Search

The behavior characterizing choice response decision-making was studied in monkeys to provide background information for ongoing neurophysiological studies of the neural mechanisms underlying saccadic choice decisions. Animals were trained to associate a specific color from a set of colored visual stimuli with a specific spatial location. The visual stimuli (colored disks) appeared briefly at equal eccentricity from a central fixation position and then were masked by gray disks. The correct target association was subsequently cued by the appearance of a colored stimulus at the fixation point. The animal indicated its choice by saccading to the remembered location of the eccentric stimulus, which had matched the color of the cue. The number of alternative associations (NA) varied from 1 to 4 and remained fixed within a block of trials. After the training period, performance (percent correct responses) declined modestly as NA increased (on average 96, 93 or 84% correct for 1, 2 or 4 NA, respectively). Response latency increased logarithmically as a function of NA, thus obeying Hick's law. The spatial extent of the learned association between color and location was investigated by rotating the array of colored stimuli that had remained fixed during the learning phase to various different angles. Error rates in choice saccades increased gradually as a function of the amount of rotation. The learned association biased the direction of the saccadic response toward the quadrant associated with the cue, but saccade direction was always toward one of the actual visual stimuli. This suggests that the learned associations between stimuli and responses were not spatially exact, but instead the association between color and location was distributed with declining strength from the trained locations. These results demonstrate that the saccade system in monkeys also displays the characteristic dependence on NA in choice response latencies, while more basic features of the eye movements are invariant from those in other tasks. The findings also provide behavioral evidence that spatially distributed regions are established for the sensory-to-motor associations during training which are later utilized for choice decisions.

1. Anatomical and single-unit recording studies suggest that the dorsolateral pontine nucleus (DLPN) in monkey is a major link in the projection of descending visual motion information to the cerebellum. Such studies coupled with cortical and cerebellar lesion results suggest a major role for this basilar pontine region in the mediation of smooth-pursuit eye movements. 2. To provide more direct evidence that this pontine region is involved in the control of smooth-pursuit eye movements, focal chemical lesions were made in DLPN in the vicinity of previously recorded visual motion and pursuit-related neurons. Eye movement responses were subsequently recorded in these lesioned animals under several behavioral paradigms. 3. A major deficit in smooth-pursuit performance was produced after unilateral DLPN lesions generated either reversibly with lidocaine or more permanently with ibotenic acid. Pursuit impairments were observed during steady-state tracking of sinusoidal target motion as well as during the initiation of pursuit tracking to sudden ramp target motion. Through the use of the latter technique, initial eye acceleration was reduced to less than one-half of normal for animals with large lesions of the dorsolateral and lateral pontine nuclei. 4. The pursuit deficit in all animals was directional in nature and was not dependent on the visual hemifield in which the motion stimulus occurred. The largest effect for horizontal tracking occurred in all animals for pursuit directed ipsilateral to the lesion. Animals also showed major deficits in one or both directions of vertical pursuit, although the primary direction of the vertical impairment was variable from animal to animal. 5. Chemical lesions in the DLPN also produced comparable deficits in the initiation of optokinetic-induced smooth eye movements in the ipsilateral direction. In contrast to this effect on the initial optokinetic response, in the one lesioned animal studied during prolonged constant velocity optokinetic drum rotation, smooth eye speed increased slowly over a 10- to 15-s period to reach a level that closely matched drum speed. These results suggest that pathways outside the DLPN can generate the steady-state optokinetic response. 6. Saccades to stationary targets were normal after DLPN lesions, but corrective saccades made to targets moving in the direction ipsilateral to the lesion were much more hypometric than similar prelesion control saccades. 7. The pursuit deficits produced by lidocaine injections recovered within 30 min. The ibotenic acid deficits were maximal approximately 1 day after the injection and recovered rapidly thereafter over a time period of 3-7 days.(ABSTRACT TRUNCATED AT 400 WORDS)

