983)2.
Campos: Let me also try to answer Dr von Noorden • s

question. From a clinical point of view data on inflow
now available make the whole picture even more
complicated. It would certainly be ideal for us to
simplify our models as basic scientists always try to
do. Unfortunately this is not the case. On the other
side there are facts. I don't think we have elements now
to explain the role of inflow. Only speculations are
possible. If indeed inflow is a long term calibrator,
then it needs the short term information which is given
by the outflow. In eyes which have been blind for a long
time the short term information is lacking so that the
long term calibrator is uneffective.
Lennerstrand: Could the spindle endings supply the

position signal? They would be slackened by the
operative procedures that Dr Steinbach described and
probably more so after a reoperation and a myotomy than
from the simple recession of a muscle.
Steinbach: I cannot rule out contributions from muscle

spindles. Our data force us to the musculotendinous
region and to the tension receptors located there, but
there is an obvious problem with trying to get position
information from tension: it can be the same for two
very different positions of the eye in the orbit
(Collins, 1975) 3 • I wonder if there are other
mechanoreceptors in the retrobulbar space to monitor the
position of the optic nerve, or tactile receptors on the
cornea and conjunctiva which provide eye position
information.
Singer: Until recently one way of treating trigeminal
neuralgia has been coagulation of the ganglion of
Gasser. I wonder whether anybody has cared to
investigate those patients.
Camoos: We have been trying for years to get patients
who had this type of procedures not too long before. It
turns out that unfortunately for us, but fortunately for
them, the great majority of patients with trigeminal
neuralgia can be effectively treated with drugs
(Tegretol). Thus, this procedure is nowadays very rare.
2) Leigh, R.J. and Zee, D.S. (1983). The Neurology of
Eye Movements,, F.A. Davis, Philadelphia.
3) Collins, C.C. (1975) The human oculomotor control
system. In Basic mechanisms of Ocular Motility and their
Clinical Implications (Edited by Lennerstrand G. and
Bach-y-Rita, P.) ,pp. 144-180. Pergamon Press.
DISCUSSION: PSYCHOPHYSICS RELATED TO STRABISMUS & AMBLYOPIA 395
Moreover, the neuralgia of the first branch, i.e. the

ophthalmic branch is very rare compared to that one
involving the other branches. I have only one patient in
which I could confirm my herpes zoster ophthalmicus
results.
Steinbach: I found one patient in Toronto with the
ophthalmic division of the trigeminal sectioned.
Disrupting binocular vision (patching either eye for an
eight-hour period) caused this patient to lose some
information about the position of the deafferented eye
(Steinbach, 1986) 4 , a result that parallels those
reported by Dr Maffei here and Gutrie et al. (1982) 0 •
Kommerell: As Dr Campos directly referred to the data

obtained by Otmar Bock and myself (Bock and Kommerell,
19866; Kommerell and Bock, 1987 7 ) I asked for permission
to show them in a summary slide. As you can see, the
pointing shifts correlate quite well with the change in
eye position induced by surgery, i.e., the prediction of
the outflow theory. There was only one patient who did
not show any pointing shift. In contradiction to Dr
Steinbach's findings, the amount of pointing shifts did
not depend on whether or not a previous operation has
been carried out. The patient who did not show a shift
had in fact been operated before. We think that the
pointing experiments before and after strabismus
reported sofar are compatible with the outflow theory,
and do not require to invoke inflow information on
postoperative eye position. There are other ways of
explaining the variability of the results. First, one
has to consider that egocentric localization as measured
in the pointing task may be ambiguous in strabismus
patients. Indeed, in our patients the results of the two
tests performed before surgery differed by up to 10.8
deg. This is not surprising since it is well-known that
strabismus patients can show anomalous retinal
4) Steinbach, M.J. (1986). Inflow as a long-term
calibrator of eye position in humans. Acta Psychologica,
63, 297-306.
5) Guthrie, B.L., Porter, J.D. and Sparks, D.L. (1982).
Role of extraocular muscle proprioception in eye
movements studied by chronic deafferentation of
intraorbital structures. Soc. Neurosci. Abstr. ~. 156.
6) Bock, 0. and Kommerell, G. (1986). Visual
localization after strabismus surgery is compatible with
the "outflow" theory. Vis. Res. '26, 1825-1829.
7) Kommerell, G. and Bock, 0. (1987). Pastpointing after
strabismus surgery: lack of proprioceptive information
about postoperative eye position. In: Proc.
International Orthoptic Congress, Harrogate. In press.
396 E.C. CAMPOS
correspondence with different angles of anomaly, even

when the same test procedure is repeated.
It is likely that such multiple values of relative
visual localization are related to multiple modes of
egocentric visual localization. Second, one has to
consider the instrumentation as a possible source of
non-visual clues. For instance, it may be important that
Dr Steinbach's patients were lying on their backs. This
provides a non-visual cue from "straight ahead", i.e.
the position of the centre target. The use of only three
target positions, instead of the nine in our study, may
have further helped the patients to avoid pastpointing.
As far as the apparatus used by Dr Campos is concerned,
one has to ask whether information about eye position is
at all necessary, be it inflow or outflow. The whole
visual field was available to the patients including the
semicircular cardboard. When a rod was shown to the
patient he could localize simply by relating it to the
visual context. It is probably important in such
investigations that the patient doesn't see anything but
the target. Another uncertainty in the interpretation of
Dr Steinbach's findings is that he measured the angle of
squint with the non-operated eye, instead of the
operated eye fixating (personal communication). As
muscle surgery unavoidably impairs the motility of the
eye to some extent, the effort required to move the
operated eye to certain orbital positions is increased,
and the prediction of the outflow theory is calculated
too small in the case of surgical undercorrection, and
too large in the case of overcorrection. So, there are
several possibilities of explaining the data other than
by invoking inflow information on postoperative eye
position.
Steinbach: I am compelled to say that we must consider

inflow. You are concerned because by testing in the
supine position we might give the patient a gravity cue.
That I agree could happen but that argues for the
absence of any shift in localization. In our studies we
are always comparing two groups. One group doesn't show
much of a, shift (the newly operated patients) and the
other group does show a shift (the reoperated patients
or the myotomized patients). You cannot turn gravity on
or off to explain our results. Your point about the
effort required to move the operated eye is well taken.
But many of the arguments that I make from the data that
we have collected hinge on the pointing responses made
using the non-operated eye. These measurements are
always taken before we take the measurements from the
operated eye, so the operated eye is still under a
bandage and has never had the need to fixate on any
target and therefore the innervation to that operated
eye has never had to be increased. So the data that we

have from non-operated eyes is crucial and still
supports inflow. The third point I wish to make is that
we hav.e significant difference in the way the patients
were treated during surgery. Our patients were under
general anaesthesia and your patients were awake and
under local anaesthesia with retrobulbar injections of
topical anaesthetic. What we must resolve is whether or
not your patients were proprioceptively deafferented
because of the local anaesthetic back in the orbit. In
addition to knocking out pain fibers it may also have
knocked out proprioceptive fibers during a crucial
period after the surgery. In addition, Porter 8 just
reported that retrobulbar anaesthetics have a myotoxic
effect.
Kommerell: Significant pointing shifts using the non

operated eye were not found in our patients. Our post-
operative data were obtained one day after surgery, and
we assume that the effects of local anaesthesia on nerve
conduction had worn off in about 2 hours.
Bagolini: I would like to make a general consideration.

We have to distinguish between egocentric and
oculocentric localization. Absolute stereopsis is in the
realm of egocentric localization: relative stereopsis
deals, on the contrary, with oculocentric localization.
In fact, it is through disparate retinal elements that
we localize the depth of an object wiht respect to
another. The work of Drs Maffei, Campos and Steinbach
deals only or mainly with egocentric localization. Now,
the main problem is: has egocentric localization some
interference with oculocentric localization? According
to some comments by Drs Campos and Maffei, it seems that
egocentric localization has some interference with
oculocentric localization. This is difficult to accept
for many of us but there is a point brought out by Dr
Maffei. He says that by cutting the ophthalmic branch of
the fifth nerve he found a decrease in binocular cells
in the visual cortex. This finding could create a link
between egocentric and oculocentric localization.
Campos: Dr Maffei, do you want to comment on this fact

and also on the possible relation between the loss of
binocular cells, the alteration of vergence movements
described in monkeys and the behavioural changes found
in animals.
8) Porter, J.D., Edney, D.P. and McMahon, E.J. (1987)
Extraocular myotoxicity of the retrobulbar anaesthetid
Bupivacaine. Invest. Ophthalmol. Vis. Sci, 28, 153, ARV~
Suppl.
398 E.C. CAMPOS
Maffei: The loss of binocular units in the visual cortex

after section of the ophthalmic branch of the 5th nerve
is not necessarily correlated with the impairment of
stereopsis that we observe in our behavioural
experiments. At this moment to distinguish between an
alteration of vergence movements and loss of binocular
units as first cause of the impairment of steropsis is
practically impossible. In addition we should consider a
3rd possible cause, namely an alteration of the
disparity detecting mechanisms, an alteration which is
not necessarily correlated with the loss of binocularity
of cortical cells.
I notice, with some surprise, that there is quite a bit
of resistance to accept the evidence from clinical and
animal experiments in favour of a role of inflow
proprioceptive signals. I like to comment that in
science to be resistant to new findings and theories is
quite a good attitude, to be blind however can be quite
dangerous in the way to future experiments.
Campos: If there are no further comments on the

intriguing problem of proprioception I would just
briefly add that the literature on physiology of vision
particularly as far as space perception is concerned, is
based on old results which are still valid. It is also
based on theories which have not been necessarily
proven. People, however, usually have a lot of problems
in rejecting theories they were brought up with. Let me
take an example. In classical textbooks it has been
written that you cannot have strabismus without
suppression and that anomalous correspondence (ARC) is
not sufficient as an antidiplopic mechanism. There are
real facts, however, which show that ARC is in some
cases sufficient as an antidiplopic mechanism.
Nevertheless, there are still people who believe that at
least a very little suppression scotoma has to exist
because this would be consistent with what they learnt.
von Noorden: Using the Bagolini glasses one finds a

fixation point scotoma in many patients with anomalous
correspondence.
Campo~ I do not deny obviously that patients exist who
do show a co-existence of suppression and anomalous
correspondence. On the other hand it is impossible to
say that there are no patients in which anomalous
correspondence is the only antidiplopic mechanism.
Helveston: To support the concept of strabismus and
anomalous retinal correspondence occurring without
suppression assumes that we know how to determine
suppression. Perhaps it is inappropriate for us to say
suppression does not exist simply because it did not

