INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The vestibuloocular reflex (VOR) functions to stabilize images on the retina during head movements by rotating the eyes. The angular vestibuloocular reflex (AVOR) is evoked by angular rotations of the head that stimulate the vestibular semicircular canals. The eye movements evoked by the AVOR vary inversely as a function of viewing distance (Chen-Huang and McCrea 1998b; Snyder and King 1992; Viirre et al. 1986).

Several observations suggest that the increase in AVOR eye movement gain during near viewing reflects changes in the processing of vestibular signals in brain stem VOR pathways. During near target viewing, sudden step changes in head velocity evoke eye movements with an enhanced gain at latencies that are shorter than even the shortest visual feedback-driven eye movements (Crane and Demer 1997, 1998; Snyder et al. 1992; Viirre et al. 1986). A viewing distance-related increase in AVOR gain has been demonstrated to occur before the convergent eye movements that produce a change in viewing distance (Snyder et al. 1992). This gain increase persists in the absence of a target in complete darkness (Chen-Huang and McCrea 1998b; Hine and Thorn 1987; Snyder et al. 1992). Finally, viewing distance gain increases in the AVOR are significantly reduced when irregular vestibular afferents are silenced (Chen-Huang and McCrea 1998b). These observations have led to the suggestion that some, or possibly all of the central vestibular signals that drive the VOR are multiplied by an internal estimate of vergence angle or viewing distance (Chen-Huang and McCrea 1998b; Paige and Tomko 1991; Viirre et al. 1986).

If the changes in the AVOR are due to parametric changes in central pathways that mediate the reflex, it is not clear how or where these changes occur. Recently, McConville et al. (1996) carried out a single-unit recording study in rhesus monkeys in which the AVOR responses of eye movement-related units were recorded during viewing of near and far targets. They found that some classes of eye movement-related units were more sensitive to viewing distance than others. Specifically, the rotational responses of eye-head-velocity (EHV) units (whose response to vestibular stimulation typically reverses when the VOR is voluntarily canceled or suppressed) were significantly enhanced during near target viewing. On the other hand, the responses of position-vestibular-pause (PVP) units were usually unaffected by viewing distance, once the eye position sensitivity of those cells was taken into account. The rotational responses of units that were not sensitive to eye movements were not affected by viewing distance (Tomlinson et al. 1996). Although based on a relatively small sample of cells, these observations suggest that viewing distance-related changes in the AVOR may be mediated by a specialized subset of the available VOR pathways.

In this study we examined in further detail the viewing distance-related changes in the angular head rotation sensitivity of different classes of horizontal canal-related secondary vestibular neurons. We focused particularly on eye movement-related neurons in the rostral part of the medial vestibular nucleus of the squirrel monkey, because such neurons are likely to be related to the VOR (Cullen et al. 1993; Scudder and Fuchs 1992). We report that the signals generated by most of the horizontal canal-related neurons putatively involved in mediating the VOR change as a function of viewing distance. The changes in sensitivity to head rotation appear to be more than sufficient to produce viewing distance-related changes in the VOR. We hypothesize that viewing distance modifications in the AVOR are produced by modifying the gain of indirect inhibitory vestibular afferent inputs to VOR pathways.

Some of these results have been previously described in a preliminary report (Chen-Huang and McCrea 1998a).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surgical procedures

Two squirrel monkeys were prepared for chronic recording of bilateral eye movements, single-unit activity and for electrical stimulation of both labyrinths. Surgeries were carried out under sterile conditions on anesthetized animals (pentobarbital sodium, 20 mg/kg, initial dose; supplemental doses of 1-2 mg/kg administered as necessary). A flattened stainless steel bolt, to be used for head restraint in the plane of the horizontal semicircular canal, was attached to the occipital bone with dental acrylic. A Plexiglas cylindrical recording chamber (15 mm diam), was stereotaxically implanted on the surface of the left parietal bone for introducing recording microelectrodes into the brain stem. A search coil was sutured to the sclera of the right eye, and the twined leads of the coils were soldered to a connector that was cemented to the skull.

On a subsequent day, a second search coil was implanted on the left eye, and bipolar stimulating electrodes were implanted into both labyrinths. One of each pair of labyrinthine stimulating electrodes (chlorided silver wire, 0.25 mm diam and Teflon insulated to within 1 mm of tip) was inserted into the perilymphatic space through a hole in the bony promontory midway between the round and oval windows. The second electrode was placed in the ventral aspect of the middle ear. Leads from both electrodes were led out of the middle ear near the insertion of the tympanic membrane and soldered to a skull-mounted connector.

Experimental recording conditions

During experiments the monkey was seated in a Plexiglas chair on a vestibular turntable. Each animal was trained to fixate two types of visual targets. Rewards were contingent on vergence angle and vertical eye position. In some experimental paradigms (Fig. 1A) the monkey fixated small targets that were projected onto a blank cylindrical screen 90 cm distant from the axis of rotation. In other paradigms the monkey looked through a door in the screen at an earth-stationary far target 1.52 m distant from the axis of rotation (Fig. 1). All experiments were carried out in a dimly lit room (2 × 4 m in size) in which the monkey was isolated. Squirrel monkeys typically fixated targets for brief periods (0.5-2 min) on demand for periods of 6 h or more.

View larger version (16K): [in this window] [in a new window]   Fig. 1. A: arrangement of visual targets. Squirrel monkeys were seated on a vestibular turntable facing a cylindrical screen 90 cm away. A position-controlled, turntable-mounted laser beam (dashed line) was focused on the screen for evoking smooth pursuit eye movements and vestibuloocular reflex (VOR) cancellation to identify different functional classes of vestibular units. The effects of viewing distance on the AVOR was studied by opening the screen door, which was located directly in front of the monkey, and having the monkeys fixate light-emitting diode (LED) targets that were 150 or 10 cm away from the eyes. The door opening was 42.8°. B and C: average viewing distance-related changes in the gain (B) and phase (C) of horizontal eye movements recorded concomitantly with the single-unit recordings in this study. The 3 frequencies of turntable rotation were 0.7, 1.9, and 4.0 Hz. Standard errors for data at 0.7 and 1.9 Hz were smaller than symbols.