The cerebellar vermal cortex in monkey was electrically stimulated through a microelectrode with brief pulse strains synchronized to the onset of selected saccadic eye movements. The delivery of these intrasaccadic microstimuli were randomly interposed in groups of saccadic eye movements made by trained monkeys to known target locations. The trajectory of the saccades with intrasaccadic stimuli were then compared to those made to the same target without stimulation. Contralaterally directed saccades were consistently slowed from their normal trajectory (normally at 12 ms after stimulus onset) and were made hypometric. Ipsilaterally directed saccades were not affected. Vertical eye movement components were also affected at some stimulating sites. Changing the stimulus train duration from 10 to 80 ms had no effect on the change in velocity or amount of dysmetria produced in contralateral saccades. Taken together these results suggest that a different set of cerebellar elements were being stimulated by the intrasaccadic mode as contrasted to stimulation at the same site at random times. Furthermore, the short period of time during an ongoing saccade presents a unique instant in which to test the relationship of various central structures to the saccadic system. Based on current concepts of the organization of the brainstem saccadic pulse generators, these data are interpreted to suggest that: (1) the cerebellar vermis has a rather direct influence on the contralateral saccade generator, (2) this influence can act to recode (diminish) the generator's representation of the size of motor error, (3) the effect occurs at a point in the generator where motor error is spatially coded.

A small region in the dorsal midline portion of the nucleus reticularis tegmenti pontis (NRTP) in monkeys contains neurons that respond to focal visual stimuli or during saccadic eye movements or both. None of these cells or any others in this region respond to the motion of large visual fields (optokinetic stimulation), although such responses were specifically sought. Thus, this group of NRTP neurons forms a completely different set of cells than those previously described in more rostral but closely adjacent portions of the pontine nuclei which respond well to optokinetic stimulation. The most frequently encountered cell type in this region of NRTP (153 neurons) produced a high-frequency burst of discharges during saccadic eye movements. Neural discharge (burst intensity or duration) was not related to saccade metrics. Instead, peak burst frequency and/or the number of spikes in a unit's burst reached a maximum when the saccade moved the eyes to a circumscribed region (movement field) of the animal's visual field. There were two subtypes of these burst neurons. In one type (44%) the movement fields were smaller and entirely contained within the oculomotor range. In the other type (56%) the movement fields consisted of a whole sector (some as wide as 180 degrees) of the entire oculomotor range. All the neurons in this sample that we were able to test in total darkness continued to produce bursts of discharges of similar profile during spontaneous saccades into their movement field. All the movement fields were retinotopically organized, although a few cells (22%) showed a marked variation of burst metrics with initial eye position. Another small group of cells in NRTP (8 neurons) responded to small spots of light turned on within a circumscribed region of the visual field while the animal maintained fixation on a separate spot of light. These visual neurons produced no saccade-related discharge. A larger group of neurons (24 out of 52 tested cells) produced both a visual response and a saccadic burst. The visual field of this type of cell was always smaller and was contained within the movement field of the cell. The response of both types of NRTP visual neurons was enhanced when the visual stimulus was to be the target for a saccadic eye movement. On double-saccade trials the visual stimulus was never present in the hemifield containing the cell's visual field.(ABSTRACT TRUNCATED AT 400 WORDS)

The behavior of single vestibular nerve fibers from the lateral semicircular canal was recorded during sinusoidal oscillations of the head, during optokinetic stimulation with the head stationary, and during spontaneous oculomotor behavior in the alert monkey. The response of similar fibers to adequate vestibular stimulation was also studied in some of the animals under deeply anesthetized conditions. In the alert animals all units were spontaneously active and their discharge was modulated only by adequate vestibular stimulation. Ipsilateral horizontal rotations of the head were excitatory for all units. No modification of this basic vestibular response by visual stimulation including full-field striped drum rotation was observed. Furthermore no correlation of unit activity with oculomotor function including voluntary saccadic and pursuit eye movements was found in any of the units. The regularity of spontaneous discharge was the most consistent characteristic that differentiated the unit response into types. Most units were very regular in discharge, but a few were very irregular. The averaging of unit discharge over several cycles of oscillatory head rotation showed that the irregular type units were also consistently modulated by adequate vestibular stimulation. Both regular and irregular type units were found in the anesthetized animals. Unimodal distributions of the quantitative values for unit resting discharge rate, sensitivity, and phase relationship were found. The distributions for these three parameters were similar in the units recorded in the anesthetized animals. Thus at least these characteristics of semicircular canal response seem not to be affected by the vestibular efferent system which should be altered or eliminated in the case of the anesthetized animals.