exist in a certain patient or on a specific test that we
determined arbitrarily was the ultimate test for
suppression scotomas. I suggest that if a patient has
strabismus and no diplopia then anomalous retinal
correspondence is invariably present and can be
confirmed if tested for with a nondissociating test.
Moreover, suppression will also be found as it is the
mechanism of relief for diplopia and the stimulus for
anomalous retinal correspondence.
Camoos: Whenever one talks of anomalous correspondence

or of suppression one has to specify the way in which
this condition was actually tested. If I have a test
with which I am able to show anomalous correspondence, I
have no reason which compels me to think that a
suppression scotoma, at least a very little one, has to
exist, because of what is written in the textbooks or
classical papers. For example Harms in 1938 examined the
binocular cooperation in strabismics and stated that in
all patients at least what he calls a fixation point
scotoma was present. I used an experimental paradigm
which matches the casual seeing condition of the
patient. I examined patients with Harms' technique and
found the same results with his techniques. Using my
technique the results were different. So, I am not
denying the existence of this classical building which I
think nobody could. The concept of the necessity of a
suppression scotoma can be transferred to the physiology
of binocular vision. The binocular sensory condition of
strabismic patients is very likely an exageration of a
physiological situation. In turn, we would therefore
have to think that even in normals suppression cannot be
excluded. And in fact it has been claimed that the
binocular visual act is achieved through alternating
suppression. Therefore, you have to accept the presence
of suppression also for normals. Let us move now to the
general problem of binocularity.
Aslin: I would like to ask Dr Campos a question about
the issue of suppression. In your first presentation
this morning, you mentioned that it seems incongruous
that one could interpret VEP data as indicating complete
summation, i.e. an addition of inputs from the two eyes
as if they were acting independently. You presented very
convincing data that in normal adults who are undergoing
rivalry there is a reduction in VEP summation consistent
with the presence of rivalry or suppression. The reason
I think this notion that the inputs from the two eyes
could sum together appears unlikely, is that we cannot
imagine an organism that has no rivalry or suppression
process. I think your comment was directed to data
400 E.C. CAMPOS
collected by Shea, McCullough and myself in normal

infants prior to four months of age. We reported the
presence of what looks like total VEP summation of the
inputs from the two eyes. Those data suggest that
rivalry or suppression mechanisms prior to about 4
months of age may not be working in the same way as in
normal adults. Data from Dr Held's lab, which he did not
have time to present in his talk, provide psychophysical
support for the absence of rivalry in norma.l infants
prior to 4 months of age.
Campos: I am perfectly aware of Dr Held's data which are

most interesting for me because they show that the
rivalry mechanism goes through a maturation process. I
did not enter into the question of amount of summation.
I fully agree with Dr Aslin that there is a large
variability of the amount of summation both in normals
and in strabismics. Therefore one can not make
statements on the amount of summation. I want to point
out simply that summation is absent in the presence of
artificially induced or spontaneous diplopia when an
input to the two eyes is still present. There is no
suppression and yet integration of information from the
fixation areas does not take place. The same happens in
patients in which you are able to find suppression with
striated glasses.
von Noorden: Dr Campos, what does it mean in functional
terms when the VEP is greater with both eyes open than
with one eye closed? Also, as a patient does not
suppress the entire retina, some input is received from
parts of the retina that are non-suppressed. How do you
explain then that you get a larger binocular VEP .from
the patient with ARC than from the one that was
suppressing?
Campos: I did not mean to elaborate on the or1g1n of
summation. I used VEP to demonstrate objectively
psychophysical findings and was able to find ~ good
correlation between the two methods. I have no way of
knowing the areas of which summation is an expression. I
fully agree with Dr von Noorden when he says that the
presence of suppression in strabismus does not involve
the whole retina. I am implying that one speaks of
suppression when it involves the fixation area in a
patient tested with striated glasses. Summation was
found in patients in which ARC was present with striated
glasses. Lack of summation was detected in those
patients in which suppression was present at the same
test. Finally, the peripheral retina, although not
suppressed, provides very weak or no VEP responses with
usual derivation.
Lennerstrand: I am a bit concerned over the use of VEP

for an objective evaluation of binocular functions. We
know that the stimulus parameters are very important
both in studies of binocular summation and interaction.
It has been hard to convincingly demonstrate orientation
selectivity or spatial selectivity with the interaction
technique in normal subjects (Jakobsson, 1985) 9 •
Therefore a word of warning is warranted although I do
not dispute the interpretation that you have made of the
VEP recordings that you showed.
Campos: Thank you Dr Lennerstrand. I am certainly aware

of your work done with dichoptic stimulation which is a
completely different experimental situation as compared
to mine. I am not correlating VEP summation with given
cortical functions. I showed that you cannot
differentiate normals from strabismics on the basis of
presence or absence of summation because you find
summation both in normals and in strabismics. I would
like to stress that summation was present in patients
with ARC as tested with striated glasses, whereas lack
of summation was present in the other ones. Summa.tion,
however it could be interpreted, has different
characteristics in normals and in strabismics with ARC.
Harcourt: I believe that some patients with manifest

strabismus of at least 5 degrees at the fixation
distance can make positive responses indicating
binocular depth pe:rception using the Lang two pencil
test. Does Dr Helveston believe that this simple test
assesses depth perception in the same way as other
stereoscopic tests? We have already heard that
stereopsis is unlikely to be present in patients with
manifest strabismus of more than 2-3 degrees on account
of the organization of the visual cortex and associated
areas. I would also like to draw attention to. the new
random dot stereogram test developed by Dr J. B. Weiss
(CERES, Paris) which can be used at distance, a very
useful advantage. The disparity is some 90 seconds of
arc at 3 metres.
Helveston: I can answer your question by saying that

both tests are reliably unreliable. I do not use the
Lang two pencil test routinely. I use it only if I want
to demonstrate to the family the deleterious effect of
loss of stereoacuity. I am neither an expert at carrying
out the test not interpretating it. On the other hand,
all stereo-tests that we do are quite contrived in that
9)Jakobsson, P. (1985) Binocular interaction in human
visual evoked potential. Linkoping University Medical
Dissertations, No. 192.
402 E.C. CAMPOS
they have monocular clues and they have testing

artifacts. Factors such as experience on the part of the
person being tested severely diminish their scientific
reliability. For me the ultimate test for stereopsis
occurs in the operating room. If one ocular is off one
and a quarter diopters, all sense of depth is lost while
operating in the anterior chamber. Perhaps for
scientific work we should rely more heavily on real
depth stereo testing.
Bagolini: I wish to project a picture taken from a paper

by Cynader et al. 10 related to anomalous retinal
correspondence. I want to remind you again what
anomalous retinal correspondence is. In strabismic
patients, for example in the case of esotropia, the
patient should see the image that he is looking at as
double; the image belonging to the deviated right eye
should be seen at the right side. The patient does not
see double because he is suppressing. After a certain
period of time he localizes the second image
superimposed on the one seen by the fixing eye.
Anomalous correspondence has then developed. I think
that Dr Cynader and co-workers beautifully demonstrate
electrophysiologically what happens in anomalous retinal
correspondence. The receptive fields for stimulation of
the two eyes are nearly superimposed in the cortex in a
region which does not correpond to the retinal area
stimulated in the deviated eye of a cat rendered
exotropic. This appears to be the first
electrophysiological proof of anomalous correspondence.
The explanation of ARC given by me (Bagolini, 1967)1 1
seems to be valid according to their observation.
Bicas: Stereopsis is not a necessary consequence of a

good binocular function and, definitely, is not a cause
of stability or coordination of eye movements. It is
not, either, a necessary or a sufficient cause of
perception of depth. So, 1) are not the tests for
binocular dis pari ties (i.e. , stereopsis) merely
laboratorial, without indicating how the patients really
recognize their surroundings? and 2) do monocular,
amblyopic and/or strabismic patients develop
compensations to their perception of space?
Campos: Dr Helveston do you want to comment this?
10) Cynader, M.S., Gardner, J.C. and Mustari, J.M.

(1984). Effects of neonatally induced strabismus on
binocular responses. Exp. Brain. Res., 53, .384-399.
11) Bagolini, B. (1967). Anomalous correspondende:
definition and diagnostic methods. Doc Ophthalmol. 23 ,
346-386.
Helveston: I agree there are many monocular mechanisms

that can be used in place of stereopsis. However, these
are effective primarily at distance. On the other hand
there are certain things primarily found close at hand
that only stereopsis will allow us to accomplish. Ocular
surgery may be a good example. Dr Campos, objected to my
comment in the text that stereopsis was the anchor for
binocular vision. He objected specifically to the
metaphor implying a positive active nature for
stereopsis as present in the term anchor. Perhaps
instead we should think of stereopsis as keystone of
perfect binocular vision. This implies a little more
passive nature but it also denotes an important role if
you happen to be standing under the arch. I would
therefore propose that we are standing under the arch of
vision and that the keystone is stereopsis. In our own
research our goal is to find out whether stereopsis
stands beside, behind, or leads the band with regard how
our eyes work together. My initial impression is that
stereopsis or a parastereoscopic phenomenon is the key
to normal binocular vision.
Haase: I am sorry to argue again on methods. My qustion

is directed to Dr Helveston. I see a problem in the Lang
stereo test with random dots and I would like to ask you
whether you share my objection. The test contains 3
figures and each figure differs from the others
regarding size and disparity. Changes in two parameters
seem not to be a good idea to examine stereopsis.
Helves ton: I included the Lang test for completeness.

This is an important test because it is the only one
that uses random dots but needs to anaglyph glasses.
Camoos: I can only add that the Lang test is incapable

of differentiating microstrabismus patients from
normals.
Held: I would like to take what is for me a strange role

here and advocate the clinical use of redundant
stereograms in testing strabismics for stereopsis. I do
not know why this seems to be a forgotten technique.
Does everybody know what I am talking about? I mean what
is sometimes called the wallpaper effect where you have
repeated patterns and the exact angle of vergence is not
critical. You get periodic reproduction of corresponding
images and that should be a useful procedure for testing
where you have an angle of anomaly.
Blakemore: I think it is important to emphasize that

harmonious anomalous retinal correspondence (HARC) need
not require a system of binocular neurons with receptive
404 E.C. CAMPOS
fields on geometrical non-corresponding retinal points.

