04 Neurocortex Merzenich Et Al PDF
04 Neurocortex Merzenich Et Al PDF
04 Neurocortex Merzenich Et Al PDF
AbsTracT. Recent experiments have shown that the effectiveness of the coupled
excitatory inputs and intrinsic inhibitory processes controlling the overt
excitatory responses of cortical neurons a re dynamically alterable throughout
life. Many neoco rtical functional plas ticity studies demonstra ting this alterability
have been conducted in the somatosensory cortex. In this review , we briefly
describe some aspects of th e results of these studies. Features of the underlying
processes of excitatory "input restriction" or " input selection" reveal ed by
them are described. Some implications of these adaptive cortical processes for
the origin s of behavior a re discussed.
41
42 M.M. Merzenich et al.
BEFORE DIFFERENTIAL
STIMULATION
AFTER DIFFERENTIAL
STIMULATION
1mm
Fig. !-Outl in es of the area 3b cortica l territories representing the surfaces of the
digits at two epochs in the life of a normal adu lt monkey. Maps were derived prior
to initiation of a special behavior and about 3 months later, during which time this
monkey had performed a task that involved differential, heavy stimulation of the
tips of the distal phalanges of digits 2, 3. and occasionally 4 for about one hour/
day. The differentially stimulated skin surface on digit 2 is indicated by stippling in
the hand drawing at the right. Representational "maps'' were drawn from definition
of receptive fields defined within several hundred microelectrode penetrations into
this area 3b zo ne at both experimenta l stages. Outlined territories are zones over
which receptive fields were centered on digits I to 5. The zones of representation
of the struck skin surface on the distal phalanx of digit 2 is stippled in both drawings,
i.e., in this zone , all defined receptive fields overlapped onto this heavily stimulated
surface. Note that after a period of differential digit tip stimu lation in this and in
five ot her monkeys studied with this paradigm , there was a substantial enlargement
of its territory of representation. In this monkey, most of the gain in territory was
from across the area 3b- 3a border as defined functionally prior to the initiation of
this digital stimulation behavior. On average , about half of accrued territory of
representation in these monkeys is gained internally from within area 3b and about
ha lf from across the functionally defined rostral border of the koniocortical cutaneous
hand surface representation . Redrawn from Jenkins and Merzenich ( 1987).
Cortical Representational Plasticity 45
Fig. 2-Some features of th e ha nd rep rese nta ti o n (center, top) in a n adult owl
mo nk ey in which di gits 3 a nd 4 were fused by a syndact yly surgica l procedure fo r
7 mo nth s prior to th e deriva ti o n of a hi ghly detail ed cortical map of this a re a 3b
zo ne . In th a t map, all receptive fi elds defined within a la rge re gion (marked by
diagonal stripin g) between th e te rrit o ri es of the re prese ntatio ns of th e digits had
recepti ve field components on both of th e m . Two such fields defin ed at the sites
indicated by the black dots within this zo ne are illustrated o n the syndactyl di gits
in th e ha nd drawing at th e left. By contrast, rece ptive fi e lds defin ed for ne urons
ac ross norma l sha rp digital rep rese ntati o nal "discontinuities'' responded exclusive ly
to stimul ati o n of o ne or th e o ther of the two adjace nt digits as illustrated , fo r
exa mple by the two typica l nearby pe net ra tions across th e di git 2-digit 3 bo rd e r,
also sho wn in th e ha nd drawing at th e left.
A t th e e nd of th ese experiments , the synd actyl digits were surgicall y se pa rated,
a nd receptive fi e lds again defined within th e zo ne of do ubl e-di git receptive fie lds.
Two such fields , typ ical of nearly all acute ly reco rd ed within this zone (recorded a t
th e locati o ns ma rk ed by stars in the map) are sho wn in the hand drawing at th e
ri ght. Both had rece ptive field compo ne nts o n th e two adj ace nt digits, a respo nse
feature enco unt e red infreque ntl y in no rm al ma ps in a rea 3b in adult owl mo nk eys.
46 M.M. Merzenich et al.
en~
10
< ·.!:"'
w
a:: as 8
<!
U..CII
a::U.. 6
a:
o-
w
N
.
.§ 4
:i.,
ctGI
:=:a
a::u..