Turntable rotation and head movement recording

Passive head rotations in the plane of the horizontal semicircular canals (15° nose down from the stereotaxic plane) were produced with a position servo-control system whose output modified the command of a velocity servo turntable controller (Inland model 832). The axis of turntable rotation, which was parallel to earth-vertical, passed through the squirrel monkey's midsagittal plane, where it intersects with the line connecting the two external auditory meatuses. The responses of single units were typically recorded during rotations at several different frequencies and velocities of table rotation (most commonly, 0.7 Hz, peak velocity 20-40°/s; 1.9 Hz, 10-20°/s and 4.0 Hz, 6-10°/s). Each stimulus frequency produced a slightly higher peak head acceleration (88, 119, and 150°/s² for the 0.7-, 1.9-, and 4.0-Hz stimuli, respectively) and a significant decrease in head displacement (4.55, 0.84, and 0.24° for the 0.7-, 1.9-, and 4.0-Hz stimuli, respectively). Angular table velocity was recorded with an angular velocity transducer (Watson Industries).

Eye movement recording

Eye movements were measured with a magnetic search-coil system (40 cm diam, Neurodata Instruments) that was mounted on a movable superstructure on top of the vestibular turntable. The system was linearly related to eye position throughout the monkey's oculomotor range of ±20°. The root-mean-square (rms) noise level of the system was equivalent to ±0.2° (0-7 kHz). The gain of each eye coil was calibrated independently by assuming that the AVOR gain recorded in the light in response to a 1.9-Hz, 10°/s sinusoidal head rotation had a gain of unity. Zero horizontal eye position for each eye was defined as the position of the eye during fixation of a target 8 mm lateral to the midline (half of the interpupillary distance) when the other eye was patched. These calibrations were checked periodically by recording smooth pursuit eye movements evoked by a 0.7-Hz, 20°/s moving target. The vergence angle was obtained from the difference between right and left eye positions. Negative values indicate convergence.

Visual target characteristics and behavior control

Smooth pursuit eye movements were evoked with a laser projected onto a cylindrical screen 90 cm from the monkey with a pair of mirrors attached to position controlled galvanometers. Two retractable ceiling-mounted earth-stationary visual targets were used to study viewing distance-related changes in the VOR. Each target was a red light-emitting diode (LED) affixed to the tip of a motorized telescoping rod attached to the ceiling. The far and near LEDs were 5 mm (0.2°) and 1 mm (0.5°) in diameter, respectively. The far target was positioned on the midline 152 cm from the monkey. The near target was positioned on the midline 12.25 cm from the interaural plane (10 cm from a plane passing through the meridian of the 2 eyes). A third head-stationary near target was attached to the monkey chair with a hinge that allowed it to be positioned 10 cm from the eyes. The monkey was intermittently rewarded for fixating a visual target with drops of vanilla-flavored milk. The milk rewards were typically given at a frequency of ~1/s when the horizontal and vertical position of both eyes converged within 1° of the target.

Single-unit recording techniques

The techniques used for obtaining single-unit recordings have been previously described (Chen-Huang et al. 1997; Cullen and McCrea 1993). Briefly, Epoxy-insulated tungsten microelectrodes (4-7 M impedance) were advanced into the cerebellum with a manually operated micromanipulator. A hydraulic microdrive (Trent Wells) was then used to move the microelectrode into the vestibular nuclei. The location of each electrode probe into the brain was determined with respect to a skull-mounted reference point inside the recording chamber and verified by the location of vestibular field potentials evoked after electrical stimulation of the ipsilateral vestibular nerve (0.1-ms monophasic perilymphatic cathodal pulses, 200 µA). Single isolated vestibular unit potentials were conventionally amplified and discriminated. The output pulse of a window discriminator was used to trigger the event channel of a CED 1401 data acquisition system, which stored the time of the spike at a 100-µs resolution (see Data acquisition and data analysis).

SYNAPTIC ACTIVATION OF VESTIBULAR UNITS FROM THE VESTIBULAR NERVE. For most of the units encountered, the ipsilateral vestibular nerve was electrically stimulated (0.1-ms pulses, 50-300 µA) to determine the presence and latency of synaptically evoked responses. Units whose response latency was <1.4 ms were considered to have monosynaptic inputs from the vestibular nerve. The lack of a response to vestibular nerve stimulation was not considered to be strong negative evidence for the lack of direct inputs from the vestibular nerve. The maximum currents used were not sufficient to activate all of the fibers in the vestibular nerve, and, in a few cases (particularly with those units not tested) the unit potential was too small to recognize in the presence of the vestibular field that was concomitantly evoked by the stimulus.

ANTIDROMIC IDENTIFICATION OF ASCENDING TRACT OF DEITERS NEURONS. In one monkey, a concentric bipolar electrode (platinum-irridium; 150 µm OD, 75 µm ID; tip separation 300 µm) was placed ~400 µm dorsal to the left ascending tract of Deiters (ATD) after determining the location of the left medial longitudinal fasciculus at a rostrocaudal level that was 0.5 mm caudal to the caudal pole of the trochlear nucleus. Some of the data from these units have been presented in a preliminary report (Chen-Huang and McCrea 1998a).

Single-unit recording protocol

Only isolated units whose firing behavior was related to head rotation in the horizontal semicircular canal plane (either during fixation of an earth-stationary target or during fixation of a head-stationary target) were studied. Once encountered, horizontal semicircular canal-related units were tested for their responses following electrical stimulation of the ipsilateral vestibular nerve and/or the ipsilateral ATD (see ANTIDROMIC IDENTIFICATION OF ASCENDING TRACT OF DEITERS NEURONS). The response of each unit was then recorded in the following paradigms (~30-60 s per paradigm):

	1
	AVOR during fixation of a far target (152 cm distant; 0.7 Hz, ±20°/s; 1.9 Hz, ±10°/s; 4.0 Hz, ±6°/s)
	2
	AVOR during fixation of a near target (12.25 cm distant; 0.7, 1.9, and 4.0 Hz)
	3
	Smooth pursuit of a moving target (90 cm distant; 0.7 Hz, ±20°/s)
	4
	Cancellation of the AVOR; head stationary far target (90 cm distant; 0.7 Hz)
	5
	Spontaneous eye movements in the absence of table rotation
	6
	Cancellation of the AVOR; head stationary near target (12.25 cm distant; 0.7 Hz)

Data acquisition and data analysis

DATA ACQUISITION. Eye positions, eye velocity, head velocity, and table position were acquired at 16-bit A/D resolution with the CED 1401 system at 500 Hz. Single-unit spike data were stored as event times (100-µs resolution) using the CED 1401 event channel for off-line analysis. Off-line analysis was carried out on a Macintosh microcomputer with the aid of IGOR, a data analysis and programming software (WaveMetrics).

UNIT FIRING RATE. Unit clock values were converted to firing rate values that were assigned to each 2-ms A/D bin using a time-symmetric algorithm. All of the quantitative analysis in this study was carried out using these records of unit firing rate. In some illustrations, the bin size was increased from 2 to 10 ms to improve graphic presentation; this change in bin size is noted in the figure legend.

