SECTION IX

Eye–head–neck coordination
Progress in Brain Research, Vol. 143
ISSN 0079-6123
Copyright ß 2004 Published by Elsevier BV.

CHAPTER 35

Current approaches and future directions to

understanding control of head movement

Barry W. Peterson*

Department of Physiology, The Feinberg Medical School, Northwestern University, Chicago, IL 60611, USA

Abstract: This chapter reviews four key issues that must be addressed to advance our knowledge of control of head
movement by the central nervous system (CNS). (1) Researchers must consider how the CNS utilizes the multiple
muscle patterns that can produce the same head movement in carrying out tasks in an optimal way. (2) More attention
must be paid to the dynamics of neck muscle activation that are required to implement head movements and show they
are produced by CNS circuits. (3) Research is required to determine how the multiple pathways that impinge upon neck
motor centers are utilized in a variety of tasks including eye–head gaze shifts, smooth head tracking, head stabilization
and manipulating objects with the head. These pathways include corticospinal, vestibulospinal, reticulospinal (three
subdivisions), fastigiospinal, tectospinal and interstitiospinal tracts. (4) Further analysis is needed to understand how
vestibular signals are modulated during each of the above-mentioned tasks. This ambitious agenda is justified by the
fact that the head–neck motor system is an ideal model for understanding issues of complex motor control.

Introduction weaker than in the arm. The head also participates in

gaze shifts that align the optic axes with an object to
This chapter provides a review of the current state be viewed. Their stereotypic nature makes such gaze
of work on head-movement control. What follows is shifts easier to understand than arm movements
a conceptual look at this system, as illustrated by whose goals and trajectories are exceedingly diverse.
diagrams rather than data from the author’s and The presence of both eye- and arm-related control
others’ laboratories. systems is reflected in the wide variety of pathways
An interesting property of the head-movement impinging on regions of the cervical spinal cord that
control system is that it shares features of both the control the neck musculature. These include vestibu-
eye and arm-movement control systems. Like the eye, lospinal, reticulospinal, corticospinal, tectospinal and
the head is used to control the direction of gaze. It fastigiospinal pathways. As described below, motor
thus requires mechanisms to keep its orientation commands traversing these pathways are selectively
stable in the face of external perturbations. As in the deployed to accomplish various head-movement
eye, evolution has solved this problem by provid- tasks.
ing strong, short-latency vestibular reflexes, the This review examines four aspects of head motor
vestibulocollic reflexes (VCRs). Muscle-spindle- control that shed light on the special properties of
driven reflexes, which serve to adjust compliance of this neuromuscular system. The first two are the
the arm, are also present in the neck (cervococollic kinematics and dynamics of the head–neck neuro-
reflexes) but their monosynaptic portion is relatively muscular system. The author then examines supra-
nuclear commands for voluntary head movements
*Corresponding author: Tel.: þ 312-503-6216; carried by the various descending pathways that
Fax: þ 312-503-5101; E-mail: [email protected] impinge upon neck motor nuclei. The fourth section

369
DOI: 10.1016/S0079-6123(03)43035-5
370

examines neural control strategies for generating eye–

head gaze shifts. A final section will discuss future
directions for research on head-movement control.

Kinematics of the overcomplete head–neck

musculoskeletal system

Various lines of evidence suggest that the head–neck

system is overcomplete in that it has more than the
minimum number of muscles required to control the
motions of its seven joints (Pellionisz, 1988). This
means that there will not be a single unique pattern of
muscle activation associated with every head move-
ment. Rather, a particular movement can be executed
by many different muscle-activation patterns.
Figure 1 illustrates the contrast between kine-
matics of the oculomotor and head-movement Fig. 1. Kinematic transformation matrices for eye and neck
motor plants. In A and C, rows indicate muscle inputs, and
systems. In the oculomotor system, three pairs of
columns show movement outputs. Values in A, which indicate
muscles control the three rotational degrees of the movement produced by each muscle pair, are from
freedom of the eye (yaw, pitch and roll). This Robinson (1982). B is the inverse of A. Here, rows indicate
corresponds to a situation with three unknowns and movements, and columns show muscles must be activated to
three equations. As a result, one can calculate a single produce them. Movements produced by each muscle in the neck
are still in doubt and thus the matrix elements in C have been
unique set of activation strengths of each muscle pair
left blank. Muscle abbreviations: LRMR, lateral and medial
that are required to produce each motion (Robinson, rectus; l and r SCM, left and right sternocleidomastoid; l and r
1982). Spi, left and right splenius; l and r Sem, left and right
Matrix A in Fig. 1 represents the eye movements semispinalis; SOIO, superior and inferior oblique; SRIR,
that each muscle pair will generate. For instance, the superior and inferior rectus. Other abbreviations: P, vertical
or pitch eye movement; P1, P2 head pitch about joint 1 and 2,
middle row indicates that activation of the superior/
respectively; R, torsional or roll eye movement; R2, Y2, roll and
inferior rectus pair (superior rectus activation pitch of head about joint 2.
coupled with suppression of inferior rectus) will
produce negligible horizontal rotation, strong
upward rotation (
0.91) and significant torsional
rotation (0.42). Because the matrix is invertible, one considered as acting separately, resulting in six
can also compute the pattern of muscle activation unknown activation strengths. Two types of joint
needed to produce a particular motion as represented are included. One can move only in the pitch
in the inverse matrix, B. For instance, the middle row direction, as does the joint between C1 and the skull.
indicates that a downward pitch motion requires The other can move in all three directions as in the
activation of inferior rectus and inactivation of case of lower cervical joints. This results in four
superior rectus with a net strength of
0.81 and also equations—too few to completely specify the activa-
activation of the oblique muscle pair with a strength tion of muscles required to produce a particular
of 0.43 to counter the roll motion that the vertical motion. Rather, there are many different solutions
recti would otherwise produce. There is also a very that can all produce the same resultant motion. We
weak inactivation of LRMR. can fill in the matrix elements specifying the motions
The matrix in Fig. 1C portrays a simplified head/ each muscle would produce but the 4  6 matrix is
neck system with three muscles on both the right and not invertible. Thus, we cannot specify a pattern of
left sides and two joints. Because the muscles do not muscle activation that is necessary to produce a given
align into push/pull pairs as in the eye, each must be motion.
 371