In my experience, artificial squint of any significant
angle destroys binocularity in the cortex and does not
cause a corresponding shift in relative positions of the
receptive fields of binocular cells. However, HARC could
still occur with systems of independent monocular
cortical cells, as long as each single object in space
gave rise to the same sensation of egocentric direction
through the two eyes, despite the presence of the
deviation. However, such a system of monocular cells
could not, presumably, support true stereopic vision.
Aslin: I would like to direct this question to Dr

Crawford. A surprising aspect of your data is that
monkeys with essentially no binocular neurons have
straight eyes. Your data from deprived monkeys are
analogous to what we think is going on in normal
infants, i.e. they do not have a functional binocular
system until 4 months of age; yet their eyes are
straight. I can think of four possible explanations for
your finding. First, you did not begin the deprivation
until 30 days after birth so each animal had 30 days of
post-natal experience that could possibly have provided
some residual information to maintain eye alignment even
after long-term deprivation. Second, your monkeys did
have some cells that were binocular. Although it seems
implausible that the loss of 90% of the binocular
neurons would still allow the system to maintain
straight eyes, I suppose it is logically possible that
even a single binocular neuron would be sufficient.
Third, binocular information may proceed to the
oculomotor system without passing through the visual
cortex. Fourth, it seems possible that disparity
information could be carried by neurons that are
monocular according to definitions of ocular dominance
that have been used in electrophysiology. We know from
the work of Barlow, Blakemore and Pettigrew and Bishop
that disparity selectivity is determined by an
inhibitory interaction between the alignment of cortical
receptive fields. Perhaps a neuron classified as ocular
dominance group one or seven may have some residual
inhibitory influence that simply is not detected unless
you do the kind of experiment that measures the
correspondence between receptive fields.
Crawford: There are a couple of points that can be made.
The first being that even in the case of starting the
optical dissociation at 30 days of age some cell has to
be carrying the message to keep the eyes aligned. The
second point being that one can always say that there
must be ~ cells remaining. I do not think that there
is any way around that. Yet, to me it seems implausible
to invoke that explanation because a system that
requires the sort of precision as does the vergence

system would certainly require a greater signal
fidelity. Assuming that you start out with 100% of the
binocular neurons making a contribution and reduce that
to some 10% it seems unlikely that you still have
anything approximating the fidelity in the oculomotor
system that these animals appear to have. I would like
to stress the "appear to have", since we have not in
fact tested fine vergence movements in these animals.
That is in process. It is a distinct possibility that
the foveation process itself is sufficient to get the
system within some fusion range that we cannot detect by
either clinical or casual laboratory observation. These
animals may yet prove to have some vergence defects and
I want to express that reservation. But I do not have an
explanation for the fact that they seem to be so normal.
Yet I am very confident that they have very few
binocular cells. They have very few binocular cells in
two cortical areas, the striate and area 18 cortex. As
to the point that you make, that it is possible to make
up, de novo, a new set of binocular cells from monocular
cells at some higher order center. That would be a
possibility I suppose, but there is no evidence for it.
So I am assuming that if there is a paucity of binocular
cells in area 17 and area 18, and knowing that this
route happens to be the gateway to the rest of the
cortical brain, it seems unlikely that you are going to
find them in some other part.
Singer: I may sound repetitive, but even alternators are

capable of recognizing immediately with the left eye an
object which they have seen first through the right eye.
Thus, there must be convergence between the two eyes at
some level of processing. In response to Dr Crawford's
comment I wish to point out that there is experimental
evidence that strabismic cats which lack stereopsis and
binocular neurons in area 17 and 18 may preserve
binocular neurons in pre-striated cortical areas. There
is also evidence that animals raised under conditions
which disrupt binocularity such as alternating occlusion
nevertheless preserve the ability to fuse dichoptically
presented patterns and to distinguish features that are
only detectable when the two eyes are used
simultaneously. They preserve this ability despite
complete disruption of binocularity in area 17.
Bagolini: I want to add something to Dr Blakemore's

comments. Concerning the experiment of Cynader 10 , I must
stress that anomalous retinal correspondence is a purely
binocular affair. When the patient looks only with one
eye the directional localization of the retinal points
becomes normal. Dr Cynader should have made these
406 E.C. CAMPOS
experiments binocularly but from reading the paper I

could not understand if they were binocular or not. If
he has not done the experiment binocularly, perhaps his
results may be an artifact. But anyway this experiment
gives me the opportunity to draw attention to this
neuroelectrophysiological problem. I would also like to
say that in strabismus, and especially in rather large-
angle strabismus, a retinal point, say in the temporal
side, may become correspondent in an anomalous way to a
nasal point of the contralateral eye. Evidently, the
information must be transcallosal if we accept Cynader's
experiment. Even in relative stereopsis the information
may sometimes be transcallosal.
Haase: Why are we so unsatisfied with our tests? Because

they do not reflect patients reports after surgery.
Patients improve in behavioural tasks - running up and
down stairs, riding bycycles and so on - but we cannot
find stereopsis. In this context, the Lang pencil test
could be useful, but it has to be standardized and
transferred to a quantitative procedure.
Campos: We have now to move on towards the last subject

of our session, namely amblyopia.
Maffei: I have a question to Dr Levi. At a certain point

of your presentation, you pointed out that amblyopia
could be explained on the basis of a change in the
magnification factor. Could you elaborate more on that?
Levi: I do not think I said that directly. However, I

drew a parallel between the peripheral visual field of
normal observers; and the central field of strabismic
amblyopes. Many of the functions are both qualitatively
and quantitatively similar in the two visual systems.
For example, we find similar decoupling of vernier and
grating acuity in the peripheral visual field as we do
in strabismic amblyopes. For this reason we have
suggested that in a sense the fovea of strabismic
amblyopes is like the normal peripheral field, and it
might be so because it is spatially under sampled, i.e.
the density of cortical receptive fields is lower than
it is in the normal fovea.
Campos: Thank you Dr Levi. Let me ask both Drs Levi and
Westheimer why the visual cortex provides this very fine
hyperacuity. Can you speculate on the fact that relative
localization may be more important than other functions?
Westheimer: I think what is needed is a change of
perspective. We used to think that visual acuity was the
final grain of the visual space but it is limited by
diffraction phenomena, the point spread functions and so

on. If we look at the other way round, however, then the
fine localization ability can become the standard and
its low threshold is not surprising.
Levi: I agree fully that hyperacuity represents a
measure of the spatial grain of the visual system as Dr
Westheimer pointed out. A related question is why do we
have such a fine spatial grain? One possibility is that
the mechanisms which perform hyperacuity have a more
general task of segmenting the visual scene in an
analyzing form.
Aslin: I do not think Dr Levi had quite enough time to

present the last schematic of the different receptive
field sizes in different types of amblyopia and I am
curious to hear that explanation.
Levi: The figure showed three sets of receptive fields.

The first set were small receptive fields that had
excitatory zones of one minute, and were separated by
one minute. I schematized those as being normal. They
are small receptive fields that are finely sampled. The
second row showed receptive fields that were larger and
were also sampled at larger intervals. They were twice
as big and sampled at twice the distance. We schematize
these as the kind of mechanisms that might be present in
anisometrophic amblyopes where reduced contrast
sensitivity eliminates the smallest receptive field. The
remaining mechanisms are larger, but they are efficently
sampled. They are sampled in the same proportion as the
small receptive fields found in the normal fovea. The
third row showed receptive fields that were also
enlarged by a factor of two, showing the kind of
convergence that Colin Blakemore mentioned yesterday,
but their separation was increased by a factor of four.
These mechanisms are undersampled. What I suggested is
that for detection of stimuli, the spacings of receptive
fields is not very important. However, for fine spatial
discrimination, the grain of the receptive fields might
indeed be quite important.
Campos: I would like to ask the audience whether all
those various functional changes which have been found
in amblyopia, namely the ones that were described by Dr
Levi and also the ones described by Bedell and Flom
(1981)1 2 are typical of amblyopia or whether they could
12) Bedell, H.E. and Flom, M.C. (1981). Monocular

spatial distortions in strabismic amblyopia. Invest.
Ophthalmol & Visual Sci. 20, 263-268.
408 E.C. CAMPOS
be an expression of an altered visual performance. Which

elements do we have that some of the partitional errors
present are typical of amblyopia? Probably Dr Levi and
Dr Haase may want to answer this.
Haase: We found that crowding phenomenon in organically

based low vision patients is not expressed in such an
amount as it is in amblyopes. Using the procedure
suggested by Bedell and Flom (spatial distorsion) we
examined organically based low vision patients. They did
not demonstrate a dis torsion comparable to the results
obtained in strabismic amblyopes.
Levi: I think we cannot really answer your question at

this point. For instance we can say that resolution and
hyperacuity are both abnormal in amblyopes of all types.
They are probably abnormal in many disease processes
too. The relationship between resolution and hyperacuity
in anisometropic amblyopes follows what one might expect
from blur. In strabismic amblyopes we see a different
relationship between resolution and hyperacuity. We can
reproduce this in the normal periphery, so I would not
want to make any arguments that these are pathognomonic
of amblyopia. However, they might provide us with some
idea of the types of dysfunctions, and the kind of
neural elements that might be abnormal.
Helveston: For 5 year old children we have used 20/40

vision in one eye or more than one line difference
between the eyes as the level of failure for vision
screening. Children are also required to appreciate a
gross stereopsis target (3000 sec disparity) to pass.
Campos: Would anyone comment on this?
Levi: Amblyopia represents a syndrome. Bedell and Floro's
results have in fact suggested that you can find some of
these amblyopic symptoms in patients who appear to have
normal acuity. So there is a degradation of vision that
is probably very fine, and where you draw the line is
the question that you are asking. The real question is
what you want to draw the line for. If it is for
screening purposes, then you want to draw a line that
would distinguish patients who have significant degrees
of loss, but not be so fine as to catch everybody. I do
not think anybody has really done a well designed study
looking at false-alarms and hits for an amblyopia
screening test of that sort.
Siostrand: Dr Levi, you have described the

psychophysical consequences of image degradation
primarily at the threshold level. Can you comment upon

the consequences under suprathreshold conditions of the
different types of amblyopia.
Levi: It is well known that there is abnormal contrast

sensitivity in amblyopes, and that at high contrast
levels amblyopes show very little deficit in contrast
perception. However, in our hyperacuity experiments, the
stimuli were all very visible. What is near threshold,
is the visibility of the fine changes in relative
position. The consequences for amblyopes in every day
life may be the kinds of things that Dr Haase presented,
where amblyopes have difficulties in doing fine detailed
discriminations, and in reading if they are forced to do
so with the amblyopic eye.
Sj6strand: Just one extra comment. One very important

function in daily life is reading. How much do we know
about the relation of impaired acuity to reading
difficulties under monocular conditions?
von Noorden: I would like to ask Dr Levi whether he has

ever compared his findings in amblyopes with a group of
patients whose visual acuity is reduced by similar
amounts due to organic reasons. That would quickly give
the answer whether we are dealing here with a
disturbance specific for amblyopia or something that can
be explaiend merely by the reduced visual acuity.
Campos: I fully agree with Dr von Noorden on this. The

difficulty from the ophthalmological point of view is to
be able to find an organic pathology which takes place
so early in life as amblyopia and gives an important
visual acuity decrease without unsteady fixation. In a
macular dystrophy, vision is either so low early that
nystagmus develops or the basic condition is present but
the visual acuity drop takes place later on in life and
therefore it is difficult to compare this condition with
amblyopia.
Hohmann: Visual acuity must first be standardized with a
population of normal sighted subjects of different ages.
This is not done for most of the visual acuity tests.
Our secreening criterion is at 0. 8 visual acuity for
optotypes with >30 min of ARC spacing (=single
optotypes) for children from 3 and half years onwards.
If you put your screening criterion low as 0.5 you will
miss a lot of amblyopes. Regarding reading ability, it
is rather difficult to measure it precisely. Testing
visual acuity with very little inter symbol space (~2 to
3 min of arc) by asking the patient to "read" symbol by
symbol is a mimic situation to reading in children when
410 E.C. CAMPOS
they are about to start learning.