UNIT STATIC EYE POSITION SENSITIVITY. Static eye position sensitivity was assessed from firing behavior generated during periods (500 ms to 2 s) of spontaneous fixation in the absence of a target. Multiple regression estimates of the sensitivity to horizontal and vertical eye position were computed from at least 20 stable eye positions (>100 ms after a saccade). For units that demonstrated nonlinearity in the rate-position relationship, multiple fits of firing rate were made to subranges of eye position where the relationship appeared to be linear by visual inspection. In some units whose firing rate approached zero during off-direction gaze, eye position coefficients were determined with an iterative fitting technique that used changes in residual variance to estimate the point at which the fit became nonlinear. Static eye position coefficients were used to correct records for eye position-associated signals.

ANALYSIS OF UNIT DATA OBTAINED WITH SINUSOIDAL STIMULI. All of the unit data related to smooth pursuit, VOR cancellation, and the VOR were obtained during sinusoidal target or turntable rotations. Typically, unit responses during two or more 30-s duration trials were obtained for each of the behavioral paradigms described above. Only cycles in which the positions of both eyes were within 2° of the target were included in the analysis. In the analysis of records related to the VOR, cycles were rejected if an appropriate vergence angle was not maintained during the cycle. The included cycles were concatenated, desaccaded, averaged, and fit with sinusoidal functions.

Methods for calculating viewing distance-related changes in unit and eye movement responses

The rotational responses of many vestibular units were significantly different during near target viewing (AVOR_n) than during far target viewing (AVOR_f), but the phase of these responses did not change significantly as a function of viewing distance. It was convenient to define a scalar quantity (GVD) that was an estimate of change in response gain (G) related to viewing distance (VD) of a unit, based on the difference in responses recorded during AVOR_n and AVOR_f. A ratio of a unit's VOR response gains during AVOR_n and AVOR_f (N/F ratio) was also calculated.

The vergence angles required to fixate the far and near targets were 0.2 and 9.6°, respectively. The ideal gain of the VOR required to maintain image stability during head rotation calculated from the equations described by Viirre et al. (1986) was 1.02 for the far target and 1.22 for the near target. Thus the change in VOR gain required by kinematic considerations was 0.20. In previous behavioral studies we found that the AVOR gain change was inversely related to the frequency of turntable rotation, and even at low stimulus frequencies it had a variable gain that rarely matches the kinematic requirement (Chen-Huang and McCrea 1998b). In the present study, the change in AVOR gain recorded concomitantly with unit recordings varied from 0.0 to 0.25. The mean gain change was 0.17 ± 0.02 (mean ± SE) during 0.7-Hz table rotations, 0.09 ± 0.01 during 1.9-Hz rotations, and 0.03 ± 0.01 during 4.0-Hz rotations (Fig. 1B). These values were slightly smaller than those recorded in the absence of single-unit recordings. Because the target was located on the midline, very little change in the phase of the eye movements was predicted, or, on average, recorded (Fig. 1C). Variances in both eye movement and unit responses reported in figures, text, and tables are means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Location and classification of vestibular units

The effects of viewing distance on the responses of 105 vestibular nucleus neurons to sinusoidal angular head rotation in the plane of the horizontal semicircular canals were recorded in two squirrel monkeys. Every unit encountered was classified on the basis of its firing behavior during 1) sinusoidal rotation while the monkey fixated an earth-stationary LED target, 2) VOR cancellation evoked by fixation of a head-stationary target projected onto a cylindrical screen, and 3) sinusoidal smooth pursuit of a target projected onto the screen. Units whose firing rate was not related to horizontal rotation or horizontal eye movements during pursuit were discarded. Most of the units tested (84/105) were activated at monosynaptic (77/84) or disynaptic (1/84) latencies following electrical stimulation of the ipsilateral vestibular nerve (Fig. 2, A and B).

View larger version (31K): [in this window] [in a new window]   Fig. 2. Examples of the synaptically evoked responses of a position-vestibular-pause (PVP) unit (A) and an eye-head-velocity (EHV) unit (B) that were activated at monosynaptic latencies following electrical stimulation (arrows) of the vestibular nerve. The illustrated responses were evoked with currents that were slightly above threshold for evoking spikes. Note the triphasic field potentials evoked at 0.5-ms latency, which correspond to vestibular nerve action potentials, and the slow negative synaptic field potential recorded at 0.7- to 1.3-ms latency. The presence and relative amplitude of the negative field potentials aided assessment of the location of the recording microelectrode in the vestibular nuclei. C: map of electrode tracks in one animal in a horizontal section through the vestibular nuclei in a plane pitched nose upward 22° from the stereotaxic plane. Tracks on both sides of the brain are superimposed on the same map. Solid circles indicate tracks in which monosynaptic field was larger than 0.4 mV. Arrowheads indicate probes in which the units illustrated in A and B were recorded. Starred tracks indicate probes in which units included in this study were recorded. VN, vestibular nuclei; PH, nucleus prepositus hypoglossi; ABD, abducens nucleus; N.IV, trochlear nucleus. Dashed line A1 indicates stereotaxic anterior plane +1.0 mm rostral to the interaural plane.

Most of the units in this study were located in the rostral half of the vestibular nuclei. The estimated location of the tracks in one monkey is illustrated in a diagrammatic horizontal section through the vestibular nuclei in Fig. 2C. The diagram is a dorsal view of the vestibular nuclei constructed in a plane orthogonal to the orientation of electrode tracks. Reconstruction of electrode tracks suggested that neurons in the medial, superior, and lateral vestibular nucleus were included in our sample. The majority of horizontal canal-related secondary vestibular neurons were encountered in the ventral lateral vestibular nucleus and adjacent regions of the rostral medial vestibular nucleus. Four units recorded on one track were included in this study even though the reconstruction suggested they were located in the rostral prepositus nucleus [starred track at rostral end of nucleus prepositus hypoglossi (PH) in Fig. 2C]. These burst-position units were included because the synaptic field potential recorded in that track was large and because adjacent nonhorizontal canal-related units were activated at monosynaptic latencies following stimulation of the vestibular nerve.