Experimental examination of muscle-activation Correspondingly, extraocular motoneuron (MN)

patterns associated with head movements has activity has a burst component related to desired
provided examples where multiple activity patterns eye velocity and a tonic component related to eye
were associated with a particular movement and position. Mechanics of the head/neck system are
others where a single pattern was observed. A single dominated by the inertia of the head and secondarily
pattern across time and individual subjects was by viscoelastic properties of the muscles. Correspond-
observed in highly constrained tasks such as ingly, neck muscle activity should be triphasic with
stabilization of the head by the VCR during body pulses of activity in agonist and antagonist muscles to
rotation (Keshner et al., 1992; Banovetz et al., 1995; accelerate and brake the head at movement onset and
Uchino, this volume) and generation of isometric termination with possible tonic activity related to
forces with the head (Vasavada et al., 2002). In head deviation. Such signals were in fact reported in
contrast, multiple patterns were observed during early studies (Zangemeister et al., 1982). The
tracking of a target with the head (Keshner and dynamics of MN and premotoneuron activity,
Peterson, 1988; Choi et al., 2000) and when subjects however, has received relatively little attention in
had to stabilize the head against loads applied with a recent studies of eye–head gaze shifts, which have
weight-pulley system (Keshner et al., 1989). This emphasized instead the use of common error signals
suggests that the redundancy inherent in the over- to drive both the eye and head. This leaves
complete head/neck musculoskeletal system allows unanswered questions regarding where in the final
the central nervous system (CNS) to choose different output pathway the appropriate dynamical signals
patterns of muscle activation to execute the same are generated?
movement in ways that optimize some aspect of the In contrast, studies of head stabilization have paid
movement (see also Sugiuchi et al., this volume). a great deal of attention to head/neck dynamics.
Obvious distinctions are between rapid, predictive Early recordings in the decerebrate cat revealed the
and fine, visually driven tracking and between force expected second-order behavior with head position
generation with low and high stiffness. The observed signals at low frequencies and head acceleration
behavior suggests that different central neural path- signals at high frequencies (Bilotto et al., 1982).
ways are used to drive movements that have such Angular accelerations of the head activate
differing properties. This is an important question to regularly and irregularly firing afferents from the
be addressed in studies recording central neural vestibular semicircular canals. It is the latter
activity associated with head movement. Can we that preferentially drive vestibulocollic pathways
indeed show that different premotor pathways are (Highstein et al., 1987). Their dynamics include a
active when a movement is made utilizing different pole term that converts angular acceleration to
muscle synergies? Also, can we define the way in angular velocity at lower frequencies and a zero term
which each pattern is optimal and relate it to the task that converts back to acceleration at higher fre-
being performed when the pattern appears? Answers quencies above 1 Hz (Fernandez and Goldberg,
to these questions will greatly advance our under- 1971). These terms form the first part of the VCR
standing of motor systems. transfer function in Fig. 2.
Wilson et al. (1979) used electrical polarization to
bypass this first afferent-related term allowing them
Dynamics of head–neck system and to observe the further processing of VCR signals in
its neural controllers the CNS. As shown in the transfer function, this
consists of a second pole-zero pair. The second pole
Dynamics refers to the way in which amplitude and converts the velocity signal produced by the first to
timing (phase) of neural or movement signals vary an angular position signal. Again above 1 Hz, the
with frequency of the head movement or of second zero reverses this effect, yielding an
the stimulus that is eliciting it. Mechanics of the angular acceleration signal at high frequencies.
eye are dominated by viscosity of orbital tissues Thus, semicircular canal inputs are processed
and spring-like forces of the extraocular muscles. exactly as required to produce a second-order
372

Fig. 2. Dynamics of horizontal head stabilization about a single joint. The symbol s is the LaPlace variable. I, B and K are inertia,
viscosity and stiffness, respectively, of neck. Kc and Kv are constants weighting the action of the CCR and VCR, respectively, on neck
musculature. See text for discussion of time constants (T).

position/acceleration control signal on neck MNs. Supranuclear commands for voluntary head
Peterson et al. (1985) also observed the dynamics of movements
responses to rotation of the body about an earth-
fixed head, finding again a second-order signal. This Commands for generating head movements reach
behavior, which is represented by the cervicocollic neck MNs from a variety of sources that are shown in
reflex (CCR) transfer function in Fig. 2, likely reflects Fig. 3.
the properties of neck muscle spindles. Several classes of vestibulospinal neurons project
Measurement of the transfer functions of system to neck motor pools. One group includes the
components opened the way for dynamical modeling bifurcating position-vestibular-pause (PVP) neurons
of the head stabilization system in cats (Goldberg and that project to both extraocular and neck MNs (Isu
Peterson, 1986) and humans (Peng et al., 1996a,b). et al., 1988). Their name denotes the fact that they
These models have proven remarkably effective carry an eye position signal and a vestibular
in replicating and explaining the complex dynamic semicircular canal signal and pause during saccadic
behavior of the head during horizontal rotations gaze shifts. Perlmutter et al. (1998a) established that
of the body (Keshner and Peterson, 1995; Keshner the canal input to these neurons is from a single
et al., 1995). Several conclusions can be reached canal. Responses of identified bifurcating neurons to
from these models: (1) the VCR is not very effective linear head motions have not been studied but it is
in stabilizing horizontal head position in space likely that they also receive input from otolith
at frequencies below 0.5 Hz (gain  0.1); (2) the afferents sensitive to linear acceleration as observed
VCR and CCR are vital for preventing head in unidentified PVP neurons (Angelaki et al., 2001).
oscillations in the 1–3 Hz range; and (3) a static The likely signal carried by these neurons is a
vertical VCR appears to be important for stabilizing command for stabilizing eye and head movements
the head in an upright posture against the force of aligned with the coordinate frame of the three
gravity. It is likely that the VCR also plays a role in semicircular canals. This would then be distributed
stabilizing the head during linear motions of the body with the proper weights to extraocular and neck MNs
(Lacour and Borel, 1989; Borel and Lacour, 1992; to generate the required motions.
Borel et al., 1994). This is a topic for future There are also vestibulospinal neurons that project
investigation. only to the neck. These typically receive convergent
 373