Haase: We compared reading ability and acuity. People
who are experienced in reading and can already read well
show better performance in the reading than in the
acuity test. Reading tests cannot be used to detect
amblyopia in very young children.
Campos: When speaking of reading I did not refer to the

interpretation of words. I was pointing to the speed at
which the patient is able to read one letter after the
other of an optotype. This does not take into account
the association problem that you pointed out.
Maffei: I noticed in Dr Haase's drawing that the colour

was somewhat different when the drawing was done with
the good or with the amblyopic eye open. The saturation
of colour was different, the contrast was different. So
my question is: is there such a thing as colour
amblyopia?
Haase: Several authors made investigations on this

topic. They did not find any specific loss of colour
vision in amblyopic patients. I assume that "pale"
painting reflects just a reduced sensitivity.
Levi: Most clinicians will agree that strabismic and

anisometropic amblyopia differ in their clinical
char3cteristics and I think that there is psychophysical
evidence for functional differences. I am wondering how
those differences might be implemented neurally. There
is clearly a difference in the input to binocular
neurons on corresponding points. In anisometropia one
eye has a strong input, whereas the other eye has a weak
input but they are both spatially coherent. In
strabismus the stimuli to corresponding points at least
to physiologically and anatomically corresponding
points, are spatially and temporally non-coherent so
that it is a clear possibility for neural differences.
However, the same difference occurs in alternating
strabismics who do not have amblyopia and a critical
question is why some strabismics alternate while others
develop strabismic amblyopia when there is no refractive
imbalance. I think of several possibilities. One is that
in fact those strabismics have a non-comitant deviation
which is not evident when they come in to be tested, so
there is in fact a reason that they prefer to fixate
with one eye. Another possibility is that there might be
some developmental differences in the time of arrival of
afferents from the two eyes to the cortex that gives one
eye a competitive advantage over the other.
Lindstrom: I have just a question to the clinicians.

Could the eye preference somehow be similar to the
handedness problem?
von Noorden: This was studied by Dr Burian who
established that there is coincidence between handedness
and eye dominance in only about 80%. There is no firm
correlation between handedness and ocular dominance.
Campos: This session now comes to its end. I think it

was shown that all the aspects involved with
psychophysics of binocular and monocular perception in
strabismus and amblyopia are of utmost interest. We have
at least been able to raise some questions which need
further clarification.
33
CONCLUDING REMARKS
GERALD WESTHEIMER
Listening to three days of presentation from clinicians,

clinician-scientists, and scientists in the field of strabismus,
amblyopia and binocular vision provokes certain general
observations. Since the individual contributions are already
summaries of much more extensive work, further abstractions are
probably not in order because they would often lead to
misrepresentations -- a single, simple statement would not do
justice to the scope of the findings, nor allow for the reservations
that need usually be attached to them in order properly to delimit
their general range of validity.
It is possible, however, to come away from such a meeting with

an altered, perhaps even enhanced, perspective of the state of the
art of the subject, both in the clinic and in the experimental
laboratory -- and by no means do I see a contradiction in terms when
I use the word art in connection with either of these pursuits.
First of all, it is essential to realize what separates the

nature of our meeting from one, in say, physics, chemistry or even
molecular biology. There is no theory in our·subject worth the
appelation. I am grateful to the organizers for having spared us
detailed mathematical calculations. I can see no reason for this
kind of formulation, neither in the clinic nor in the binocular
vision laboratory. Nothing that we do warrants the assumption of
the rigor inherent in mathematics and I foresee no immediate change
in this. This is an experimental science, pure and simple, and the
results we deal with are soft, we might as well admit it and not
give ourselves airs of theoretical physicists or biocyberneticists.
This does not mean, however, that our research is not, or at

least should not be, analytical. It is here that, I think, the
greatest advances are still to come. On a first look in a new
subject we cannnot be exhaustive, though there is never an excuse to
be purely superficial. As an example, let us consider the situation
relative to binocular cortical units. The question of the
413
414 G. WESTHEIMER
confluence of signals from the two eyes to the cortex was taken up
on several occasions in the history of neuro-ophthalmology. Both
from anatomy and from neural injury, expecially after the first
world war, it was known what happens to vision in various parts of
the visual field when there are gross lesions of various parts of
the cortex. But the microcircuitry was left open. The first
electrophysiological recordings from the cat cortex by Hubel and
Wiesel showed many units that responded to stimuli from both eyes.
Anatomists still did not rise to the bait and it required these
workers another decade to demonstrate ocular dominance columns. Now
we know, of course, that the two eyes' signals are kept apart
through the geniculate and, in the monkey, into a layer of mono-
cular, center/surrounds cells in layer IVc. Electrophysiologists,
psychophysicists and clinician-scientists did rise to the bait,
however, and proceeded to draw a great number of conclusions, many
of them quite unwarranted. That lots of single units respond to
stimuli to both eyes, does not necessarily mean that the substrate
of single binocular vision and stereopsis resides in these units.
For disparity processing, for example, the contrary can be argued,
for a single monocular stimulus can be paired with one to the left
of it coming from the other eye for the purposes of detection of
crossed disparity, and another to the right, for the purposes of
uncrossed disparity. And these can be many degrees apart and with
quite a different orientation. Yet I know of no better indication
of good binocularity than the detection of disparity for the
purposes of fusional movements and stereopsis.
This example is intended to stress the need for calm and

detached analysis of experimental findings to put them into the
appropriate frame for an understanding of the functioning of the
complete, intact organism and the application of this understanding
to diagnosis and therapeutics.
Unfortunately there is a further factor to be considered,

namely the sophistication of modern research procedures. Once the
first cursory descriptions are in, the situation in our state of
knowledge often becomes muddied again rather than gaining in clarity
and validity. An example of this was our discussion of the nature
of extraocular muscles and proprioceptions. Those of us brought up
on the paradigm of clarity in textbooks, Cogan's Neurology of the
Ocular Muscles, had no difficulty about the subject. We knew about
the fibers, about grapelike endings, about lack of proprioception.
Now we have six types of fibers, orbital and global layers, patients
who do not past-point, or at least not as much, and so on. There is
no immediate solution to this dilemma. We are only in the beginning
of a long haul to an eventual thorough knowledge of the field.
Patience and stamina are needed.
Rather than becoming angry, or impatient, at the scientist, or

ignoring him, the clinician should employ quite a different
approach. He should try to understand exactly what the laboratory
results really say (not just what the titles of the papers say),
CONCLUDING REMARKS 415
what the strength of the claims are and what their possible
application in the clinic. Usually one cannot hurry laboratory
research, particularly in days of reduced funding, opposition from
animal rights groups, and brain drain to highly appealing subjects
such as molecular biology and artificial intelligence. But
clinicians, particularly, should insist on recognition, by these
scientists who claim to care, of the validity of their observations
on patients. This is one of the most important talking points
between clinicians and scientists. Clinical observations can be
strengthened by clear measurements, by statistics, by proper double-
blind protocols and so on. Much needs to be done to improve and
standardize such measurements as eye position, eye movement, image
quality, e.r.g., visual evoked potentials and so on. But once made
convincing, such clinical results cannot be ignored in our global
picture. No claim can be made about the universality of a
formulation that does not include such clinical findings. A
sterling example of what I mean is anomalous retinal correspondence.
This syndrome was seen as important early on in the history of
binocular vision, for example in Tschermak's voluminous writings
(though it should be added that he himself had this condition). But
little has been heard of anomalous retinal correspondence since,
even in discussions of the development of cortical connections.
It needs, however, to be remembered just how few animal models

are really appropriate. In oculomotorics, for example, the
beautiful work on rabbits can be related to cats to only a limited
extent. The same lack of transference applies to the superior
colliculus of the cat as compared to the monkey. Can lid suture in
monkeys really be regarded as paradigmatic of strabismic or
anisometropic deprivation in man? Perhaps there is yet another
technical round to be joined, where modern opto-electronic means can
be found to simulate the environmental change to which animals can
be subjected so as to mimick most effectively the visual environment
of an afflicted patient.
Scientists are, of course, never overwhelmed by the teaching of

clinicians. There are good reasons. We have alluded to the
frequent mushiness of the observations. More disconcerting,
however, is that the teaching is often rigid and dogmatic. It is
refreshing, therefore, to learn at meetings such as this, that some
things lose their sacredness and immutability. The AC/A is a noted
victim this time.
I have another reason for urging patience. At this meeting we

heard only passing mention of multiple kinds of synapse and their
pharmacological characterization. In fact we are in the middle of
an explosive growth in knowledge in neurobiology. There is a
cascade of new findings in genetics, in developmental neurobiology,
in the molecular biology of receptor molecules, in pathway tracing
by novel neuroanatomical techniques, in neural net theory, in
rigorous psychophysical delineations of animal behavior. Once this
knowledge has become firm, it is certain to be fruitfully employed
416 G. WESTHEIMER
in strabismus and amblyopia and give meaningful substance to the

title of today's symposium - Experimental Basis for Advances in
Clinical Management. In the meantime we can take real pride in the
role played by clinical findings on strabismus and amblyopia in the
advance of knowledge on broad scientific fronts of brain and mind.
34
FINAL DISCUSSION
Moderator: GERALD WESTHEIMER
Westheimer: We will take up a series of topics in

sequence. The first one concerns the etiology of
strabismus and, in particular, genetic factors. Are
there any comments?
Campos: One could say from the clinical point of view

that there are some factors which predispose to
strabismus and which are hereditary. For example,
hyperopia is a condition which could be hereditary and
an accomodative type of esotropia could develop from
it. Therefore, if one has a parent with this type of
condition one has to refract the child early in life in
order to prevent the onset of strabismus.
Si6strand: Since we know that hyperopia is a high risk

factor in creating squint, I think it's very important
to look longitudinally on refraction changes from birth
during the first years and study the factors involved.
Deeper knowledge of the emmetropization process seems
very important.
Helveston: Dr Lennerstrand on the first day of this

meeting showed some MRI studies of patients with
hydrocephalus, Arnold-Chiari malformation and "A"
pattern. We have a large group of patients with
rnyelorneingocele and we observe that these patients have
a very high incidence of A pattern, overaction of the
superior oblique muscles and dissociated vertical
deviation. In addition they have a specific abnormality
of the palperal fissure. This is not unlike the
relationship of Down's syndrome patients to each other.
We have not been able to characterize exactly what is
417
418 G. WESTHEIMER
wrong with the fissure except that the lower lid

appears to be straight and the fissure to be
antimongoloid. This might be a genetic type of
strabismus.
Harcourt: A-pattern esotropia in patients with

meningomyelocele is common when associated with
previous episodes of uncontrolled hydrocephalus, and
could be the result of persistent inferior rectus
weakness following partial recovery of bilateral third
cranial nerve palsies.
Levi: We have heard tantalizing hints from Dr Helveston

and others of possible central sensory anomalies that
might lead to early onset strabismus, and it might be
interesting to consider the development of tests for
the parents to see if there might be predictive factors
for such abnormalities which may have a genetic basis.
von Noorden: There is another explanation for the "A"

pattern in hydrocephalus. Bossing of the forehead leads
to anterior displacement of the trochlea that changes
the relationship between the planes of the inferior and
superior obliques, causing overaction of superior
obliques that causes an "A" pattern.
Bagolini: I would summarize this problem this way.