Horizontal canal-related vestibular neurons were classified on the basis of their firing behavior during the VOR, VOR cancellation, smooth pursuit eye movements, fixation, and spontaneous saccadic eye movements. The classification scheme and nomenclature used here are similar to that previously described for segregating the responses of primate vestibular neurons (Cullen and McCrea 1993; Miles 1974; Scudder and Fuchs 1992; Tomlinson and Robinson 1984). These categories include the following:

	1
	Type I vestibular only (VI) units (n = 10)
	2
	Burst-position (BP) units (n = 10)
	3
	PVP units (n = 47)
	4
	EHV units (n = 22)
	5
	Position-vestibular (PV) units (n = 6)
6	Other eye movement-related vestibular units (n = 10)

All of the VI, PVP, and PV units, and the majority of the EHV units exhibited type I (ipsilateral on direction) rotational responses during fixation of earth-stationary targets. Most eye movement-related BP units, some EHV units, and most of the unclassified units exhibited type II responses (see below). The predominance of PVP and EHV units in this sample was due to the fact that a concerted attempt was made to record from regions of the vestibular nuclei that contained horizontal canal-related secondary vestibular neurons that project to the abducens nucleus or the medial rectus subdivision of the oculomotor nucleus. Six EHV units and one PVP unit included in this study were antidromically activated following electrical stimulation of the ascending tract of Deiters. Some of the firing characteristics of these cells have been described in a preliminary communication (Chen-Huang and McCrea 1998a). In the following paragraphs the characteristic firing behavior of each class of units will be described.

Vestibular-only units

The firing behavior of a typical VI unit during smooth pursuit, VOR cancellation and during the VOR evoked while fixating the far target (AVOR_f) and while fixating the near target (AVOR_n) is illustrated in Fig. 3. In each paradigm the stimulus frequency was 0.7 Hz and the peak velocity was 20°/s. Figure 3, A1-D1, shows sample records obtained from the cell. Figure 3, A2-D2, shows the averaged, desaccaded responses of the same cell.

View larger version (37K): [in this window] [in a new window]   Fig. 3. Firing behavior of a VI unit. A1: smooth pursuit eye movements. B1: VOR cancellation. C1 and D1: during the VOR-evoked during far (AVORf) and near target viewing (AVORn), respectively. The stimulus frequency was 0.7 Hz, 20°/s in each paradigm. Sample records of the responses of a typical VI unit are shown in A1-D1. Averaged responses recorded from that cell during ~1 min of recording are illustrated in A2-D2. Green dashed traces represent target position in A, and dark blue dashed traces are table velocity in B-D. Solid red lines are horizontal eye position. Eye or head movements toward the recording side are illustrated as downward traces. In D, vergence angle (dotted green line) and right eye center position (light blue hatched line) are also shown. The bottom histograms in A-D are of instantaneous unit firing rate (10 ms/bin). The calibration in A1 applies to the target position trace and all eye position traces in the 1st column. The calibration in A2 applies to the target position trace and all eye position traces in the 2nd column. Unit firing rate and time calibrations in D1 and D2 also apply to the other unit responses in that column.

The firing characteristics of VI units are summarized in Table 1. All of the VI units tested (8/8) were activated at monosynaptic latencies following stimulation of the vestibular nerve and were sensitive to ipsilateral head velocity during sinusoidal turntable rotation (mean gain = 0.96 ± 0.16 spikes/s/deg/s at 0.7 Hz). Their head velocity sensitivity was slightly higher during VOR cancellation (mean gain = 1.08 ± 0.15 spikes/s/deg/s). VI units were not sensitive to eye position and saccades and were not modulated during sinusoidal smooth pursuit eye movements.

The rotational responses of most VI units were not significantly affected by viewing distance. One VI unit's response gain decreased from 1.4 spikes/s/deg/s to 1.1 spikes/s/deg/s during near target viewing (22% decrease in gain). This exceptional unit caused the average rotational responses of VI units to be slightly lower during near target viewing (Table 2).

Burst position units

Burst-position units were usually encountered in probes through the medial aspect of the rostral medial vestibular nucleus. They are included in this study because most of them were located in the vestibular nucleus and because they presumably could be involved in producing the AVOR. BP units were sensitive to horizontal eye position, usually (9/10) in the ipsilateral direction, and generated bursts of spikes during on-direction saccades. Only the contralateral eye movement-related BP unit could be monosynaptically activated following electrical stimulation of the ipsilateral vestibular nerve. Sample records and averaged responses of an ipsilateral BP unit are illustrated in Fig. 4. The firing characteristics of ipsilateral on direction BP units are summarized in Table 1.

View larger version (35K): [in this window] [in a new window]   Fig. 4. Firing behavior of an ipsilateral on-direction burst-position (BP) unit. Sample records are shown in A1-D1. Averaged, desaccaded responses are illustrated in A2-D2. Calibration and labeling conventions of different traces are the same as in Fig. 3. Note that the background firing rate in D (when the mean position of the contralateral eye was ipsilateral to center position) was slightly higher than in C (when the mean position of both eyes was near the center), which suggests that this unit was primarily related to the position of the contralateral eye.

BP units were sensitive to eye position during steady fixation (static eye position sensitivity, K_s = 2.6-32; median K_s = 11.1 spikes/s/deg; Fig. 5A) and to eye velocity during pursuit. In six units the background firing rate during near and far target viewing was significantly different, which presumably reflected the preference of the units to the position of one eye. The eye position signals appeared to be related primarily to the contralateral eye in three units and to the ipsilateral eye in three other units. In the remaining four units, the difference in background firing rate was too small, or inconsistent, to allow an estimate of ocular dominance. BP units were typically more sensitive to on-direction saccades than to off-direction saccades. The unit illustrated in Fig. 4, A1-D1, for example, generated bursts of spikes during leftward, on-direction saccades and quick phases of nystagmus, but was weakly, and inconsistently inhibited during off-direction saccades.

View larger version (19K): [in this window] [in a new window]   Fig. 5. Static eye position sensitivity of vestibular units. Polar plots of the directional sensitivity of the four classes of vestibular unit whose firing rate was related to eye position during steady fixation. Each point is an estimate of a unit's vertical and horizontal eye position sensitivity obtained from multiple regression analysis. In many EHV units, the eye position vector was determined only from the part of the oculomotor range in which a linear estimate could be made. No symbols are attached to units whose eye dominance could not be determined. Solid symbols were units whose eye movement sensitivity was related more to the contralateral eye; open symbols were units more related to the ipsilateral eye. The few EHV units whose eye position direction reversed at the center position are not illustrated.

BP units were strongly modulated during sinusoidal smooth pursuit eye movements (Fig. 4, A1 and A2). Their mean sensitivity re eye velocity was 3.40 ± 0.55 spikes/s/deg/s; n = 10). During VOR cancellation (Fig. 4, B1 and B2) the rotational response of these cells was related primarily to the small, unsuppressed eye velocity.