Fig. 3. Scheme showing processes involved in generating and controlling eye–head gaze shifts. Commands for saccadic gaze shifts
originate in association areas of frontal and parietal cortex (F-PCx) and activate output systems originating in motor cortex (MCx) and
SC. SC signals reach neck motoneurons (NMn) primarily via reticulospinal neurons (RSN) and oculomotor nuclei (OM) via excitatory
and inhibitory burst neurons (EBN, IBN). Both head and eyes are stabilized in space via pathways originating in semicircular canals
(SCC) and relaying in the vestibular nuclei (VN). The VN are also the relay for smooth pursuit commands relayed via area MST and
cerebellar vermis and flocculus (Cb). The cerebellum also regulates eye and head movements via its fastigial nucleus (FN). Also
included are three mechanisms that regulate gaze shifts: the eye saccade gate implemented by omnipause neurons (OPN); the resettable
neural integrator (RNI) that turns off saccades when the correct goal is reached; and, a head-movement estimator (HME) that cancels
the portion of VN activation due to commanded head movements. HMS are head-movement sensors—possibly muscle spindles. Solid
arrows indicate excitatory connections; open circles, inhibitory connections; open arrows, mixed actions.

input from multiple canals that produce a unique strongest direct input from otolith afferents sensitive
preferred rotation direction that is likely aligned with to linear accelerations (Angelaki et al., 2001). The
that of the specific motor pools to which they project signals carried by these neurons are likely to mediate
(Perlmutter et al., 1998a). It is not known how such postural-stabilizing responses.
vestibulocollic neurons respond to linear motions. Another major source of input to neck MNs is the
Observations by Boyle et al. (1996) suggest that this reticulospinal system, which originates from several
population may include neurons that receive input groups of neurons in the medial mesencephalic,
from the cerebellar flocculus and exhibit an eye–head pontine and medullary brainstem (Robinson et al.,
velocity discharge pattern. They could provide one 1994). The best understood neurons are those that are
pathway for mediating visually guided smooth head located in the paramedian pontine reticular forma-
movements. tion and send their axons to the spinal cord via the
Finally, neck MNs also receive input from medial reticulospinal tract (mRST). The paramedian
collateral branches of vestibulospinal neurons pro- pontine reticular region is the final premotor relay
jecting to lower levels of the spinal cord (Perlmutter in the network that controls horizontal gaze shifts.
et al., 1998b). These long vestibulospinal neurons In addition to mRST neurons, it also contains
exhibit a wide variety of preferred angular rotation excitatory burst neurons (EBNs) that drive horizon-
directions including directions that activate semicir- tal eye movements via projections to abducens MNs
cular canals on the opposite side. Interestingly, it is and internuclear neurons (Strassman et al., 1986a,b).
such Type II neurons that appear to receive the Both mRST and EBN neurons carry a signal related
374

to gaze velocity. They receive strong excitatory inputs strongly to inputs from otolith organs. This raises the
from the superior colliculus (SC), a structure that possibility that the reticulospinal system may contain
specifies the desired metrics of a gaze shift by a spatial separate channels for mediating visually driven
code wherein discharge of neurons in a specific orienting responses, and slow head stabilization and
location on a retinotopic map leads to a saccade tracking movements. This suggestion requires further
appropriate to foveate a stimulus falling on the examination in animals with free heads.
corresponding portion of the retina. Transformation A third reticulospinal projection arises from the
from the spatial code in the SC to the gaze velocity region of central mesencephalic tegmentum corre-
code in the reticular formation is thought to involve sponding to nucleus subcuneiformis (Robinson et al.,
weighted projections from SC to the reticular 1994). Some neurons in this region fire prior to
formation and regulation by cerebellar circuits that saccades while others discharge later, at the point
project to the reticular formation via the cerebellar when head movement is underway (Waitzman et al.,
fastigial nucleus (Optican and Quaia, 2002). (Further 1996, 2002). Inactivation of the region results in
aspects of this system are considered below). Neurons hypermetric contralateral saccades (Waitzman et al.,
projecting in the mRST also receive input from 2000), suggesting that it is involved in the regulation
frontal cortex and the vestibular system (Peterson of gaze-shift amplitude. An intriguing possibility is
et al., 1974; Peterson and Abzug, 1975), raising the that this region plays a role in apportioning the SC
possibility that they may participate in voluntary gaze command signal between eye and neck motor
control of the proximal musculature and postural plants.
reflexes. Corticospinal projections to the neck arise from
Less well understood are the projections to neck several cortical areas including primary motor cortex,
MNs from neurons in the gigantocellular reticular the dorsal and ventral premotor areas and the
formation of the medial medulla. These descend supplementary and cingulate motor areas (He et al.,
primarily in the ipsilateral and contralateral lateral 1993, 1995). In primates, some of these projections
reticulospinal tracts (lRSTs) and have both excitatory appear to terminate on neck MNs (Dum and Strick,
and inhibitory actions on neck MNs (Peterson et al., 1996). In the cat, however, the most direct excitatory
1978). The latter pay a key role in the atonia cortico-neck MN pathway was disynaptic with a
associated with rapid eye-movement sleep (Terzuolo relay in the pontomedullary reticular formation.
et al., 1965). Like mRST neurons, these neurons Stimulation of corticospinal fibers below a transected
receive inputs from frontal cortex, SC, fastigial pyramid gave rise only to disynaptic inhibition
nucleus and vestibular and somesthetic afferents followed by longer latency excitation (Alstermark
(Peterson and Felpel, 1971; Eccles et al., 1975; et al., 1985, 1992). It is likely that cortico-
Peterson et al., 1980). This raises the possibility that reticulospinal connections are stronger than direct
they serve a wide variety of functions related to corticospinal connections in primates, as well. They
voluntary head movements, orienting and postural would allow cortical access to orienting responses
stabilization. Kitama and colleagues (Kitama et al., involving synergistic activation of large sets of axial
1995a,b), however, have reported that these functions MNs and to postural adjustments that form a vital
may be separated. Neurons that play a key role in part of many voluntary movements. The latter may
orienting movements produced by the tecto-reticu- also involve cortico-vestibulospinal pathways. The
lospinal system (Grantyn and Berthoz, 1985, 1987; special role of the direct pathways remains to be
Grantyn et al., 1987, 1992a,b; Grantyn et al., this explored experimentally.
volume; Sasaki, this volume) appear to receive Another descending pathway arises from the
relatively little vestibular input whereas other interstitial nucleus of Cajal, which functions as a
neurons in the pontine and medullary reticular controller for vertical gaze position (Fukushima,
formation respond vigorously to horizontal rotations 1987; see also Fukushima et al., Chapter 37 of this
and also receive input from the SC. These likely volume). There are also projections from the SC and
include the reticulospinal neurons examined by cerebellar fastigial nucleus that appear to terminate
Bolton et al. (1992), which responded especially on interneurons within the neck segments of the
 375