2
Oa::
z~
0
0 100 200 300 400 500 (min)
TIME
Fig. 4--Receptive field sizes of neurons defined at eight cortical loci within the
forelimb and hindlimb representational zones in area 3b in an adult cat , plotted as
a function of the duration of 20 pps electrical stimulation of the ulnar nerve in the
distal foreleg. On average, roughl y five-fold increases in the extents of receptive
fie lds are induced throughout the contralateral body surface representation by such
stimulation . Note that in the hindlimb representational zone , there is no directly
evoked neural response to electrical stimulation of this forelimb nerve . A series of
controls demonstrate that such receptive field size changes are a direct consequence
of electrical stimulation of the peripheral nerve (Recanzone et al. 1986, 1987).
Cortical Representational Plasticity 49
Such studies indicate that the anatomical map of the skin surface is far
cruder than the functional map of overtly effective excitatory inputs defined
by the use of simple, natural stimuli at any given moment in the life of the
adult primate or carnivore.
5) Inputs from far broader sources than those that overtly drive cortical
neurons commonly evoke excitatory postsynaptic potentials within them
(Zarzecki et al. 1983; Zarzecki and Wiggin 1982) . Such functionally
demonstrable excitatory inputs are recordable by stimulation of different
nearby peripheral nerves on distance scales indicating a spread of excitatory
inputs representing given peripheral skin locations in the somatosensory
cortical fields of carnivores of up to at least several millimeters.
6) Individually labelled ventroposterior cell afferent arbors spread
relatively widely within the middle cortical layers in even the koniocortical
field , area 3b, of monkeys and carnivores (Garraghty and Sur 1987; Landry
and Deschenes 1981). Physiological mapping of the arbors of single
ventroposterior thalamic units , representing specific skin loci referenced to
the cortical zone of overt representation of that skin locus , reveal a mismatch
(Snow et al., submitted; e.g., Fig. 5). Both electrophysiological arbor-
mapping experiments and single fiber labelling studies in New World
monkeys indicate that individual arbors may spread up to more than 2
mm across the horizontal dimension in the koniocortical somatosensory
representation , area 3b.
Electrophysiological mapping studies also suggest that arbors from across
the isorepresentational axis of the ventroposterior thalamus (see Jones et
al. 1982; Kaas et al. 1984) project with substantial divergence into cortical
fields (Snow et al., submitted). This dispersion of anatomical inputs
representing common peripheral skin locations - along with the relatively
broad distribution of individual afferent fibers - underlies the generation
of the relatively crude ("degenerate"; see Edelman 1978) "anatomical maps"
in this system.
The limits of anatomical divergences of projections presumably account
for observed distance limits for representational reorganization (Merzenich
1987; Merzenich , Jenkins et al. 1984; Merzenich , Nelson et al. 1984). That
is, the site of representation of a given skin surface can shift within the
neocortex to only a limited extent (see Fig. 6). Representational changes
up to about 1.5-2 mm have been recorded in cortical area 3b, but shifts
are usually not greater than about a thousand microns. Greater changes
have been recorded in the more divergently connected cortical area
1 (unpublished observations), and substantially greater representational
translocations have been recorded in the principal Sll field in macaques
(Pons 1987).
50 M.M. Merzenich et al.
-
O .lmm
Fig. 5-Surface view of the territory of distribution of the cortical arbor of a single
ventroposterior nucleus neuron representing a restricted cutaneous receptive field
on the surface of digit 3 in an adult cat (dotted). The arbor boundaries were defined
to within a few tens of microns by low-amplitude antidromic microstimulation of
this neuron in multiple cortical penetrations (dotted). The territory over which
cortical multiunit receptive fields overlapped with this single thalamic neuron 's
receptive field is indicated by the cross-hatching. This order of mismatch between
the distributions of arbors and the cortical zone of representation of the skin surface
represented by studied thalamic neurons (in this case, the arbor extended into the
zones of representation of both adjacent digits) was common. Redrawn from Snow
et al. (1988).