The rotational responses of most (9/10) BP units increased as a function of inverse viewing distance (e.g., GVD ranged from 0.3 to 1.4 spikes/s/deg/s at 1.9 Hz). One unit exhibited a significant decrease in rotational gain during AVOR_n (GVD = 0.5 spikes/s/deg/s). On average, BP unit sensitivity to head velocity increased from 2.82 sp/s/deg/s during far target viewing to 3.68 spikes/s/deg/s during near target viewing at 1.9 Hz, whereas response phase was not significantly affected by viewing distance. A large increase in the mean BP unit response gain was also observed during 0.7-Hz rotation. These rotational responses are summarized in Table 2. The mean BP GVD, of 0.71 and 0.86 spikes/s/deg/s at 0.7 and 1.9 Hz, corresponded to an average increase in modulation of 23%. The 23% increase in modulation was larger than 17% increase in VOR gain recorded concomitantly, and the average increase in modulation (16 spikes/s) was larger than the increase that would be predicted from their sensitivity to pursuit eye velocity and static eye position (12 spikes/s).

Position-vestibular-pause units

The effects of viewing distance on the AVOR responses were recorded for 47 PVP units that were sensitive to ipsilateral head rotation in the plane of the horizontal semicircular canal. One unit was antidromically activated following electrical stimulation of the ipsilateral ATD, and thus presumably projected to the medial rectus subdivision of the ipsilateral oculomotor nucleus. Sample records and averaged responses of one PVP unit are illustrated in the left and middle columns of Fig. 6. The gain and phase of the responses of each PVP unit during 0.7-Hz smooth pursuit eye movements, VOR cancellation, and the VOR are illustrated in the polar plots in Fig. 6, A3-D3. The firing characteristics of PVP units are summarized in Table 1.

View larger version (41K): [in this window] [in a new window]   Fig. 6. Firing behavior of a PVP unit. Sample records of the responses of a typical PVP unit are shown in A1-D1. Averaged, desaccaded responses are illustrated in A2-D2. Calibration and labeling conventions of different traces are the same as in Fig. 3. The polar plots in A3-D3 show the gain and phase of the responses of all of the PVP units tested at 0.7 Hz in each paradigm. Dashed line in each plot is in phase with ipsilateral eye velocity (E') in A3 and ipsilateral head velocity (H') in B3. Each point represents the response gain and phase of a PVP unit during the experimental paradigm (+90 corresponds to a response in phase with ipsilateral acceleration, and 90 to a response in phase with ipsilateral position). Gain of 4 spikes/s/deg/s is indicated by the dashed arcs in each plot.

Most of the PVPs tested (38/40) were activated at a monosynaptic latency following electrical stimulation of the ipsilateral vestibular nerve (Fig. 2A). During steady fixation in the absence of a target, the tonic firing rate of these units was related to eye position (range 0.4-10.6 spikes/s/deg; mean K_s = 4.1 ± 0.4 spikes/s/deg). The spatial tuning of this static position signal calculated from multiple regression analysis varied considerably from cell to cell (see Fig. 5B) but was always contralateral. On the basis of their background firing rate during near and far target viewing, the eye position sensitivity of most PVPs (23/25) appeared to be better related to the position of the contralateral eye than to the ipsilateral eye (Fig. 5B). PVPs were either inhibited or stopped firing altogether during saccades; although the inhibition was typically weak or nonexistent when the saccade was in the unit's on-direction. A few (6) PVP units generated bursts of spikes during on-direction saccades.

PVP units tended to be more sensitive to head rotation during the VOR than to eye movements alone. The modulation in their firing rate during horizontal smooth pursuit of a moving visual target (0.7 Hz, 20°/s) was 1.52 ± 0.15 spikes/s/deg/s re eye velocity (Fig. 6, A1-A3). This pursuit response phase-lagged contralateral eye velocity by an average of 40.2 ± 5.6°. When the monkey fixated a far, earth-stationary target, rotation of the turntable at the same velocity and frequency evoked a response whose gain was 1.71 ± 0.14 spikes/s/deg/s re ipsilateral head velocity and contralateral eye velocity during the VOR. The PVP unit responses phase-lagged head and eye velocity by 13 and 12°, respectively. During VOR cancellation (Fig. 6, B1-B3) the gain of PVP rotational responses decreased by ~50% (mean gain re head velocity = 0.88 ± 0.09 spikes/s/deg/s; see Table 1).

The rotational responses of most (34/47) PVP units increased as a function of viewing distance (Fig. 7), although in several of those units (9/34) the increase was small and statistically insignificant. On average, PVP head velocity sensitivity increased from 1.73 spikes/s/deg/s during far target viewing to 2.04 spikes/s/deg/s during near viewing at 0.7 Hz (see Table 2 and Fig. 7). The average PVP GVD was 0.28 ± 0.08 spikes/s/deg/s; a 20% change in gain that was comparable with the concomitantly recorded VOR gain change of 17%. This small unit gain change was statistically significant (P < 0.01; t-test for paired observations) in spite of the fact that the viewing distance-related changes in response gain of individual PVP units varied considerably (Fig. 7, B and C). Thus the change in PVP unit population response was appropriate for generating the viewing distance-related changes in the AVOR that was concomitantly recorded, even though a significant fraction of units of this type were insensitive to viewing distance.

View larger version (25K): [in this window] [in a new window]   Fig. 7. Viewing distance-related changes in the sensitivity of PVP units to head rotation. A: PVP unit near target AVOR gain (AVORn) vs. far target AVOR gain (AVORf). Solid line indicates equal response gains during AVORn and AVORf. Dashed lines indicate near responses that deviated by >2 SE from the average AVORf response. B: distribution of GVD in PVP units. Positive values are increases in rotational gain during near target viewing. C: distribution of N/F response ratio in PVP units (mean = 1.20).

Eye-head-velocity units

The effects of viewing distance on the AVOR responses of 22 EHV units that were sensitive to head rotation in the plane of the horizontal semicircular canal were recorded. Six of these units were antidromically activated following electrical stimulation of the ipsilateral ATD. Sample records and averaged responses of one EHV unit are illustrated in the left and middle columns of Fig. 8. The gain and phase of the responses of each EHV unit during 0.7-Hz smooth pursuit eye movements, VOR cancellation, and the VOR are illustrated in the polar plots in Fig. 8, A3-D3. The firing characteristics of EHV units are summarized in Tables 1 and 2. During VOR cancellation, 10 EHV units increased their firing rate during ipsilateral head rotation (EHVI units, + in Fig. 8, A3-D3), and 12 units increased their firing rate during contralateral rotations (EHVII units, square symbols in Fig. 8, A3-D3). Most of the EHVI and EHVII cells that were tested were activated at monosynaptic latencies following electrical stimulation of the ipsilateral vestibular nerve (16/17 tested; Fig. 2B).