spinal cord (Robinson et al., 1994). These may be the gaze shift while the eyes are moving in the opposite
head-movement analog of projections of these direction under the influence of the VOR. In fact, eye
structures to EBNs that drive eye saccades. velocity typically reverses direction before the end of
the gaze shift, thereby reflecting an opposing
interaction between gaze drive and the partially
Neuromuscular strategies underlying suppressed VOR (Tabak et al., 1996). It was
eye–head gaze shifts originally thought that the VOR was normally active
throughout a gaze shift. This would require the
The most extensive analysis of head movements saccadic system to generate a much larger drive in
concerns their role in eye–head gaze shifts. Figure 4 order to overcome the vestibuloocular eye move-
summarizes key aspects of this behavior. ments that would oppose the gaze shift (Bizzi et al.,
The lower traces in Fig. 4A indicate the move- 1976). There is now wide agreement, however, that
ments of the eyes (in the head) and of the head in the VOR is suppressed during a gaze shift
space during an 80 gaze shift. Gaze is the sum of the (Tomlinson and Bahra, 1986a; Ron et al., 1993;
other two signals and reflects the point in space Tabak et al., 1996; Roy and Cullen, 1998). Careful
toward which the optic axes of the eyes are directed. measurements by Tabak et al. (1996) indicate that the
In this example, the eyes and head begin moving at degree of suppression in the human increases with
the same time (gaze shift onset indicated by the first amplitude of the gaze shift but is never complete. The
dotted line) but this can vary (Zangemeister et al., drawing at the top of Fig. 4A is typical, showing
1981, 1982; Zangemeister and Stark, 1982). For small a suppression of 50%. As discussed further below,
movements, the head is often delayed whereas for there is also evidence that VCRs are suppressed
large movements or highly predictable movements, during active, voluntary head movements (McCrea
the head may begin to move before the eyes (Ron et al., 1999; Roy and Cullen, 2001).
et al., 1993). In this case, the VOR drives the eyes in Figure 4B illustrates another aspect of eye–head
the opposite direction, thereby holding gaze stable behavior. Head-movement components of gaze shifts
until gaze onset. depend on eye position in the head (Guitton et al.,
The gaze shift ends 0.3 s after it began. At this 1984; Tomlinson and Bahra, 1986b; Guitton and
time, the head is still moving in the direction of the Volle, 1987; Freedman and Sparks, 1997; Goossens

Fig. 4. Eye–head behavior during gaze saccades. (A) Traces at bottom show trajectories of head in space, eye in head and of their sum
(gaze) during an 80 gaze saccade. Top trace plots VOR gain as measured by Tabak et al. (1996). Vertical lines indicate gaze shift onset,
gaze shift end and head-movement end. (B) Plot of head-movement contribution to gaze shifts as a function of the predicted eye
eccentricity at the end of the gaze shift if the head did not move (the sum of initial eye position in head and the change in gaze), as
analyzed by Stahl (1999). Axes are in degrees.
376

and Van Opstal, 1997; Stahl, 1999; Ceylan et al., activity on the colliculus during the gaze shift toward
2000). If a gaze shift can be accomplished by eye regions that code progressively smaller gaze shifts
movements alone, leaving the eyes near central (Munoz and Guitton, 1985; Munoz et al., 1991;
position, the head-movement component will be Munoz and Wurtz, 1995) but the current models
minimal. The abscissa on the Fig. 4B graph, which assign a key role in the process to the cerebellar
plots the sum of initial eye position and the gaze shift, vermis (Optican and Quaia, 2002) and its output
reflects this behavior. If the initial eye position (Ei) is pathway via the fastigial nucleus (FN).
already in the direction of the gaze shift, Ei þ
G will The eye-movement component of the gaze shift
be large, and a larger head movement will be planned passes through a gate representing the inhibitory
than if initial eye position is in the direction opposite action of omnipause neurons (OPN) before reaching
to the gaze shift. The range of allowed eye position EBNs and inhibitory burst neurons (IBNs). The
varies between individuals—those with large allowed presence of this gate explains why activity sometimes
ranges are eye movers while those with small ranges reaches head-movement circuits prior to activation of
are head movers. The range can also be adjusted to EBNs and IBNs. Activity of these bursters then
match cognitive demands of the task. For instance, drives oculomotor neurons (OM). It also charges the
when subjects view the world through pinhole glasses, RNI and the eye-position integrator that maintains
the range is greatly reduced so that gaze shifts are eccentric eye positions between gaze shifts (not shown
largely carried out by moving the head. This allows in Fig. 3). A final action arises from the projection of
the target to remain visible through the pinhole IBNs to the vestibular nuclei where they are thought
(Ceylan et al., 2000). The simple form of apportion- to inhibit PVP-type vestibuloocular neurons. This
ing gaze shifts can be accomplished by adding an eye would account for the fact that the pause exhibited by
position offset signal to the main reticulospinal these neurons is precisely aligned with the time of the
pathway carrying gaze commands to the neck as saccadic gaze shift.
shown in Fig. 3, and as observed experimentally The head-movement component of the SC-
(Grantyn and Berthoz, 1987). It is not clear, however, controlled gaze shift reaches neck MNs primarily
how the range is adjusted in accordance with task via reticulospinal neurons (RSNs). As explained
demands. above, there are separate populations of RSNs
Figure 3 summarizes current knowledge of neural responsible for horizontal and vertical head move-
circuitry involved in generating eye–head gaze shifts. ments located in the mesencephalic, pontine and
A major source of commands for such shifts is the medullary tegmentum. The box shown in Fig. 3 is
intermediate-to-deep layers of the SC, which receives a composite of these populations. It receives SC-,
inputs from many sources, including gaze centers in motor cortical- and vestibular input, which may be
the frontal and parietal cortex (F-PCx). SC discharge segregated in different populations of RSNs (Kitama
is in the form of a location code where firing of a et al., 1995a). This box also receives an eye-position
particular neuron commands movements to a specific signal, which has been observed in medial RSNs of
point on a retinotopic map. Firing intensity varies the pons (Grantyn and Berthoz, 1987; Grantyn et al.,
somewhat with the motivation of the animal and the 1992a). As explained above, this signal varies the
corresponding velocity of the resulting gaze shift, but activation of RSNs and hence the head-movement
this variation is a secondary effect. Output of the component of gaze shifts depending on the position
colliculus is thus a gaze displacement command. This of the eyes in the head.
must be converted to a gaze velocity command that A prominent input to EBNs, IBNs and RSNs
declines to zero as gaze reaches the target. Figure 3 arises from the fastigial nucleus (May et al., 1990;
shows conceptually how this can be done. A Robinson et al., 1994). Knowledge of the complex
resettable neural integrator (RNI) keeps track of signals carried by fastigial neurons, and the way in
how far gaze has shifted since the saccade began and which they are distributed to different premotoneur-
its output is subtracted from the SC signal to yield the ons, is currently quite limited. There is a similar lack
desired gaze velocity command. There is evidence of information on the role of pathways from motor
that this process may involve shifting of the locus of and premotor cortical areas of the frontal lobe
 377