Smooth pursuit eye movements in the monkey were studied by measuring the average eye acceleration in the first 100 msec of the ocular response to suddenly moving constant velocity discrete visual targets. The tracking target's motion was initiated on or eccentric to the fovea against a very dimly lit homogeneous background or an illuminated, highly textured, large-field background. The effect of the textured background was always to reduce the initial eye acceleration as compared to that present in the homogeneous background case. Movement of the background in the opposite direction to target spot movement also reduced the initial eye acceleration. In contrast, steady-state tracking measured several hundred msec after the onset of pursuit under the same conditions was only marginally reduced by the presence of the textured background.

The cerebellar vermis (lobules V to VII) was focally stimulated (maximum current = 300 microA) through microelectrodes in alert monkeys trained to fixate on visual targets located at different positions in the eye movement field. Microstimulation of this area evoked saccades whose direction and amplitude were dependent on the spatial locus of the vermal point stimulated and on the position of the eye at the time of stimulation. Stimulation evoked saccades on most (70%) of the electrode penetrations. By alternatively stimulating and recording through the electrode as it advanced through the depth of the vermis, it was possible to map the threshold for evoking saccades along a penetration as well as to ascertain the type of tissue (white matter or Purkinje cell layer) situated at the stimulation points. On penetrations where saccades were evoked, there was generally one (90%) and sometimes two (10%) low-threshold region(s). These low-threshold regions were located in fiber tracks and not in the Purkinje or other cellular layers of the cerebellar cortex. The direction and size of the evoked saccades were dependent on position of the eye in the orbit. At a few sites, even the presence or the absence of an evoked saccade depended on the initial eye position. Postsaccadic drifts after termination of evoked saccades were also a common feature (50% of the tracks) associated with vermal microstimulation. The presence or absence as well as the direction of the postsaccadic drift were also dependent on initial eye position. These observations suggest that the vermal stimulations evoked saccades by the antidromic activation of mossy fiber afferent inputs that emanate from the brain stem saccadic burst generator. Furthermore, stimulation in this manner would actually bypass the cerebellar circuitry and produce saccades without the usual modifying influence of the cerebellum.

The vestibulo-ocular reflex (VOR) was measured in congenitally blind, adventitiously blind, and normally sighted adults to determine how it was affected by loss of vision. VOR gain and phase were measured in subjects rotated sinusoidally in total darkness, while concentrating on an imaginary earth-fixed target. Gain was lower in adventitiously blind subjects than in sighted subjects. The gain reduction in blind subjects was accompanied by an increase in the amount of phase lead at low frequencies. The dominant time constant was typically 3 sec for adventitiously blind subjects and 16 sec for sighted subjects. No convincing vestibulo-ocular responses were measured in congenitally blind subjects except at the highest test frequencies. These findings demonstrate that vision is necessary early in life for development of the VOR, and that vision is also necessary throughout life for ongoing maintenance of the VOR.

Head velocity-related modulations in Purkinje cell activity were observed in lobules VI and VII of the cerebellar vermis. The sensitivities and phase shifts with respect to head velocity of Purkinje cells were comparable with vestibular responses observed in the fastigial nuclei. The results support the hypothesis of a vermal target velocity correlate and necessitate the inclusion of vermal lobules VI and VII among the cerebellar regions involved with vestibular information processing.