View larger version (39K): [in this window] [in a new window]   Fig. 8. Firing behavior of EHV units. Sample records of the responses of an EHVII unit are shown in A1-D1. Averaged, desaccaded responses are illustrated in A2-D2. Calibration and labeling conventions of different traces are the same as in Fig. 3. The polar plots in A3-D3 show the gain and phase of the responses of all of the EHV units tested at 0.7 Hz in each paradigm. Open squares are EHVII units, crosses are EHVI units. A gain of 2 spikes/s/deg/s is indicated by the circle in each plot. Conventions for illustration are the same as Fig. 6.

Most EHV units were sensitive to eye position during steady fixation, but the preferred on direction and ocular preference were idiosyncratic for each cell (Fig. 5C). Several EHVI units were more sensitive to vertical eye position and vertical smooth pursuit eye movements than to horizontal eye movements and position. EHVI units tended to be weakly related to ipsilateral conjugate horizontal eye position (mean gain = 1.1 ± 0.5 spikes/deg), whereas EHVII units were related to contralateral conjugate horizontal eye position (mean gain = 1.8 ± 1.0 spikes/deg). The ocular preference of 3 EHV units was for the ipsilateral eye, and for 10 other EHV units the contralateral eye. The tonic firing rate of EHV units that projected into the ATD, and presumably to the ipsilateral medial rectus subdivision of the oculomotor nucleus, was not necessarily related to the ipsilateral eye. The ocular preference of two ATD units was for the ipsilateral eye, and four others were related primarily to the contralateral eye (Chen-Huang and McCrea 1998a). EHV units also exhibited idiosyncratic responses during saccades. Most units (13/22) were either inhibited or stopped firing altogether during saccades in the oculomotor off-direction. Four EHV units generated bursts of spikes during on-direction saccades, and the firing rate of six others was not related to saccades in any direction. EHV units tended to be more sensitive to smooth pursuit eye movements (mean gain = 1.94 spikes/deg/s eye velocity) than to head velocity during the VOR (mean gain = 1.35 ± 0.18). The head movement on direction of this class of cells reversed during VOR cancellation, although the gain re head velocity during cancellation was, on average, less than half that recorded during the VOR evoked by fixation of an earth-stationary target (mean gain = 0.60 ± 0.08).

Two peculiar features of many EHV units were that their sensitivity to static eye position was frequently nonlinear, and that the eye movement-related signals exhibited during smooth pursuit eye movements were often different from the signals generated during steady fixation. Figure 9 illustrates an analysis of the firing behavior of three ATD EHV units in the left vestibular nucleus during steady fixation (top panels) and during smooth pursuit eye movements (bottom panels). Note that the modulated discharge of all three units led contralateral eye velocity during 0.7-Hz ocular pursuit, but the signals generated by each cell during steady fixation (between spontaneous saccades) were idiosyncratic. The firing rate of the cell in Fig. 9A was correlated with ipsiversive eye position, the cell in B was related to contraversive eye position, whereas the cell in C was related to both ipsi- and contraversive eye position. Similar observations have been reported for secondary vestibular units that receive inputs from the ipsilateral flocculus (Lisberger et al. 1994).

View larger version (26K): [in this window] [in a new window]   Fig. 9. Static and dynamic eye movement signals generated by 3 different EHVII units that projected into the ATD. Top graph in each column shows the relationship between firing rate and eye position recorded during steady fixation. Bottom panels show averaged, desaccaded responses of the corresponding units during smooth pursuit eye movements. Solid line is an eye position record, and the filled histogram is unit firing rate. Ipsilateral eye position is expressed as negative values in rate-position plots. Downward deflections in the smooth pursuit records are ipsilateral. During smooth pursuit, each unit's response led the contralateral eye velocity. However, the static eye position sensitivity was ipsilateral for the unit illustrated in A, contralateral for the unit in B, and bilateral for the unit in C.

The effects of viewing distance on the AVOR responses of EHV units were not uniform. The response gain of most (13/22) EHV units increased during near target viewing (Figs. 8 and 10, Table 2). However, near viewing had no significant effect on the AVOR response of five units, produced a decrease in gain in three units, and a reversal of response phase in the response of another. On average, the GVD of EHVII units was 0.75 ± 0.24 spikes/s/deg/s at 0.7 Hz, which corresponded to a gain increase of 38%. EHVI units tended to be less sensitive to head velocity during the VOR, and consequently their GVD was smaller on average (0.41 ± 0.20 spikes/s/deg/s at 0.7 Hz). However, the average increase in rotational gain of EHVI units was 49%. On the whole, the viewing distance-related change in firing rate was significantly larger for EHV units than for PVPs (t-test, P < 0.01).

View larger version (22K): [in this window] [in a new window]   Fig. 10. Viewing distance-related changes in the sensitivity of EHV units to head rotation at 0.7 Hz. A: EHV unit AVORn gain vs. AVORf gain. Solid line indicates equal response gains during AVORn and AVORf. Dashed lines indicate near responses that deviated by >2 SE from the average AVORf response. B: frequency histogram of EHV unit GVD. C: frequency histogram of EHV N/F ratio (mean = 1.44).

Viewing distance significantly affected the response phase of most EHV units. In most cases the near response significantly phase lagged the response recorded during far target viewing. At 0.7 Hz, the AVOR_n response phase lagged the AVOR_f response by 34 ± 6° in the nine EHVII units that exhibited a significant GVD increase. The response phase of EHVI units was more mixed. The AVOR_n responses of most of these units also lagged the AVOR_f responses. However, the AVOR_n response of four EHVI units phase led their AVOR_f responses by 11-54°. All of those EHVI units were sensitive to vertical eye movements and had AVOR_f responses that lagged contralateral head velocity by >70°.

In sum, viewing distance-related changes in the AVOR response gain of EHV units were, on average, larger than the change in response recorded from any other class of vestibular unit, and more than twice as large as the concomitant changes in eye movements.

Position vestibular units

PV units were encountered in the same region where PVP units were found. Most of the PV units (5/6) were activated at a monosynaptic latency following electrical stimulation of the ipsilateral vestibular nerve. The firing rate of these cells was correlated with static eye position but was unaffected during saccadic eye movements. The responses of PV units during smooth pursuit, VOR cancellation, and the far or near viewing VOR are summarized in Tables 1 and 2. In general, the smooth pursuit eye movement sensitivity of PV units was much lower than PVP and EHV units. The gain and phase of the rotational responses of PV units were similar to PVPs when the animal was fixating an earth-stationary target or was canceling its VOR (Table 1). On average, PV units were more sensitive to changes in viewing distance than PVPs. Their mean GVD was 0.38 ± 0.53 (spikes/s/deg/s), and their N/F ratio was 1.41 ± 0.25 during 0.7-Hz rotation. In sum, PV units were not as sensitive to eye movements as EHV and PVP units, but the sensitivity of their rotational responses to viewing distance was quite high.