(He et al., 1993, 1995; Dum and Strick, 1996) in (Boyle et al., 1996; McCrea et al., 1999; Roy and
producing head movements. Anatomical and electro- Cullen, 2001). These neurons do not cease dischar-
physiological observations indicate that these path- ging as do PVPs. Furthermore, they continue to be
ways have powerful actions on RSNs and neck MNs responsive to externally produced head movements.
(Peterson et al., 1974; Alstermark et al., 1985, 1992) There is thus selective cancellation of just that
but the signals they carry during behavior have not portion of the vestibular signal produced by a
been observed. There is also considerable controversy voluntary head movement. This cancellation extends
regarding whether the frontal eye field, a region throughout the duration of the head movement, thus
traditionally thought to be the primary premotor continuing beyond the end of the gaze shift. In Fig. 3,
cortical center for eye movements, also acts upon the this cancellation is mediated by the box labeled ‘head-
neck musculature. Given its strong excitatory action movement estimator’ (HME). This mechanism
on the SC, it would seem likely that such an action processes efference copy signals of motor commands
should exist but some groups have observed it (Tu from the SC and motor cortex channels, and head-
and Keating, 2000) while others (Chen and Sparks, movement feedback signals (HMS) to arrive at an
2001) have not. estimate of the vestibular signal that would be
The vestibular nuclei are an essential component produced by a voluntary head movement. This signal
in organizing and driving stabilizing eye and head would then inhibit vestibular neurons. The cancella-
movements that compensate for externally imposed tion appears to be less complete during slow tracking
motions of the head. They do this both via powerful head movements (K. Cullen, personal communica-
excitatory and inhibitory projections directly to tion), perhaps because the HME does not have access
extraocular and neck MNs and via projections to to signals arriving via the MST-cerebellar pathway.
reticular neurons (EBN, IBN, RSN). Also shown is (MST is the abbreviation for the cortical visual area
the direct inhibitory input from Purkinje cells of the which senses image motion.)
flocculus and nodulus. Little is known about the
latter but the former play a key role in relaying
signals that allow animals to follow small moving Future directions
targets (Lisberger et al., 1994). Smooth pursuit eye
movements are mediated by flocculus-receiving Compared to the wealth of knowledge available on
neurons in the vestibular nuclei that exhibit eye– oculomotor control, our knowledge of head-move-
head, velocity type behavior (Zhang et al., 1995a,b). ment control is at a very early stage. Significant
Some of these neurons appear to project to the neck, progress can be made by studying the head-movement
as well (Boyle et al., 1996), and may be one of the system using paradigms developed for oculomotor
pathways that contributes to smooth pursuit head studies. We need to know how neurons in each of the
movements. There are also projections to and from sites that project to the neck behave during several
the fastigial nucleus whose roles are poorly under- head-movement behaviors. Ideally, neural activity
stood. One such role may be in organizing postural would be compared across eye–head gaze shifts,
adjustments during such behaviors as locomotion. smooth head tracking, head stabilization and reaching
As in the case of the VOR, there is evidence that for food with the mouth. Such studies will be much
the VCR is suppressed during gaze shifts. Part of this more valuable if the neurons investigated are identified
involves the pause in discharge of PVP neurons that (using electrical stimulation) as projecting to the neck,
was discussed. Many of these neurons bifurcate to and perhaps also according to areas of the CNS, which
project to both eye and neck motor pools (Isu et al., project to them. This will be a major undertaking since
1988). Therefore the portion of the VCR due to their it involves recordings from several cortical areas, the
action will be absent during the time of the gaze shift SC, three reticular regions [central mesencephalic
(but not the entire duration of the head movement). reticular formation (CMRF), paramedian pontine
Several groups have also observed cancellation of the reticular formation (PPRF), gigantocellularis], the
portion of vestibular input related to a voluntary medial, lateral and inferior vestibular nuclei and the
head movement in other types of vestibular neurons cerebellar fastigial nucleus. It would be valuable to
378