There are certain risk factors which are inheritable,
e.g. hyperopia and anisometropia, and they have already
been mentioned. Then there are children from parents
which have strabismus but not hyperopia and
anisometropia: one may suspect an inheritable lack of
fusion as suggested by Worth. I think this is a real
possibility. Another factor is related to muscle
abnormalities: for example Duane's syndrome is
inheritable but unfortunately we have no way to prevent
it. On the contrary in the previously mentioned cases
we may have some tool to prevent it. Muscle fibrosis
and some inheritable variation in skull development may
facilitate strabismus and constitute a risk factor. In
many inheritable neurological diseases there is a risk
for strabismus but then strabismus may become a minor
problem.
Westheimer: How about animal models? Are there any
comments, especially with regard to the future? Can we
rely on suturing eyelids and raising up animals in the
dark or should we go further and if so, how?
Hoffmann: First a more general comment. I think we

really have to abandon to think of a "phylogenetic"
line rat-rabbit-cat-monkey-man. On the other hand there
FINAL DISCUSSION 419
have to be common principles and mechanisms in these

species that could be investigated and the
clarification of them could be very helpful. With
respect to the question why there are no strabismic or
so few naturally strabismic animals or monkeys, I think
one has to keep in mind that an animal or a group of
animals which developed frontal eyes and stereopsis had
an advantage by doing so. That also means that those
individuals developing strabismus were at a severe
disadvantage and there was a very strong selection
pressure against them. But I am convinced that the
faulty mechanisms leading to strabismus are present in
the gene pool of these species. To me it would seem
feasible to select and specifically breed monkeys with
infantile esotropia.
von Noorden: It's not quite so that all conclusions

presented at this meeting were based on data from
lidsutured monkeys or those raised in the dark. We were
successful in producing strabismic and anisometropic
amblyopia in monkeys. We have also produced alternating
esotropia and showed that there is no amblyopia. The
problem with the strabismic model is that it cannot be
exactly equated to strabismus in humans because it was
artificially induced. Perhaps the very factor that
gives rise to infantile esotropia was not present in
these monkeys. We need to make other disciplines aware
of our need for a species where strabismus occurs
naturally.
Held: I think it would be a great mistake to make our

appreciation of animal research contingent upon the
availability of a model of a specific pathology. After
all, most of our basic insight have come from animal
research that had nothing to do initially with the
possibility of using the results as a model of a
specific disease syndrome.
Crawford: I would like to make one comment on the topic
of generalization to clinical problems from the results
obtained from a variety of animal species. I applaud
the work employing a large number of animal models that
have provided irreplaceable evidence on various
functions that have implications for clinical problems.
However, many of these results cannot be uncritically
extrapolated to clinic application. There are some
critical differences between animal models that need to
be recognized, for example results obtained in the cat.
The cat and the monkey clearly have different central
nervous system projections. That is not to say that the
vast amount of information that we have gained from
studying the physiology of the cat central nervous
420 G. WESTHEIMER
system is not of value. The point is that the pathways

to the central nervous system are different between cat
and monkey, and is indeed different from that of man.
Westheimer: We have as our next topic a challenging

one. How can we enhance the acceptance and validity of
clinical findings, for example anomalous
correspondence?
Campos: As far as anomalous correspondence is

concerned, it has both advantages and disadvantages. It
does support an anomalous binocular cooperation but it
is a drawback for recreation of a normal binocular
cooperation. This is what happens in humans. It would
be interesting to look for a correlate of ARC in
animals. Something should be done in this direction.
This is very important particularly because anomalous
correspondence and small angle deviation is the final
condition in which many patients usually end up after
treatment. A final comment on psychophysical testing.
It is very different to test a patient than a subject
in a laboratory of experimental psychology. This is
also a subject we should discuss, i.e. how to suggest
ways of testing reliably but so that the subject is not
exhausted.
Crawford: A comment on the question as to what the

clinician might do to make his clinical experience more
pursuasive. I am concerned at the meeting about a
number of things that have to do with a disparity
between the kinds of data that are collected in the
clinic, and some of the physiological results reported
here. For example there are many reports from tests
done in the clinic citing examples of stereo acuity in
patients with strabismus, and anomalous retinal
correspondence. A number of reports of that sort raises
questions about the manner in which these data are
actually collected. For example we have heard the
comment that random dot stereograms (RDS) really should
be a tool of the laboratory, rather than a tool for
clinical usage. However, one of the things that the RDS
actually does, and does well, is that it separates the
variable of horizontal disparity from a large number of
other confounding variables and allows one to make a
very clear discrimination based upon horizontal
disparity alone. A number of other tests used for
determining stereo acuity have other factors that
contaminate the results.
Sjostrand: I think it is important that the clinician
carries out longitudinal instead of transsectional
studies to elucidate the relation between refractive
errors and squint or amblyopia. It is also crucial for

us clinicians to separate factors that could be related
to underlying biology (i.e. sensitive periods) and
other factors that are social due to complicance
problems during treatment. Dr Haase presented the
results of amblyopia treatment as related to the age at
which treatment started. I think these results could be
explained equally well as a result of problems with
compliance. When the children grew older they have
greater objections to the occlusion treatment. We have
found in an ongoing study trying to relate the results
of amblyopia treatment to age that you can almost
normalize the vision up to the age of seven in 95% of
the children if you have a good compliance to
treatment.
Blakemore: I think that psychophysical studies will be

a crucial intermediate between clinical work and animal
research. For instance psychophysicists could point out
appropriate kinds of test that critically distinguish
clinical conditions and which are sui table for use in
the clinic and the animal laboratory, thus providing
the opportunity for direct comparison of result.
Lennerstrand: Correlation between strabismus and other

ocular motor dysfunction is very important particularly
in analyzing infantile esotropia in greater detail. We
need better equipment for accurate recordings of eye
movements in young children. We also need to correlate
strabismus with central nervous system lesions or
malfunction. The latter can be determined by means of
detailed neurological analysis of strabismic patients,
and the brain imaging techniques that we now have
available (CT-scans, MRI, PET-scans, evoked potential
mapping and so on).
Held: I have been an observer for many years of the
sociology of the research scene where psychophysicists
work with ophthalmologists. As far as I can see the
real successes will come when psychophysicists are
brought into the clinical setting as more or less equal
colleagues and participate as such. In general, with
some exceptions, ophthalmologists are not researchers
and to improve the quality of research they need the
expertise of trained researchers. But these
psychophysicists must be equal participants in the work
that goes in within a clinic if the outcome is to be
successful.
Bagolini: This is a minor point, but perhaps not so
minor. When we have to compare animal experiments with
clinical experience the physiologist should care not to
422 G. WESTHEIMER
have animal with paralytic strabismus: they should try

to develop in animals a comi tant type of strabismus.
Certain aspects that we want to study such as anomalous
retinal correspondence can develop more easily in
comi tant strabismus and less easily or not at all in
paralytic strabismus.
Westheimer: We move on to the next topic. What is on
the horizon in therapeutics?
Steinbach: The variability of strabismus surgery is
puzzling. Are there other factors (like inadvertent
damage to proprioceptors) that could be playing a role?
Could a good resection result come about because all of
the palisade endings are removed; could a good
recession be one that leaves these endings untouched? I
certainly do not know the answers but I wonder if
attention to what the surgery is doing to a muscle
(beyond the simple mechanical analysis) might be
helpful in understanding predictability of the results.
Helveston: There has been a significant improvement in

techniques of strabismus surgery in recent years. Among
these improvements is the abandonment of marginal
myotomy as a primary weakening procedure. Better
sutures and needles, more emphasis on anatomy of both
muscle and fascia, improved anesthesia, and early
diagnosis and treatment have also been important
factors. Considering surgery is done in the most
capable of hands, it is my opinion that the next
important break through in the treatment of strabismus
will come from our colleagues in the basic sciences.
Campos: I agree with Dr Helveston on the fact the
certain technical improvements have been made in the
last 25 years. The main problem is not so much the
technical aspect, but the indication for surgery, and
this is still in the hands of the ophthalmologists. We
have seen during this meeting that there are mechanical
as well as innervational aspects involved in the
deviation. Probably there may be some peripheral
aspects involved as well. All these elements may
explain the unpredictability of our surgery.
Harcourt: We should consider the questions of
prophylaxis in relation to the etiology of strabismus,
especially accommodative esotropia. It is still not
really clear to what extent early prescription of
spectacles for young patients with significant degrees
of hypometropia and/or anisometropia decreases the long
term risk of strabismus, especially when there is no
relevant family history.
Bicas: I entirely agree with Dr Campos. Surgery of the

extra-ocular muscles may be very easy to perform but
still very difficult to plan. I think that we still
know very little about how to evaluate innervational
and mechanical factors correctly, and that we can
improve the use of passive forces for the surgical
treatment of strabismus.
von Noorden: There is also a need to look into
pharmacological means to prevent and to reverse the
deleterious effects of visual deprivation and binocular
interruption in infancy.
Eggers: I think we have perhaps just begun to improve