Other vestibular units

Ten units were encountered in the vestibular nuclei that could not be readily classified in one of the five cell classes described above. Nine of the units were activated at monosynaptic latencies following electrical stimulation of the vestibular nerve. Six of the cells were eye movement-related type II neurons (EVII units), whose firing rate was modulated in phase with contralateral head velocity during rotation and with ipsilateral eye velocity during smooth pursuit. The rotational responses of EVII units were also sensitive to viewing distance. During 1.9-Hz table rotations, where the result for every unit was available, the mean GVD increased 24 ± 19%. The responses of EVII units are summarized in Tables 1 and 2.

Two monosynaptically activated units were related to contralateral smooth pursuit eye velocity but not to static eye position. One monosynaptically activated unit did not respond to any type of eye movements but showed large differences between VOR cancellation and AVOR_f. One other unit paused during saccades but exhibited no other signals related to eye movements. The rotational responses of the pausing neuron and one of the two pursuit units increased during near viewing. The rotational responses of the other two units were unaffected by viewing distance.

Effects of stimulus frequency on AVOR GVD

Most of the results described above were obtained with a sinusoidal vestibular stimulus (0.7 Hz, 20°/s) that matched the stimulus frequencies used to evoke smooth pursuit eye movements and to evaluate the responses of units during VOR cancellation. The AVOR responses were also recorded during higher frequencies of vestibular stimulation in many (87/105) units. Figure 11 illustrates examples of the averaged AVOR responses of a PVP unit and an EHVII unit evoked at three rotation frequencies (0.7 Hz, 20°/s; 1.9 Hz, 10°/s, and/or 4.0 Hz, 6°/s).

View larger version (55K): [in this window] [in a new window]   Fig. 11. AVORf and AVORn responses of PVP (A1-D1) and EHVII (A2-D2) units at different stimulus frequencies. Averaged, desaccaded records of example unit responses at different frequencies of turntable rotation are illustrated in A-C. Top traces in A-C are averaged eye velocity during far target viewing (dashed light green traces) and near target viewing (solid blue traces). The responses recorded during AVORf (thick green lines with light shading) are superimposed on responses recorded during AVORn (dark blue shaded histograms). All traces are aligned with the start of contraversive sinusoidal table velocity; upward is contralateral. D1 and D2: mean gain (solid lines, left axis) and phase (dotted lines, right axis) of PVP (D1) and EHVII (D2) responses during AVORf (green open symbols) and AVORn (blue solid symbols). Note that AVORn responses consistently phase lagged AVORf responses and the difference in response gain during AVORn and AVORf progressively decreased as a function of frequency.

Graphic summaries of the responses of 13 PVP and 7 EHV units that were tested at all 3 frequencies of rotation are illustrated in Fig. 11, D1 and D2. The average response gain and phase of both classes of cells tended to increase slightly during far target viewing as stimulus frequency increased, although there was considerable individual variability within each class. During near target viewing the average PVP unit gain decreased slightly from 2.04 ± 0.18 spikes/s/deg/s at 0.7 Hz to 1.53 ± 0.24 spikes/s/deg/s at 4 Hz, whereas the average EHVII unit gains were nearly the same at each stimulus frequency. Viewing distance had larger, significant effects on the average response gain of the 32 vestibular units tested at 0.7 Hz, but it did not affect the average response gain of these vestibular neurons recorded at 4 Hz, which corresponded to the changes in AVOR gain as a function of viewing distance at these frequencies.

The AVOR near target responses of both PVP and EHVII units slightly phase lagged their responses during far target viewing at every frequency (Fig. 11, D1 and D2). The phase lag in unit response produced by a reduction in viewing distance was remarkably constant at different frequencies.

Firing behavior of vestibular units during voluntary cancellation of AVOR_n

The responses of 16 vestibular units, including 5 EHVII and 11 PVP units, were recorded during cancellation of a head-stationary target located on the midline 10 cm from the monkey's eyes. The firing behavior of vestibular neurons during 0.7-Hz VOR cancellation in the presence of near (VORC_n) and far (VORC_f) head-stationary targets was compared, and the effects of viewing distance on rotational gains were assessed.

Cancellation of the VOR during fixation of a head-stationary near target (mean VOR gain = 0.07 ± 0.01 re head velocity) was equal or better than VOR cancellation during fixation of a head stationary far target (mean gain = 0.14 ± 0.01). EHV units, as noted above, characteristically reversed the direction of their response during VOR cancellation, whereas PVP units characteristically exhibited a reduction in their rotational response. The response of one unit during cancellation of AVOR_n and AVOR_f is illustrated in Fig. 12A. The filled, shaded histogram is the averaged response recorded during fixation of an earth-stationary near target, whereas the superimposed thin and thick traces show the averaged responses recorded during fixation of head-stationary far and near targets, respectively. The unit's modulation during VORC_n and VORC_f was similar, regardless of the distance of the target, although there was a small decrease in the background firing rate during near target viewing that was related to the more medial deviation of the contralateral eye during VORC_n.

View larger version (18K): [in this window] [in a new window]   Fig. 12. Vestibular unit responses during VOR cancellation evoked by fixation of near (VORCn) and far (VORCf) head-stationary targets. A: responses of a PVP unit during fixation of an earth-stationary near target (AVORn, shaded histogram), VOR cancellation during fixation of a 90-cm distant target (VORCf) and during fixation of a near, 10-cm distant target (VORCn). All traces are aligned with the start of contraversive sinusoidal table velocity. Note that the gain of the unit decreased dramatically in both the VORCf and VORCn conditions. B: VORCn plotted as a function of VORCf. C: change in PVP and EHV rotational gain related to viewing distance during VOR (AVOR GVD) plotted vs. VORc GVD.

In each of the PVP and EHV units tested, the response recorded during VORC_n was comparable with the response recorded during VORC_f cancellation (Fig. 12B). During VORC_n the average gain of PVP responses was 0.8 ± 0.2 spikes/s/deg/s compared with 1.0 ± 0.2 spikes/s/deg/s during VORC_f. The mean response phase led head velocity by 45° during VORC_n and 17° during VORC_f. The small differences in response gain recorded during AVOR_n cancellation and AVOR_f cancellation (VORC GVD; Fig. 12C) were not correlated with viewing distance sensitivity. The comparability of the single-unit responses during VOR cancellation might be interpreted to be a reflection of the fact that the GVD signals on secondary VOR neurons were related to their eye movement sensitivity. An alternative possibility is that the inputs that reduce or modify the rotational responses of PVP and EHV units during VORC are also modified as a function of viewing distance. The comparison of unit viewing distance sensitivity and eye movement sensitivity in the next section suggests that the latter interpretation is more likely.