record from several such sites simultaneously in order Acknowledgments

to see how their activity shifts with the animal’s
behavior (see also Kobayashi et al., this volume). The author’s work on head movements is supported
Another approach that can be adopted from our by grant number NS 41288 from the National
oculomotor colleagues is the use of neural systems Institute of Neurological Disorders and Stroke.
models like those used by Galiana and Guitton
(1992). These are indispensable tools in under-
standing the function of neural networks like that Abbreviations
shown in Fig. 3. We will also need models that deal
with the complex kinematics and dynamics of the CCR cervicocollic reflex
head–neck system. Studies reviewed above describe CMRF central mesencephalic reticular formation
the start that has been made in this direction. CNS central nervous system
Hopefully, we will eventually arrive at a neural EBN excitatory burst neuron
network model that incorporates a realistic neck F-PCx frontal and parietal cortex
motor plant and emulates what is known about its HME head-movement estimator
neural control. HMS head-movement feedback signals
Construction of the latter type of model will IBN inhibitory burst neuron
depend upon an experimental approach that is not lRST mRST, lateral and medial reticulospinal
commonly employed in oculomotor studies. This tracts
involves correlation of the activity of neck muscles MN motoneuron
and specific classes of central neurons. Two methods MST cortical visual area sensing image motion
are available. Spike-triggered averaging reveals direct PPRF paramedian pontine reticular formation
actions of the test neuron on different neck motor PVP position-vestibular-pause (neuron)
pools. Correlation of neural and EMG activity RNI resettable neural integrator
reveals the relationship between them and how it RSN reticulospinal neuron
changes across categories of behavior. Since the SC superior colliculus
required analyses can be done off-line after data are VCR vestibulocollic reflex
obtained, they can provide valuable additional VOR vestibuloocular reflex
information about each neuron studied.
This is an ambitious agenda but the author
believes that the effort will be worthwhile. As pointed References
out in Section 1, the head–neck system is in many
ways an ideal model system for studying neural Alstermark, B., Pinter, M.J. and Sasaki, S. (1985). Pyramidal
control of structures that have motor plants more effects in dorsal neck motoneurones of the cat. J. Physiol.
complex than the very simple oculomotor plant. The (Lond.), 363: 287–302.
presence of stereotyped behaviors such as gaze shifts Alstermark, B., Pinter, M.J. and Sasaki, S. (1992). Descending
pathways mediating disynaptic excitation of dorsal neck
and head stabilization has many advantages over the
motoneurones in the cat: brain stem relay. Neurosci. Res., 15:
more complex behaviors that characterize arm 42–57.
movements. But, if necessary, one can also elicit Angelaki, D.E., Green, A.M. and Dickman, J.D. (2001).
more complex behaviors with the head that explore Differential sensorimotor processing of vestibulo-ocular
features such as load compensation during voluntary signals during rotation and translation. J. Neurosci., 21:
movements or anticipatory postural adjustments. 3968–3985.
Observations during free locomotion as pioneered by Banovetz, J.M., Peterson, B.W. and Baker, J.F. (1995). Spatial
coordination by descending vestibular signals. 1. Reflex
Mori and colleagues would also be very interesting
excitation of neck muscles in alert and decerebrate cats.
(see S. Mori et al., Chapter 33 of this volume). Exp. Brain Res., 105: 345–362.
Comprehensive understanding of motor behavior Bilotto, G., Goldberg, J., Peterson, B.W. and Wilson, V.J.
certainly requires knowledge of the control of axial (1982). Dynamic properties of vestibular reflexes in the
muscle systems. The neck is an ideal place to start! decerebrate cat. Exp. Brain Res., 47: 343–352.
 379

Bizzi, E., Polit, A. and Morasso, P. (1976). Mechanisms Goossens, H.H. and Van Opstal, A.J. (1997). Human eye-head
underlying achievement of final head position. J. Neuro- coordination in two dimensions under different sensorimotor
physiol., 39: 435–444. conditions. Exp. Brain Res., 114: 542–560.
Bolton, P.S., Goto, T., Schor, R.H., Wilson, V.J., Yamagata, Y. Grantyn, A. and Berthoz, A. (1985). Burst activity of identified
and Yates, B.J. (1992). Response of pontomedullary reticu- tecto-reticulo-spinal neurons in the alert cat. Exp. Brain Res.,
lospinal neurons to vestibular stimuli in vertical planes. Role 57: 417–421.
in vertical vestibulospinal reflexes of the decerebrate cat. Grantyn, A. and Berthoz, A. (1987). Reticulo-spinal neurons
J. Neurophysiol., 67: 639–647. participating in the control of synergic eye and head
Borel, L. and Lacour, M. (1992). Functional coupling of the movements during orienting in the cat. I. Behavioral
stabilizing eye and head reflexes during horizontal and properties. Exp. Brain Res., 66: 339–354.
vertical linear motion in the cat. Exp. Brain Res., 91: Grantyn, A., Ong-Meang Jacques, V. and Berthoz, A. (1987).
191–206. Reticulo-spinal neurons participating in the control of
Borel, L., Le Goff, B., Charade, O. and Berthoz, A. (1994). synergic eye and head movements during orienting in the
Gaze strategies during linear motion in head-free humans. cat. II. Morphological properties as revealed by intra-axonal
J. Neurophysiol., 72: 2451–2466. injections of horseradish peroxidase. Exp. Brain Res., 66:
Boyle, R., Belton, T. and McCrea, R.A. (1996). Responses of 355–377.
identified vestibulospinal neurons to voluntary eye and head Grantyn, A., Berthoz, A., Hardy, O. and Gourdon, A. (1992a).
movements in the squirrel monkey. Ann. N.Y. Acad. Sci., Contributions of reticulospinal neurons to the dynamic
781: 244–263. control of head movements: presumed neck bursters. In:
Ceylan, M., Henriques, D.Y., Tweed, D.B. and Crawford, J.D. Berthoz A., Graf W. and Vidal P.P. (Eds.), The Head-Neck
(2000). Task-dependent constraints in motor control: pinhole Sensory-Motor System. Oxford University Press, New York,
goggles make the head move like an eye. J. Neurosci., 20: pp. 318–329.
2719–2730. Grantyn, A., Hardy, O., Olivier, E. and Gourdon, A. (1992b).
Chen, L.L. and Sparks, D.L. (2001). Initial head position Relationships between task-related discharge patterns and
modulates the direction and amplitude of gaze shifts evoked axonal morphology of brainstem projection neurons involved
by microstimulation in the frontal eye field of head-unrest- in orienting eye and head movements. In: Shimazu H. and
rained monkeys. Soc. Neurosci. Abstr., 27: 784.5. Shinoda Y. (Eds.), Vestibular and Brainstem Control of Eye
Choi, H., Keshner, E. and Peterson, B.W. (2000) Primate neck and Head and Body Movements. Japan Scientific Society
muscle activation patterns during tracking in two postures. Press, Tokyo, pp. 255–273.
Soc. Neurosci. Abstr. Guitton, D. and Volle, M. (1987). Gaze control in humans: eye-
Dum, R.P. and Strick, P.L. (1996). Spinal cord terminations head coordination during orienting movements to targets
of the medial wall motor areas in macaque monkeys. within and beyond the oculomotor range. J. Neurophysiol.,
J. Neurosci., 16: 6513–6525. 58: 427–459.
Eccles, J.C., Nicoll, R.A., Schwarz, W.F., Taborikova, H. and Guitton, D., Douglas, R.M. and Volle, M. (1984). Eye-head
Willey, T.J. (1975). Reticulospinal neurons with and without coordination in cats. J. Neurophysiol., 52: 1030–1050.
monosynaptic inputs from cerebellar nuclei. J. Neurophysiol., He, S.Q., Dum, R.P. and Strick, P.L. (1993). Topographic
38: 513–530. organization of corticospinal projections from the frontal
Fernandez, C. and Goldberg, J.M. (1971). Physiology of lobe: motor areas on the lateral surface of the hemisphere.
peripheral neurons innervating semicircular canals of the J. Neurosci., 13: 952–980.
squirrel monkey. II. Response to sinusoidal stimulation and He, S.Q., Dum, R.P. and Strick, P.L. (1995). Topographic
dynamics of peripheral vestibular system. J. Neurophysiol., organization of corticospinal projections from the frontal
34: 661–675. lobe: motor areas on the medial surface of the hemisphere.
Freedman, E.G. and Sparks, D.L. (1997). Eye-head coordi- J. Neurosci., 15: 3284–3306.
nation during head-unrestrained gaze shifts in rhesus Highstein, S.M., Goldberg, J.M., Moschovakis, A.K. and
monkeys. J. Neurophysiol., 77: 2328–2348. Fernandez, C. (1987). Inputs from regularly and irregularly
Fukushima, K. (1987). The interstitial nucleus of Cajal and its discharging vestibular nerve afferents to secondary neurons
role in the control of movements of head and eyes. Prog. in the vestibular nuclei of the squirrel monkey. II.
Neurobiol., 29: 107–192. Correlation with output pathways of secondary neurons.
Galiana, H.L. and Guitton, D. (1992). Central organization J. Neurophysiol., 58: 719–738.
and modeling of eye-head coordination during orienting gaze Isu, N., Uchino, Y., Nakashima, H., Satoh, S., Ichikawa, T.
shifts. Ann. N.Y. Acad. Sci., 656: 452–471. and Watanabe, S. (1988). Axonal trajectories of posterior
Goldberg, J. and Peterson, B.W. (1986). Reflex and mech- canal-activated secondary vestibular neurons and their
anical contributions to head stabilization in alert cats. coactivation of extraocular and neck flexor motoneurons in
J. Neurophysiol., 56: 857–875. the cat. Exp. Brain Res., 70: 181–191.
380