muscle surgery. With modern technology and
instrumentation we are on the verge of actually being
able to characterize the mechanical and plastic
properties of the orbital tissues for the first time.
We have the prospect of being able to do experimental
surgery on a mathematical mode, which is something we
have not been able to do before. I think we also are
facing the prospect in future years of having devices
such as artificial muscles which may improve what we
can do surgically.
Westheimer: Are there any words on non-invasive

procedures?
Schor: Now that we are beginning to develop some

comprehensive models of certain binocular disorders
such as accommodative esotropia and amblyopia, we can
develop non-invasive therapeutic procedures which
result in a functional correction of binocular vision.
The key-factor here is that these procedures must rely
on models of the disorders so that the therapy can be
aimed at solving the problems underlying them.
s iostrand: I would like to emphasize and underline Dr
Harcourt's comment that the major group of squint may
be preventable. The group of strabismus related to
hypermetropia and the effect of intervention can be
studied by epidemiological tools and such studies are
to my mind of great importance in the future.
Westheimer: There is still something to be said how to
go about developing future contribution to this area.
Is teamwork a part of this game?
Helveston: In our work with infantile esotropia it
occurred to us that the segment of the health care
delivery team that is involved in the care of the
424 G. WESTHEIMER
children during the first 6 months of life is naive

with regard to visual development and strabismus. It
has been an enlightening experience to work with
pediatricians and others who take care of the children.
I think that if we are going to make progress toward
improved treatment of our children, we should involve
our other colleagues in this awareness.
Lennerstrand: In view of our rather limited

understanding of the involvement of oculomotor control
systems in strabismus, the following areas are
suggested for further clinical research:
- Orbital mechanics, especially muscle anatomy studied
with CT-scan and NMR-imaging, and eye kinematics with
biomechanical models. Models could be used to predict
outcome of pharmacological and surgical therapy.
Extraocular muscle structure, both in normal cases
and in strabismus, using ultrastructural and
histochemical techniques. Structural changes after
denervation and reinnervation are also important to
establish.
With regard to extraocular muscle function, motor
unit participation in different kinds of eye movements
would give information on brainstem mechanisms in
strabismus. Relation between force development and eye
movements are of importance for models of eye muscle
and general orbital mechanics. Knowledge about firing
patterns of EOM in relation to force generation can
give insights into contractile functions of eye muscle.
Comparison of firing patterns in antagonistic and
synergistic eye muscles can reveal brain stem control
patterns of eye movements (exemplified by Donder's and
Listing's laws). Thus, EMG recordings, particularly at
the single motor unit level, would seem of great
importance for future functional studies of the human
oculomotor system.
- Studies of central (brainstem and higher) mechanisms
in strabismus. These studies have to be based on
analysis of eye motor function measured with accurate
eye movement recordings. The findings have to be
correlated with pathological changes in the CNS,
determined by postmortem examinations or by means of
CT-scan, NMR-imaging, PET-scan, evoked potential brain
mapping etc. Comparison between results in animal
experiments and clinical findings in humans are also
highly relevant, particularly if data from monkeys are
available. In this way details of the neurophysiology
of the vergence mechanisms and the AC/A relations and
of oculomotor plasticity can be revealed.
Effects of stimulation of extraocular muscle
afferents on oculomotor function should be studied by
means of eye movement recordings, EMG and eye muscle
tension. Activation by stretch or vibration of the

muscles could be done on normal and strabismic muscle.
- Motor control in other systems i.e. gait and postural
control and eye-hand coordination, may give insights
into the pathophysiology of strabismus. Deductions of
the type and location of the functional disturbance in
strabismus might be based on such data and complemented
with radiological studies. Also here experiments on
monkeys may be of value.
Development of oculomotor function, including all
parts from eye muscle structure and function, to
central control mechanisms and the different eye
movements systems. Such studies might disclose part of
the etiology of early onset strabismus, and possibly
also of microstrabismus and exotropia. Our knowledge on
the systems for vertical movements are restricted even
in animals.
- Vergence adaptation should be studied in normal and
in patients with different types of strabismus, and at
different ages. Studies on oculomotor adaptation to
disturbances in visual imaging induced by telescopic
lenses, dove prisms, aniseiconic lenses etc. are
suggested in normals and strabismics, adults and
children. Of great importance are also studies on
epidemiology and genetics of strabismus, and on the
long-term history of different forms of strabismus.
Westheimer: One of the real contributions of todays

meeting was to bring together scientists, clinician
scientists as well as pure clinicians and communicate.
This was done very nicely and we certainly are grateful
to the organizers of this meeting for arranging this
and for the hospitality accorded to us by the Wenner-
Gren Center Foundation.
35
GLOBAL STEREOPSIS
B. JULESZ
In the 25th year Jubilee Issue of Vision Research, published

in November, 1986, an article of mine appeared, ent~tled "Stereo-
scopic Vision." Since in this recent article I gave a review of
the main development of global stereopsis of random-dot stereograms
(RDS) from its conception (Julesz, 1960) to the present with a vast
bibliography, I feel reluctant to write a similar overview just
reworded.
Nevertheless, there are two topics that deserve more coverage

than given in the Jubilee Issue, particularly for colleagues with
interest in strabismus and amblyopia. These are: some recent
advances of the hysteresis effect of stereopsis; and some findings
on the early development of functional binocularity in human
infants.
The hysteresis effect of stereopsis under binocular retinal

stabilization (using close-fitting scleral contact lenses with
attached mirrors) was discovered by Fender and Julesz and only
recently revisited by Collewijn and Erkelens (1986), Hyson et al.
(1983) and Piantanida (1986). The main findings by Fender and
Julesz (1967) were as follows:
1) Initially the RDS had to be aligned within 6 min arc for

fusion (which is Panum 1 s area for RDS of 2 deg. arc extent) , and
then could be slowly pulled apart by 20 times more (120 min arc)
while fusion was maintained. Beyond 120 min arc of pulling fusion
would break. 2) For classical stereograms (i.e., a vertical line)
the limit of pulling would be much less (60 min arc) and refusion
would occur at about 2/3 of this limit (40 min arc). Two decades
later Piantanida (1986) repeated the Fender and Julesz experiment
but with two modifications. First he used instead of scleral
lenses SRI eye-trackers (based on the first and fourth Purkinje
images) for each eye to obtained binocular retinal stabi 1 izat ion.
Second, he only used RDS without the fiducial marks we (Fender and
427
428 B. JULESZ
Julesz) used to check breakaway. He corroborated all our results,

except that without the fiducial marks, he obtained refusion at
about 2/3 of the breakaway limit, while we had to bring within
Panum's area the RDS for refusion. Hyson, et al. (1983) found also
refusion of RDS at about 2/3 of breakaway, from which they conclud-
ed that there is a memory mechanism in stereopsis that preserves
the labeling of corresponding points even while fusion is lost.
Collewijn and Erkelens (1986) also confirmed the pulling effect of
20 times Panum's area after initial fusion, but claimed that for
refusion the RDS had to be brought back to Panum's fusional area.
Fender and Julesz (1967) did not specify whether the observed
extension of Panum' s fusional area was due to an increase of
fusional interaction or to the workings of the matching process
that eliminates false targets. In the intervening years Tyler and
Julesz (1980) found stereopsis in briefly flashed RDS in excess of
2 deg arc binocular disparities if the target was adequately wide
and the background was aligned. Obviously, in these brief flashes
convergence eye-movements could not be initiated and, therefore,
the binocular disparity of 2 deg arc matched the values found by
Fender and Julesz (1967) after slow pulling. It was still not
ruled out in the Tyler and Julesz (1980) experiments that in the
brief presentations the cyclopean target (center rectangle) did not
reach its final depth, but only appeared to jump towards the front
or the back. Therefore, Schumer and Julesz (1984) modified the
previous study by superposing a corrugated depth grating over the
cyclopean rectangle (that served as a depth pedestal). Observers
had to report the presence or absence of this corrugation, and were
able to do so for one deg arc binocular disparities.
These findings made it clear that the pulling effect in the

Fender and Julesz (1967) studies was related to the matching pro-
cess. Indeed, the small sized RDS with their fine grain had their
cyclopean square boundary at about 1 deg eccentricity, barely
enough to be inspected by shifts of focal attention (since the
images were stabilized on both retinae). This was a lucky choice
of the stimulus, since these focal attention shifts could still
separate the cyclopean square from its surround, while beyond the
square boundary the loss of visual acuity prevented inspection of
more peripheral areas. A recent experiment with large RDS (of 30
deg arc extent) by Erkelens (to be published) where only conver-
gence eye movements are stabilized, while conjunctive eye-movements
can occur confirms that maximum disparity is about 2 deg arc and
under free conjugate eye-movements there is not hysteresis. In a
way, Erkelens' technique permits conjugate eye-movements to slowly
solve the matching problem over a large field, which in the Fender
and Julesz (1967) study were solved in a small field and the slow
physical pulling played the role of conjugate eye-movements. So
the hysteresis effect found by Fender and Julesz (1967) and
Piantanida (1986) show the fine sturcture of the "label preserving
process" after binocular matching which is hidden when conjugate
movements are permitted.
GLOBAL STEREOPSIS 429
Let me note that the cooperativity of global stereopsis is

based on other experiments, too, particular on a study by Julesz
and Chang (1976) to be discussed later, who showed that a few
percent of unambiguous bias of a certain disparity would "pull" an
ambiguous RDS (that otherwise could be perceived in several depth
organizations).
In summary, the Erkelens and Collewijn (1986) studies that

permitted conjugate eye-movements missed the cortical shifting
process that preserves the labels of correctly matched areas having
different disparities. Perhaps this cortical shifting process can
be modeled by a neural shifting mechanism proposed by Anderson and
Van Essen (1987) based on an abundance of neurons in both the
magnocellular and parvocellular pathways. The magnocellular stream
includes the P-alpha class of ganglion cells, layers of the LGN,
4Ca and 4B of Vl. Layer IVB (Poggio, 1984) is the first stage at
which cells are binocularly driven. The parvocellular system
includes the P-beta ganglion cells, the parvocellular layers of
LGN, and layers of lower 4Cb, upper 4Cb, 4A and the supergranular
layers where binocularity and other integrative properties first
appear. For details, the reader should consult Anderson and Van
Essen (1987). Their shifter model thus provides a sensible explan-
ation for these millions of cells in layers heretofore, regarded as
simple relay stages.
The other topic concerns functional binocularity of infant

stereopsis. Julesz et al. (1980) measured evoked potentials (EP)
to dynamic RDS and correlograms in normal adults. They found that
if binocular disparity in a RDS alternated at a 3 Hz rate the EP
had a basic frequency of 3 Hz, too. However, for random-dot
correlograms which alternated between correlation and uncorrelation
at 3 Hz rate the basic frequency of EP was 6 Hz, thus doubled.
Petrig et al. (1981) found that in infants at about 3 l/2

months of age both RDS and correlograms yielded the ~ EP fre-
quency as the alternation frequency of the stimuli, while for 6
months of age infants the EP for correlograms doubled (as for
adults). From this we might assume that functional binocularity as
revealed by EP to correlograms develops sooner than geniune stere-
opsis. As if binocular correlation processing might be a precursor
for stereopsis and would manifest itself at an earlier stage of
development.
430 B. JULESZ
REFERENCES
Anderson, C.H. and Van Essen, D.C. ~1987). Shiftor Circuits:

A Conjunctional Strategy for Dynamic Aspects of Visual Processing
(to be published).
Erkelens, C.J. and Collewijn, H. (1986). Eye-movements in Relation

to Loss and Regaining of Fusion of Disjunctively Moving Random-Dot
Stereograms. Human Neurobiol., ~. 181-188.
Fender, D. and Julesz, B. (1967). Extension of Panum's Fusional

Area in Binocularly Stabilized Vision. J. Opt. Soc. Am., 57,
819 - 830.
Hyson, M., Julesz, B. and Fender, D.H. (1963). Eye Movements and
Neural Remapping during Fusion of Misaligned Random-Dot Stereograms.
J. Opt. Soc. Am., 73, No 12.
Julesz, B. and Chang, J.J. (1976). Interaction between Pools of

Binocular Disparity Detectors tuned to different Disparities.
Biol. Cybernet., 22, 107-119.
Julesz, B., Kropf!, W. and Petrig, B. (1980). Large Evoked

Potentials to Dynamic Random-Dot Correlograms and Stereograms
Permit Quick Determination of Stereopsis. Proc.Natn.Acad.Sci.,
U.S.A.,_I, 2348-2351.
Petrig, B., Julesz, B., Kropf!, W., Baumgartner, G. and Anliker, M.

(1981). Development of Stereopsis and Cortical Binocularity in
Human Infants: Electrophysiological Evidence. Science, N.Y., 213,
1402-1405. ~
Piantanida, T.P. (1986). Stereo Hysteresis Revisited. Vision Res.,

26, 431-437.
Poggio, G.F. (1984). Processing of Stereoscopic Information in

Primate Visual Cortex. in Dynamic Aspects of Neocortical Function
(edited by Edelman, G.M., Gall, W.E. and Cowan, W.M.), pp. 613-635.
Wiley, New York.
Schumer, R.A. and Julesz, B. (1984). Binocular Disparity Modulation

Sensitivity to Disparities Offset from the Plane of Fixation.
Vision Res., 24, 533-542.
Tyler, C.W. and Julesz, B. (1980). Stereoscopic Tilt and Size

Aftereffects. Perception, ~. 187-192.
Tyler, C.W. and Julesz, B. (1980). On the Depth of the Cyclopean

Retina. Exp. Brain Res., 40, 196-202.
INDEX
abducens motoneurons diagnosis of 381

inhibitory/excitatory 79 in humans
reticular pathways 79 animal models of 169-171
abducens nucleus 78 clinical relevance of
aberrant axonal pathways 84 169-171
aberrant development 247 in monkeys 227
abnormal binocular interaction, optokinetic reflex 89-96
evidence in humans 275-282 plasticity 230
AC/A ratio 18, 113, 136 rehabilitation strategies
accommodation 17, 112, 121-122 270
adaptation to blur 112 see also strabismic amblopia
phase-tonic organisation amblyopiogenic mechanisms 169
111-119 animal models 197, 276
accommodative accuracy 136 of amblyopia in humans
accommodative esotropia 170 169-171
early onset of 28-29 clinical relevance of 169-171
acuity 134, 232 anisometropia 169, 276-277,
see also specific types 279, 281, 387
adaptation after-effect 126 anomalous movements 118
adaptive mechanisms 82, 84, 282 anomalous retinal correspondence
adult onset strabismus 315-316 269, 311-316, 339-341
albino 171 ARC, see anomalous retinal
alignment behaviour 79 correspondence
alignment circuitry 80 area 17 177-178, 220-221
alternating strabismus, area 18 177
binocular co-operativity 267 Arnold-Chiari malformation 15
amblyopia 100, 169, 173, 179, arrest theory 280
207, 226-229, 231-232, axial length 204
247, 249-250, 254, 276,
279-280, 314-315,
369-370, 372, 375, balance 66-68, 73-75
377-378, 381 dynamic 73-75
anisometropic 386-387 balance and gait, abnormalities
clinical aspects of 381-389 of 14
deprivation 226 bilateral form deprivation 240
431
432 INDEX
binocular activation, contrast sensitivity 134, 227,

asynchronies 263 229-232, 248-249
binocular campimetry 312 control of limb movements 56
binocular competition 220, 239 convergence 17
binocular co-operation 191-192, proximal 122
267 voluntary 122-124
binocular interaction 275-276 convergence insufficiency 135
binocular rivalry 137-138, 247, co-ordinate frameworks 80
251-253, 281 cortical binocular cells 353
binocular summation 236, 260 cortical lamination 177
binocular vision 276 cortical neurons 222, 227,
anomalous 312-314 229-231, 260-262, 321
binocularity 247, 251-253 electrophysiology of 207-215
dissociation in testing 316 visual deprivation 207-215
and proprioception 339-340 cortical volume 177
bipolar cells 173-174 cortico-pretectal connections
bisection 375-376 90
bisection acuity 372 critical period 173, 179,
botulinum toxin 10, 41 192-193, 204, 235
brain damage 11 crowding 383
brainstem control 13 cytochrome oxidase 179
brainstem neuronal circuits,
phyletic organization of 77-86
brainstem neuronal matrices 81 defocus 231
brainstem neurons 77 dendrites 176-177
burst cells 118 dendritic arbor 176
dendritic lengths 177
dendritic spines 176
ca++-channels 265 loss of 178
CA/C ratio 113 denervation 41, 52
capillary vascular network 39 deprivation 385
cats 220, 225 see also specific types
Siamese 50 deprivation amblyopia mechanisms
cataracts 276-278 259-270
cerebellar flocculus 85 deprived eye 200
cerebellum 85 depth discrimination 321
cerebral cortex 82 depth of focus 136
compensatory head posture 24 descendent mammals 85
competitive mechanism 191-192 diplopia 136-138, 276, 311-312,
concomitant strabismus 7-9 315-316
treatment of 8-9 direct deprivation 240
cone photoreceptors 173 dissociated vertical divergence
cone receptors 134 24, 26, 99, 106
confusion 276, 311-312, 315 Dander's laws 18
congenital esotropia 366 Duane's syndrome 13, 29-33, 53
congenital ocular motor apraxia DVD, see dissociated vertical
15 divergence
congenital squint syndrome 99 dynamic random dot stereograms
conjugate movement 78 354
conscious awareness of position
56 eccentric fixation, bilateral
contractile speed 48 388
INDEX 433
EMG studies 53 eye movement 65-66, 69

ernmetropization 197 eye muscle afference 327
enucleation 279, 327, 329-334
EOG 365
EOM, see extraocular muscle fatigue 117
EOM development 49 fatigue resistance 48
EOM motor unit 48 filopodia 176
EOM stretch reflex 56 fixation 381
epigenetic strabismus 77 disparities of 136
esotropia 15, 118 foetus 174, 176-177
aetiology of 23-33 forces
classification of 23-33 active 65-67
clinical characteristics of passive 66-68
23-33 form vision deprivation
late onset of 7 169-170, 197, 236
see also specific types fovea 134, 173-174
essential infantile esotropia foveal cone mosaic 228
24 foveal cones, packing 228
evoked potentials (EP) foveal pit 173-174
to correlograms 429 foveal region 174
to dynamic RDS 429 frequency modulation 51
evoked visual potentials 281 Frisbee test 359, 363
exodeviations 8 fusimotor fibres 54-55
exophoria 15, 57, 135, 170 dynamic 55
intermittent 364 static 55
experimental refractive errors, fusion 139, 314
deprivation period 197-205 motor 137
extinction phenomenon 281 sensory 84, 136-138
extracellular recordings 170 fusion mechanism, abnormalities
extraocular muscle 39, 42, 170, of 11
338 fusional movement 125-129
denervation 41-43, 52 anomalous 127
fibre types 38
morphology 37-43
motor functions 47-59 GABAergic inhibition 261
normal 47-59 ganglion cells 173-174
postnatal development 39-41 gaze-paretic nystagmus 105
prenatal development 39-41 gaze pathways
psychophysical aspects vergent 77
327-334 versional 77
reinnervated 52 genetic strabismus 77
sensory functions 47-59 genetical innateness 82
strabismic 47-59 genicula-cortical connections
visual function 327-334 280
extraocular muscle insertions global stereopsis 323
80 grating acuity 247-251, 370-373
extraocular muscle gratings 382
proprioception 327-334 growth, of eye 228
clinical aspects 337-341 growth cones 176
visual functions 337-341
eye frontalization 79
eye kinetics 65-75 hair-like processes 176
434 INDEX
herpes zoster ophthalmicus 338, dendrites 176-177

341 infant 185-194
heterophoria 118 lateral geniculate nucleus 170,
human infants, normative 173, 175-178, 220-222,
oculomotor development of 227-230, 275, .281
133-140 normal 186-187
human retina 173 layer II 179
human vision layer III 179
development of 247 layer IVb 177
in infancy 247-251 layer IVc 177, 221-227, 230-232
hydrocephalus 15 LGN, see lateral geniculate
6-hydroxydopamine (6-0HDA) 264 nucleus
hyperacuity 369-372, 374, 377 Listing's laws 18
hypermetropia 170 LN, see latent nystagmus
hyperopia 198 localization 337-339, 341
hypertrophy 187 directional 125
incomitant strabismus 9 magnocellular cells 176

treatment of 10 manifest latent nystagmus 104
infant stereopsis, functional medial rectus
binocularity 429 inhibitory/excitatory 79
infantile esotropia 7, 23-33, reticular pathways 79
170 meningomyelocele 15
infantile strabismus 133 mesencephalic reticular
ocular motor phenomena 99-106 formation 117
infantile strabismus syndrome microstrabismus 8
99 midbrain neurons 83
inflow 322, 327-329, 334, middle superior temporal (MST)
337-339 area 92
inhibition 278-281 middle temporal (MT) area 92
innate crossed eyes 84 monkey eye 228
inner nuclear layers 174 monkeys 173, 177-178, 197, 207,
integrator 83 221-222, 224-226, 228-232,
internuclear neurons 80 236, 279, 353
abducens 85-86 foetal 173
oculomotor 85-86 macaque 169, 175, 177
prepositus 85-86 marmoset 175, 178
reorganisation of 267 rhesus 235
intrafusal fibres 54 monocular closure at birth 187
monocular deprivation 220-223,
225-227, 231-232
kittens 220-221 monocular segment 191
motion perception 101
motion signals, processing 269
Lang test 360, 363 motor units, distribution 49
late deprivation 191 multiple innervation conducted
late sensitivity 190 action potentials (MIC units)
latent nystagmus 24, 99, 48
102-106 multiply innervated,
lateral geniculate neurons 176 non-conduction fibres (MINC
abnormal 185-194 units) 48
INDEX 435
multipolar neurons 176 bands 222