POST NORMAL
STIMULATION
'
F H
A A
c N
E 0
1mm
POST 3a
STIMULATION
1mm
3b
..... ,
,
Fig . 6--Differences in the locations of the hand-face (upper) and area 3b-area 3a
borders (lower) defined in an ad ult owl monkey before and after a period of intense
behavioral training (see legend , Fig. 1). The stippled zones mark the minimal
territories over which receptive fields for neurons at those locations were determined
to have moved from the face to the hand, or from deep characteristically-area-3a
responses to cutaneous characteristically-area-3b responses. In this zone, there is a
direct overlap of microelectrode penetrations in wh ich receptive fields were either
(upper) on the face in the first map , and on the hand in the second; or (lower)
deep responses predominantly from muscle and joint afferents in the first map, and
low-threshold cutaneous responses in the second. Significant representational
translocations are recorded along these borders in nearly every animal from which
more than one representational map has been derived . Translocations never exceeded
a distance limit of about one mm . See text for further details and some implications .
52 M.M. Merzenich et al.
addresses should result in receptive fields covering most or all of the skin field
of the transected nerve) must involve a temporally based input selection
process. The reestablishment of map topography almost necessarily requires
operation of a coincident-based selection process. By this interpretation , details
of map topography reflect the probabilities of coincident or nearly coincident
inputs. Adjacent skin surfaces are represented in adjacent locations in large
part because there is a high probability that they will be excited nearly
coincidentally.
4) Studies of chronically coincidentally excited vibrissae conducted in the
barrel fields of adult rats are also consistent with these conclusions. If two
vibrissae are .coupled so that they are with high probability excited coincidentally ,
they are highly likely to costimulate cortical neurons over most of the cortical
zone in which they are represented (Yun et al. 1987). Such studies have been
conducted by an experimentalist who is blind to the vibrissae coupled over a
1 to 2 week period prior to derivation of the electrophysiological map. Before
the map is derived, all vibrissae are trimmed to a constant length , and the re
is no indication of which vibrissae have been coupled. Rats are maintained in
a rich tactile environment. Again, these studies demonstrate an apparent
powerful influence of the temporal correlations of inputs for shaping selected
cortical receptive fields.
A NAL Injection•
c(
..
! ~~
~-
w
a:
<_
u..
2
~
a: u..'"·"' • ._...,.:..:::::::-:;;;::: . ................
ca:::
w-
iii i iii
0
~ .§
-JU:
<a:
:E._. -2 NAL Injection•
a:
0
z
-4
-100 0 100 200 (min)
TIME
8
c(
w
a: NAL Injection
<_
!
6
u..
a: u..'"·"'
ca::: 4
w-
~ .§
-JU:
<a:
:E._. 2
a:
0
z
0
-100 0 100 200 (min)
TIME
Fig. 7-Receptive field sizes change with continuous peripheral nerve stimulation
and naloxone injections . Receptive fie ld sizes throughout the contralateral body
surface representation in area 3b of adult cats invariably en large over a seve ral- hour
period of peripheral nerve stimulation . If naloxone is introduced (at 0 on the
abscissa) whi le a forelimb peripheral nerve is being continuously stimul ated, initially
enlarged receptive fields usually shrink to near their prestimulation sizes. If naloxone
injections are repeated at half-hour intervals to maintain a high blood level , small
receptive fie ld sizes can be sustained (A). In a period well afte r a single naloxone
injection (B) , or after cessation of repeated injections (upper curve in A), receptive
fie lds again enlarge, presumab ly as naloxone binding again declines.
Cortical Representational Plasticity 59
Again , major effects are recorded over the time period in which the maximal
changes in cortical representational detail would be expected. Spatial tactile
acuity commonly improves significantly on the stump skin of amputees (Haber
1955 ; Teuber et al. 1949) . Again , the time courses and extents of such changes
in humans parallel those for observed reorganizational changes in cortical
somatosensory representation following digital amputation in monkeys. Tactile
acuity has been shown to be rapidly degraded following acupuncture (Wang
et al. 1981) ; again , we have observed substantial increases in receptive field
sizes occurring rapidly after electroacupuncture-like stimulation of a peripheral
cutaneous nerve . Changes in receptive field sizes in our animal experiments
are reversed by administration of naloxone. Two-point thresholds can be
reduced significantly by the intravenous administration of naloxone (Velasco
et al. 1984).
In all of these simple examples , observed changes in tactile acuity in humans
roughly parallel the time courses and extents of changes in the magnitudes
and grains of area 3b representations of monkeys undergoing similar
"experimental" histories.