Relationship of AVOR GVD to pursuit eye velocity sensitivity and static eye position sensitivity

The viewing distance-related changes in AVOR responses of central vestibular neurons were not well correlated with their sensitivity to eye velocity during smooth pursuit. The correlation coefficient of the linear relationship between smooth pursuit eye velocity sensitivity and AVOR GVD was 0.51 for PVP units and 0.05 for EHVII units. On the other hand, unit sensitivity to viewing distance was better correlated with their static eye position sensitivity. The correlation between eye position sensitivity and the increase in unit modulation during near target viewing was 0.76 for EHVI units, +0.85 for EHVII units, and +0.58 for PVP units (Fig. 13).

View larger version (24K): [in this window] [in a new window]   Fig. 13. Vestibular unit change in firing rate related to viewing distance vs. static eye position sensitivity (Ks). Values on both the ordinate and abscissa are expressed in terms of unit sensitivity to eye position (spikes/s/deg). Red solid line is the regression related to EHVII units (red squares), and the dashed line is the regression related to PVP units (green squares). The viewing distance sensitivity of EHVII units was strongly correlated with their sensitivity to eye position (R = 0.82), although the slope of the regression (3.4) was >3 times that expected if eye position signals alone had caused the change in unit gain during AVORn (thin black dotted line). The relationship between viewing distance sensitivity and eye position sensitivity was weaker for PVP units (R = 0.46), but the slope of the regression was similar.

The positive correlation between AVOR GVD and static eye position sensitivity was not readily attributable to the increase in the amplitude of eye movements during near viewing per se. The slopes of the regressions illustrated in Fig. 13 suggest that PVP and EHVII units tended to be three times more sensitive to the increased amplitude of eye movements during near viewing than their static eye position coefficients would predict. Subtraction of static eye position signals from vestibular unit rotational responses affected both their gain and phase, particularly at the 0.7-Hz stimulus frequency, but on average this removal had little effect on the increase in response gain recorded during near viewing. The effects of subtraction of eye position signals on the average responses of representative BP, PVP, and EHVII units during 1.9-Hz rotation are illustrated in Fig. 14, A-C. The mean eye position-corrected 1.9-Hz AVOR_n responses of BP, PVP, and EHVII units are summarized in the polar plots at the bottom of Fig. 14 (see also Table 2). Subtraction of static eye position signals produced a phase lead in both BP and PVP unit responses (Fig. 14, A and B), but the gain change produced by changing viewing distance was not significantly affected. Subtraction of static eye position signals had little effect on the gain or phase of EHVII unit rotational responses during near and far target viewing (Fig. 14C). Subtraction of static eye position signals also did not account for viewing distance-related unit responses during 0.7-Hz rotation (Table 2), although the eye position changes evoked at that frequency were larger.

View larger version (60K): [in this window] [in a new window]   Fig. 14. Eye position corrected and uncorrected rotational responses of BP (A), PVP (B), and EHVII (C) units. An example of each type of cell is illustrated. Desaccaded, averaged unit records obtained during near target viewing (blue filled histograms) are superimposed on records obtained during far target viewing (green shaded histograms). The bottom 2nd set of unit records was corrected for static eye position sensitivity. The mean responses of each class of units are plotted in polar coordinates at the bottom of the figure. Closed and open symbols represent population responses that were corrected and uncorrected for static eye position, respectively. Near target responses are blue circles; far target responses are green squares. Note that subtraction of eye position signals introduces a phase lead in every case but does not abolish viewing distance-related signals for any of the 3 classes of cells.

In sum, static eye position sensitivity and viewing distance sensitivity were correlated, but the increase in rotational gain of eye movement-related central vestibular units during near target viewing was not directly attributable to eye position and eye velocity oculomotor signals they carry.

Relationship between vestibular nerve signals and viewing distance related unit and eye movement signals

In a previous study we showed that viewing distance-related changes in the AVOR were significantly reduced when irregular afferents were selectively silenced by bilateral application of galvanic anodal currents (Chen-Huang and McCrea 1998b), whereas the AVOR_f was unaffected. We suggested that the central AVOR GVD signals might reflect the addition of an irregular afferent input to secondary AVOR pathways. Figure 15 is a polar plot of the difference between the AVOR_n and AVOR_f uncorrected 1.9-Hz responses of the three classes of neurons illustrated in Fig. 14. These phasors represent the signals added to each unit class during near target viewing. The dashed arrow corresponds to the phase of the vector difference between AVOR_n and AVOR_f eye movements. The shaded sector of the polar plot represents the response phase of squirrel monkey vestibular nerve irregular afferents at 2 Hz (Lysakowski et al. 1995). Considered together, PVP and EHV viewing distance signals lagged the mean vestibular nerve irregular afferent signals by ~91°. On the other hand the viewing distance signals generated by BP units had nearly the same phase as the viewing distance-related eye movements.

View larger version (22K): [in this window] [in a new window]   Fig. 15. Polar plot of the viewing distance-related rotational responses of different classes of vestibular neurons and the concomitant change in eye movements at 1.9 Hz. The vector difference in the response of PVP, EHV, and BP units recorded during near and far target viewing is plotted as phasors P, E, and B, respectively. PVP and EHV viewing distance-related signals phase lag ipsilateral head velocity, and lag the signals generated by ipsilateral irregular afferents by ~90°. EHV units are approximately twice as sensitive to viewing distance as PVP units. Viewing distance-related signals carried by most BP units phase lag ipsilateral eye velocity. The phasor sum of the viewing distance-related response of all 3 classes of cells at 1.9 Hz is similar in phase to the vector difference in eye movements evoked by rotation during near and far target viewing (Eye), if it is assumed that the signals carried by PVPs or EHVs are anatomically inverted while ipsilateral BP units are not. These observations suggest that the viewing distance changes in central vestibular neurons are sufficient to produce the changes in eye movement observed.

In Fig. 15, the phasor sum of the viewing distance signals related to BP, PVP, and EHV units (labeled B, P, and E, respectively) was 134° when the signals related to BP and PVP units were inverted. The inversion could occur anatomically in the case of units that project to the contralateral abducens nucleus or ipsilateral medial rectus subdivision of the oculomotor nucleus. The sum of the inverted EHV and PVP signals with BP signals had nearly the same phase as the viewing distance-related eye movement at 1.9 Hz. This suggests that inputs from these three groups of units to the oculomotor plant are sufficient to produce AVOR_n at 1.9 Hz.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

T