Keshner, E.A. and Peterson, B.W. (1988). Motor control Optican, L.M. and Quaia, C. (2002). Distributed model of
strategies underlying head stabilization and voluntary head collicular and cerebellar function during saccades. Ann. N.Y.
movements in humans and cats. Prog. Brain Res., 76: Acad. Sci., 956: 164–177.
329–339. Pellionisz, A. (1988). Tensorial aspects of the multidimensional
Keshner, F.A. and Peterson, B.W. (1995). Mechanisms massively parallel sensorimotor function of neuronal net-
controlling human head stabilization. I. Head-neck works. Prog. Brain Res., 76: 341–354.
dynamics during random rotations in the horizontal plane. Peng, G.C., Hain, T.C. and Peterson, B.W. (1996a). A
J. Neurophysiol., 73: 2293–2301. dynamical model for reflex activated head movements in
Keshner, E.A., Campbell, D., Katz, R.T. and Peterson, B.W. the horizontal plane. Biol. Cybern., 75: 309–319.
(1989). Neck muscle activation patterns in humans during Peng, G.C., Hain, T.C. and Peterson, B.W. (1996b) How is the
isometric head stabilization. Exp. Brain Res., 75: 335–344. head held up? Modeling mechanisms for head stability in the
Keshner, E.A., Baker, J.F., Banovetz, J. and Peterson, B.W. sagittal plane. Proc. 18th Annu. Int. Conf. IEEE Eng. Med.
(1992). Patterns of neck muscle activation in cats during Biol. Soc., Amsterdam.
reflex and voluntary head movements. Exp. Brain Res., 88: Perlmutter, S.I., Iwamoto, Y., Baker, J.F. and Peterson, B.W.
361–374. (1998a). Interdependence of spatial properties and projection
Keshner, E.A., Cromwell, R.L. and Peterson, B.W. (1995). patterns of medial vestibulospinal tract neurons in the cat.
Mechanisms controlling human head stabilization. II. Head- J. Neurophysiol., 79: 270–284.
neck characteristics during random rotations in the vertical Perlmutter, S.I., Iwamoto, Y., Barke, L.F., Baker, J.F. and
plane. J. Neurophysiol., 73: 2302–2312. Peterson, B.W. (1998b). Relation between axon morphology
Kitama, T., Grantyn, A. and Berthoz, A. (1995a). Orienting-
in C1 spinal cord and spatial properties of medial vestibulo-
related eye-neck neurons of the medial ponto-bulbar reticular
spinal tract neurons in the cat. J. Neurophysiol., 79: 285–303.
formation do not participate in horizontal canal-dependent
Peterson, B.W. and Abzug, C. (1975). Properties of projections
vestibular reflexes of alert cats. Brain Res. Bull., 38: 337–347.
from vestibular nuclei to medial reticular formation in the
Kitama, T., Ohki, Y., Shimazu, H., Tanaka, M. and
cat. J. Neurophysiol., 38: 1421–1435.
Yoshida, K. (1995b). Site of interaction between saccade
Peterson, B.W. and Felpel, L.P. (1971). Excitation and
signals and vestibular signals induced by head rotation in the
inhibition of reticulospinal neurons by vestibular, cortical
alert cat: functional properties and afferent organization of
and cutaneous stimulation. Brain Res., 27: 373–376.
burster-driving neurons. J. Neurophysiol., 74: 273–287.
Peterson, B.W., Anderson, M.E. and Filion, M. (1974).
Lacour, M. and Borel, L. (1989). Functional coupling of the
Responses of ponto-medullary reticular neurons to cortical,
stabilizing gaze reflexes during vertical linear motion in the
tectal and cutaneous stimuli. Exp. Brain Res., 21: 19–44.
alert cat. Prog. Brain Res., 80: 385–394 discussion 373–375.
Peterson, B.W., Pitts, N.G., Fukushima, K. and Mackel, R.
Lisberger, S.G., Pavelko, T.A. and Broussard, D.M. (1994).
(1978). Reticulospinal excitation and inhibition of neck
Responses during eye movements of brain stem neurons that
motoneurons. Exp. Brain Res., 32: 471–489.
receive monosynaptic inhibition from the flocculus and
Peterson, B.W., Fukushima, K., Hirai, N., Schor, R.H. and
ventral paraflocculus in monkeys. J. Neurophysiol., 72:
Wilson, V.J. (1980). Responses of vestibulospinal and
909–927.
May, P.J., Hartwich-Young, R., Nelson, J., Sparks, D.L. and reticulospinal neurons to sinusoidal vestibular stimulation.
Porter, J.D. (1990). Cerebellotectal pathways in the macaque: J. Neurophysiol., 43: 1236–1250.
implications for collicular generation of saccades. Peterson, B.W., Goldberg, J., Bilotto, G. and Fuller, J.H.
Neuroscience, 36: 305–324. (1985). Cervicocollic reflex: its dynamic properties and
McCrea, R.A., Gdowski, G.T., Boyle, R. and Belton, T. (1999). interaction with vestibular reflexes. J. Neurophysiol., 54:
Firing behavior of vestibular neurons during active and 90–109.
passive head movements: vestibulo-spinal and other non-eye- Robinson, D.A. (1982). The use of matrices in analyzing the
movement related neurons. J. Neurophysiol., 82: 416–428. three-dimensional behavior of the vestibulo-ocular reflex.
Munoz, D.P. and Guitton, D. (1985). Tectospinal neurons in Biol. Cybern., 46: 53–66.
the cat have discharges coding gaze position error. Brain Robinson, F.R., Phillips, J.O. and Fuchs, A.F. (1994).
Res., 341: 184–188. Coordination of gaze shifts in primates: brainstem inputs to
Munoz, D.P. and Wurtz, R.H. (1995). Saccade-related activity neck and extraocular motoneuron pools. J. Comp. Neurol.,
in monkey superior colliculus. II. Spread of activity during 346: 43–62.
saccades. J. Neurophysiol., 73: 2334–2348. Ron, S., Berthoz, A. and Gur, S. (1993). Saccade-vestibulo-
Munoz, D.P., Pelisson, D. and Guitton, D. (1991). Movement ocular reflex co-operation and eye-head uncoupling during
of neural activity on the superior colliculus motor map orientation to flashed target. J. Physiol. (Lond.), 464:
during gaze shifts. Science, 251: 1358–1360. 595–611.
 381