muscle disuse SO changes 262-266
muscle spindles 54, 319 columns 178-179
in EOM distribution 55 plasticity 232, 263-264, 267
muscular action 66, 72 stripes 222
directions 68-73 ocular mechanics 65, 68
plane 68 ocular rotation 65-70
myopia 198 ocular tilt reaction 106
myotomy 328-329 ocular translation 69, 71
marginal 328 oculomotor systems
adaptive control 140
neural plasticity 18
Na+-channels 265 physiology of 5
nasal-temporal asymmetry 99 OKN, see optokinetic nystagmus
nasal to temporal insufficiency OKR, see optokinetic reflexes
81 ontogeny 89
natural selection 78, 86 ophthalmic branch (OB) 319
natural strabismus 49 ophthalmoplegia, internuclear
neck muscles 58 14
nerve palsies, congenital sixth optic chiasm 171
cranial 29 optic nerve 177
neural integrator 113 optokinetic circuitry 83
neuron circuits 78 optokinetic nystagmus 15, 99
neuronal mechanisms 78, 84-85, asymmetrical 138-139
247, 252 by visual imagery 105
neuronal numerical density 175 optokinetic reflexes
neurons 173 in amblyopia 89-96
loss of 177 development of 90
see also specific types in strabismus 89-96
neuropil 179 optokinetic responses
neurotransmitters 281 anomalous uniocular 24
NMDA-receptor 265 induced 365
non-competitive deprivation 240 optomotor reflexes 12, 58
non-deprived eye 199 orienting behaviour 320
norepinephrine (NE) 264 outflow 322, 337-338
NOT, see nucleus of optic tract
nuclear bagl and bag 2 fibres 54
nuclear cha1n fibres 54 palisade endings 55-56, 327-329
nucleus of optic tract 89 Panum vision 281
nystagmic block 124 paralytic strabismus 9-10, 53,
nystagmus, gaze-paretic 170
component 99 parvocellular layers 170
see also specific types parvocellular neurons 176
nystagmus blockage syndrome PAT, see prism adaptation test
27-28 perimetry 279
phasic mechanisms 112-113
photopic increment threshold
occlusion 169, 279-280 236
monocular 260-262 photoreceptors 229
reverse 260, 262 maturation of 174
ocular dominance 220-227, phylogeny 81, 89
230-231 general 78
436 INDEX
physiologic diplopia 281 retino-pretectal connections 90

plasticity 202, 221, 224, 229, retrograde atrophy 275
232, 235-244 reverse deprivation 224, 229
of amblyopia 230 reverse suturing 224-226, 230
neural 18 rodents 178
pointing errors 338-339 rods 174
polyneuronal innervation 53
positional acuity 369, 372
positional uncertainty 377 saccades
postsynaptic activation 262-263 conjugate 138
pre-binocular vision 247 convergent 79
premotor neurons 117 hypometric 138
prepositus nucleus 85 non-conjugate 138
primary and secondary endings saccadic gain control 18
55 sagittalization 27
primary visual pathways sampling 375-377
normal anatomical development scotoma 278
of 173-179 scotopic spectral sensitivity
of primates 173-179 236
prism adaptation test 117-118, see-saw nystagmus 14
126 segregation 222, 224, 226, 232
proprioception 319-323, of terminals 222
327-334, 337-341 semilunar ganglion
oculomotor 319, 327-328, 333 in cats 319
proprioceptive receptors 319 in monkeys 319
position sensitivity 56 sensitive period 187, 189-190,
pseudo-Panum's area 127, 312 200, 219-232
psychophysical investigations sensitivity 238
235-236, 370, 377-378 see also specific types
pursuit 99 single fibre EMG 53
asymmetry 100-102, 139 singly innervated fibres
smooth 138-139 (SI units) 48
pyramidal cells, layer III 178 size principle 51
skew deviation 14
Snellen acuity 370-373, 376
random dot E 360 spatial contrast sensitivity
random dot stereograms 360, 427 236
Randot test 360, 363 spatial distortion 388
receptive fields 227-229 spatial localization 327,
properties of 95 333-334
recession 328-329 spatial modulation 237
reinnervation 52 spatial resolution 227-230, 232
resolution limits 238 development of 226-232
retina 91, 173, 275, 281 spatial sampling 376
development of 134 spatial transformation 81
specialisation of 79 spatial vision 134
retinal areas, anomalous spectacles 169
corresponding 127 spine loss 178
retinal rivalry 139, 280, spiral endings 319
312-313 spontaneous nystagmus 15
retinal slip 90 stereoacuity 137, 247, 251-252,
retinogeniculate synapses 177 321, 359, 363, 427
INDEX 437
stereopsis 82, 134, 139, 247, tonic adaptable components 83

251-254, 312, 314, 359-366 tonic mechanisms 112-113
global 427-429 tonus 124-125
hysteresis effect 427 trigeminal complex 56
physiological basis 345-354
strabismic amblyopia 268, 275,
277, 280, 387-389 velocity perception 102, 139
strabismus 169, 231, 276-277, velocity sensitivity 56
281, 359-366 vergence 83-84, 112
abnormal proprioceptive accommodative 79, 118, 134,
reflexes 58 136' 140
clinical manifestations of adaptation to disparity 112
37-43 clinical aspects of 121-129
extraocular muscles 37-43 disparity 78-79, 118
forms 6 fusional 134, 137, 140
heredity 7 phase-tonic organization
mechanical factors 12-13 111-119
motor dysfunction 5-19 tonic 118
natural course 19
optokinetic reflex 89-96
specific types 7 vergence accommodation 140
stretch reflexes 57 vergence adaptation 118
visual perception 311-316 vergent interocular angles 79
see also specific types vernier acuity 247, 250, 252,
strabismus surgery 327' 331' 370-373, 376
334, 339-340 versional motor system 81
strabismus treatment, prediction vertebrate evolution 81
363 vestibular geometry 81
stria of Gennari 177 vestibular nuclei 85
striate cortex 227-229, 281 vestibula-ocular reflex gain 18
of cats 219-220 visual acuity 179, 221,
of monkeys 226-227, 230-232 226-227, 382
of primates 221-226 development of 228
striated glasses 127 in humans 227
stripes 221 in monkeys 227
structural homology 78 visual axes, misalignment of 78
subcortical vergence machinery visual cortex 92, 171, 173,
82 207, 281, 319
suppression 276-278, 280, of cats 221
311-316 of humans 178-179
synapses 275 of monkeys 178-179, 219-232
formation of 173, 224 of primates 226
synaptic contacts 174 visual deprivation 92-93, 176,
synaptic density 178 179, 207-215, 275-276, 279
synaptic organization 178 visual development
deviations in 247-254
normal 247-254
temporal frequency 113 visual geometry 81
temporal vision 134 visual nervous system,
tendon organs 54 developing 369-378
Titmus test 360 visual system of cat
TNO test 360, 363 proprioception 319-323
438 INDEX
visual system plasticity
development periods 235-244
psychophysical studies
235-244
visuo-motor co-ordination 264
voluntary nystagmus 104

Strabismus and Amblyopia

Uploaded by

Copyright:

Available Formats

Strabismus and Amblyopia

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Strabismus and Amblyopia

Uploaded by

Copyright:

Available Formats

What is the title of the book?

What is the title of the book?

What are the main topics covered in the book?

What are the main topics covered in the book?

WENNER-GREN CENTER

INTERNATIONAL SYMPOSIUM SERIES

STRABISMUS AND AMBLYOPIA

Proceedings of an International Symposium held at

Gunter K. von Noorden

All rights reserved. No reproduction, copy or

No paragraph of this publication may be

Any person who does any unauthorised act in

First published 1988

Published in Great Britain by

British Library Cataloguing in Publication Data

1. Motor Dysfunction in Strabismus G. Lennerstrand 5

3. Morphology of the Extraocular Muscles in Relation to the

4. Motor and Sensory Functions of Normal and Strabismic

6. Phyletic Organization of Brainstem Neuronal Circuits and

7. Neural Basis for Changes of the Optokinetic Reflex in

8. Ocular Motor Phenomena in Infantile Strabismus

9. Phasic-Tonic Organization of Accommodation and Vergence

10. Clinical Aspects ofVergent Mechanisms B. Bagolini 121

11. Normative Oculomotor Development in Human Infants

12. DISCUSSION: Oculomotor Control and Strabismus

14. Normal Anatomical Development of the Primate Primary

15. Effects of Abnormal Visual Experience on the Morphology of

16. The Influence of the Period of Deprivation on Experimental

17. Electrophysiology of Cortical Neurons under different

18. The Sensitive Periods of the Monkey's Visual Cortex

19. Psychophysical Studies of Visual System Plasticity During

20. Normal Visual Development and its Deviations

21. Neuronal Mechanisms of Deprivation Amblyopia

22. Abnormal Binocular Interaction:Evidence in Humans

23. DISCUSSION: Normal and Abnormal Visual Development

25. Role of Oculomotor Proprioception in the Visual System of

26. Extraocular Muscle Proprioception and Visual Function:

27. Extraocular Muscle Proprioception and Visual Functions:

28. On the Physiological Basis for Stereopsis

29. Stereopsis and Strabismus E.M. Helveston 359

30. Psychophysical Consequences of Image Degradation and

31. Amblyopia: Clinical Aspects W. Haase 381

32. DISCUSSION: Psychophysics Related to Strabismus and

33. Concluding Remarks G. Westheimer 413

34. FINAL DISCUSSION Moderator: G. Westheimer 417

35. Global Stereopsis B. fulesz 427

Parallel to this upsurge of experimental work, clinical research on

Stockholm, June 1987

Richard Aslin Emilio Campos

Colin Blakemore Juan Gallo

Laurence Garey Eugene Helveston

Brian Harcourt Gerd Holmstrom

Gunnar Lennerstrand Elisabeth Schultz

Bjorn Tengroth Olof Wennhall

Per Udden Gerald Westheimer

It is a great pleasure and privilege for me to greet you all heartly

At the Stockholm meeting in 1974 it was mentioned that the meeting

After having looked through the program I am convinced that we are

The basic scientist forms his hypothesis and starts_prpjects in fields

lems. An interesting result wrongly interpreted and applied in the