Parallels to perceptual changes of body image can also be drawn. Thus, for
example , telescoping following amputation (see Haber 1955; Merzenich, Nelson
et al. 1984; Teuber et al. 1949) indicates that reorganization of a body image
map(s) occurs in these patients. Correction in the body surface image for finger
identity following an island pedicle transfer constitutes another example of an
alteration in body image representation that parallels reorganization changes
recorded in our experiments. Despite this parallelism , we believe it to be most
likely that such effects are accounted for by reorganization in somatosensory
cortical areas not yet studied with this class of experiment, in part because of
the magnitudes of implied representational changes indicated by these recovery
phenomena .
In general , it is highly likely that area 3b representational changes account
for its fractional part of the acquisition of tactual skill with practice-that the
forebrain mechanisms underlying the adaptive changes revealed in these
experiments minimally constitute the bases of at least a major neocortical
component of learning and memory.
. : :. v
:::::tff\l\l\i~t\ttiC
·······:·:···:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·:·.
•'•
X X HL •,•
-
O.Smm
REFERENCES
Allard , T.T. ; Clark , S.A.; Jenkins, W.M. ; and Merzenich, M.M. 1985. Syndactyly
results in the emergence of double-digit receptive fields in somatosensory cortex
in adult owl monkeys. Soc. Neurosci. Abst. II : 965.
Arnsten, A . F.T.; Neville, H.J. ; Hillyard , S.A.; Janowsky , D.S .; and Segal , D.S .
1984. Naloxone increases electrophysiological measures of se lective information
processing in humans . J. Neurosci. 4: 2912-2919.
Basbaum , A.l. , and Fields, H .L. 1984. Endogenous pain control systems: brainstem
spinal pathways and endorphin circuitry. Ann. Rev. Neurosci. 7: 309-338.
Baskin, D.S. ; Kieck , C.F.; and Hosobuchi, Y. 1984. Neurological reversal of
ischemic neurological deficits in man. Brain Res. 290: 289- 296.
Bear , M.F. , and Singer , W . 1986. Modulation of visual cortical plasticity by
ace tylcholin e and noradrenaline. Nature 320: 172- 176.
Chung , J.M .; Lee, K.H. ; Hori , Y. ; Endo, K.; and Willis, W.D . 1984. Factors
influencing peripheral nerve stimulation produced inhibition of primate spinothal-
amic tract cells. Pain 19: 277- 293.
Clark , S.A. ; Allard , T.T .; Jenkins , W.M. ; and Merzenich, M.M. 1986. Cortical map
reorganization following neurovascular island skin transfers on the hands of adult
owl monkeys. Soc. Neurosci. Abst. 12: 391.
Diamond , D. 1985. Physiological Plasticity of Single Neurons in Auditory Cortex
of the Cat During Learning. Irvine: Univ. of California (Ph.D. thesis) .
Dykes , R.W.; Landry , P.; Metherate, R.; and Hicks, T.P. 1984. Functional role of
GABA in cat primary somatosensory cortex: shaping receptive f1elds of cortical
neurons . J. Neurophysiol. 52: 1066--1093.
Edelman , G.M. 1978. The Mindful Brain. Cambridge: MIT Press.
Edelman , G.M., and Finkel , L.H. 1985. Neuronal gro up selection in the cerebral
cortex. In: Dynamic Aspects of Neorcortical Function , eds. E. Gall and G.
Edelman , pp. 653-695. New York: Wiley .
Finkel, L.H. ; Pearson , J .C.; and Ede lman, G.M. 1987. Plasticity of map organization
modelled using population rules for synaptic modification. Soc. Neurosci. Abst.
12: 391.
Franck , J . L. 1980. Function al reorganization of cat somatic sensory-motor cortex
(Sml) after selective dorsal root rhizotomies. Brain Res. 186: 458--462.
Gallagher , M. ; Rapp , P.R .; and Fanelli, R.J. 1985. Opiate antagonist facilitation
of time dependent memory processes: dependence upon intact norepinephrine
function. Brain Res. 347: 284-290.
Gamble , G.D ., and Milne , R.J . 1986. Analgesic effects of acupucture-like sciatic
stimulation in the non-restrained , environmentally experienced rat. Soc. Neurosci.
Abst. 16: 225 .
Garraghty , P .E. , and Sur, M . 1987. The morphology of single physiologically
identified thalamocortical axons innervating somatosensory cortex in macaque
monkeys. Soc. Neurosci . Abst. 13, in press.
64 M.M. Merzenich et al.