Roy, J.E. and Cullen, K.E. (1998). A neural correlate for Waitzman, D.M., Silakov, V.L. and Cohen, B. (1996). Central
vestibulo-ocular reflex suppression during voluntary eye- mesencephalic reticular formation (cMRF) neurons dis-
head gaze shifts. Nat. Neurosci., 1: 404–410. charging before and during eye movements. J. Neurophysiol.,
Roy, J.E. and Cullen, K.E. (2001). Selective processing 75: 1546–1572.
of vestibular reafference during self-generated head motion. Waitzman, D.M., Silakov, V.L., DePalma-Bowles, S. and
J. Neurosci., 21: 2131–2142.
Ayers, A.S. (2000). Effects of reversible inactivation of the
Stahl, J.S. (1999). Amplitude of human head movements associ-
primate mesencephalic reticular formation. I. Hypermetric
ated with horizontal saccades. Exp. Brain Res., 126: 41–54.
Strassman, A., Highstein, S.M. and McCrea, R.A. (1986a). goal-directed saccades. J. Neurophysiol., 83: 2260–2284.
Anatomy and physiology of saccadic burst neurons in the Waitzman, D.M., Pathmanathan, J., Presnell, R., Ayers, A. and
alert squirrel monkey. I. Excitatory burst neurons. J. Comp. DePalma, S. (2002). Contribution of the superior colliculus
Neurol., 249: 337–357. and the mesencephalic reticular formation to gaze control.
Strassman, A., Highstein, S.M. and McCrea, R.A. (1986b). Ann. N.Y. Acad. Sci., 956: 111–129.
Anatomy and physiology of saccadic burst neurons in the Wilson, V.J., Peterson, B.W., Fukushima, K., Hirai, N. and
alert squirrel monkey. II. Inhibitory burst neurons. J. Comp. Uchino, Y. (1979). Analysis of vestibulocollic reflexes by
Neurol., 249: 358–380. sinusoidal polarization of vestibular afferent fibers.
Tabak, S., Smeets, J.B. and Collewijn, H. (1996). Modulation
J. Neurophysiol., 42: 331–346.
of the human vestibuloocular reflex during saccades: probing
Zangemeister, W.H. and Stark, L. (1982). Types of gaze
by high-frequency oscillation and torque pulses of the head.
movement: variable interactions of eye and head movements.
J. Neurophysiol., 76: 3249–3263.
Terzuolo, C.A., Llinas, R. and Green, K.T. (1965). Mechanisms Exp. Neurol., 77: 563–577.
of supraspinal actions upon spinal cord activities: distri- Zangemeister, W.H., Jones, A. and Stark, L. (1981). Dynamics
bution of reticular and segmental inputs in cat’s alpha- of head movement trajectories: main sequence relationship.
motoneurons. Arch. Ital. Biol., 103: 635–651. Exp. Neurol., 71: 76–91.
Tomlinson, R.D. and Bahra, P.S. (1986a). Combined eye-head Zangemeister, W.H., Stark, L., Meienberg, O. and Waite, T.
gaze shifts in the primate. I. Metrics. J. Neurophysiol., 56: (1982). Neural control of head rotation: electromyographic
1542–1557. evidence. J. Neurol. Sci., 55: 1–14.
Tomlinson, R.D. and Bahra, P.S. (1986b). Combined eye-head Zhang, Y., Partsalis, A.M. and Highstein, S.M. (1995a).
gaze shifts in the primate. II. Interactions between saccades Properties of superior vestibular nucleus flocculus target neu-
and the vestibuloocular reflex. J. Neurophysiol., 56:
rons in the squirrelmonkey. I. Generalproperties in comparison
1558–1570.
with flocculus projecting neurons. J. Neurophysiol., 73:
Tu, T.A. and Keating, E.G. (2000). Electrical stimulation of the
frontal eye field in a monkey produces combined eye and 2261–2278.
head movements. J. Neurophysiol., 84: 1103–1106. Zhang, Y., Partsalis, A.M. and Highstein, S.M. (1995b). Proper-
Vasavada, A.N., Peterson, B.W. and Delp, S.L. (2002). Three- ties of superior vestibular nucleus flocculus target neurons in the
dimensional spatial tuning of neck muscle activation in squirrel monkey. II. Signal components revealed by reversible
humans. Exp. Brain Res., 147: 437–448. flocculus inactivation. J. Neurophysiol., 73: 2279–2292.