1. The observation of smooth-pursuit eye and retinal image velocity signals in lobules VI and VII of the vermis has given rise to the hypothesis that a neural correlate of a target velocity signal exists in this region of the cerebellum (29). However, activity signaling head velocity is also required to regenerate a target velocity signal. Vermal Purkinje cell activity was, therefore, recorded during the performance of paradigms designed to dissociate head movement-related responses. 2. The activity of 107 Purkinje cells was found to be related to horizontal head velocity. Of these, 52% increased their discharge rate for ipsilaterally directed passive head movement (type I), and 48% were excited by contralateral head movements (type II). 3. In five Purkinje cells in which sufficient data were obtained, cell discharge rate increased monotonically with head velocity over the range of 5-40 deg/s. The sensitivity to head velocity at 0.4 Hz +/- 25 deg/s averaged approximately 0.5 spikes.s-1/deg.s-1 in a larger sample of cells (n = 39). The sensitivities to head velocity, at this same frequency and velocity, of type I and type II Purkinje cells were comparable at 0.44 and 0.51 spikes.s-1/deg.s-1, respectively. 4. The Purkinje cell responses led head velocity by an average of 12 degrees at 0.4 Hz +/- 25 deg/s of passive head rotation. The phase shifts associated with type I and II responses were similar with phase leads of 13 and 9 degrees with respect to head velocity, respectively. 5. A linear interaction of smooth-pursuit eye and head velocity signals was observed during the performance of a variety of antiphase and inphase eye and head movement paradigms. The results support the conclusion that some Purkinje cells in lobules VI and VII of the cerebellar vermis encode a gaze velocity signal. Contributions of the head velocity signal to the regeneration of target velocity are considered in a companion paper (32).

1. Purkinje cell activity was recorded from lobules VI and VII of the cerebellar vermis during the performance of visuooculomotor tasks designed to dissociate the signals related to head, smooth-pursuit eye, and retinal image movements. Task-related modulations in the simple spike discharge rates of 157 cells were observed in three alert monkeys. 2. Of 65 Purkinje cells that were completely tested for all three signals, all exhibited smooth-pursuit eye movement-related activity. An additional vestibular or visual response was observed in 17 and 11% of the cells, respectively. Eye, head, and retinal image velocity signals were all recorded in the same unit in 52% of the Purkinje cells. The responses of 5% of the fully tested cells were associated with changes in the direction of eye, head, and retinal image movement. 3. The observed sensorioculomotor responses were direction selective in 98% of the Purkinje cells. For the Purkinje cells that were fully tested, 60% of the cells exhibited peak discharge rates for ipsilateral and 40% for contralateral eye velocity. Of these Purkinje cells, 45% exhibited eye, head, and retinal image velocity signals with equivalent direction preferences. 4. Of 42 Purkinje cells tested, 88% demonstrated some kinds of interactive responses during combined eye and sensory stimulation. The interaction of eye and head velocity signals has been discussed in a companion paper (38). The modulation in discharge rate observed during tracking in the presence of a random dot background pattern could be predicted from the dissociated responses to smooth pursuit in the dark and to movements of the background pattern during suppression of eye movements. 5. The sensitivity to smooth-pursuit eye velocity averaged 1.4 times the sensitivity to head velocity. In 80% of the Purkinje cells, however, the sensitivity to eye velocity exceeded the sensitivity to head velocity by an average of only 10%. The sensitivity to smooth-pursuit eye velocity averaged 1.6 times the sensitivity to retinal image velocity. 6. An increase in Purkinje cell discharge rate was observed during the open-loop period of the initiation of smooth-pursuit eye movements. This open-loop response was consistent with the presence of a visual signal during ocular pursuit, since these cells were also shown to be responsive to a dissociated retinal image velocity signal. Furthermore, the magnitude of the open-loop response indicated an enhancement of the sensitivity to retinal image velocity when visual information became behaviorally significant.(ABSTRACT TRUNCATED AT 400 WORDS)

The central nervous system is a parallel dynamical system which connects sensory input with motor output for the performance of visual tracking. This paper applies elementary control system tools to extend dynamical neural network models to the visual smooth pursuit system. Observed eye position responses to target motions and characteristics of the plant (eye muscles and orbital mechanics) place dynamical constraints on the interposed neural network controller. In the process of constructing a model for the controller, we show two previous pursuit system models, using efference copy and feedforward compensation, are equivalent from an input-output standpoint. We introduce a controller model possessing a potentially highly parallel implementation and offer an example with supporting neural firing rate data. Changes in time delays or other system dynamics are expected to lead to compensatory adaptive changes in the controller. A scheme to noninvasively simulate such changes in system dynamics was developed. Actual physiologic data of adaptive responses to increased time delay is presented as an example of the utility of this parallel controller. Compensatory changes in our parallel controller model are easily predicted. These results suggest a productive interaction between neural network modeling, neurophysiology, and control systems engineering.