phase relations of purkinje cells in the rabbit flocculus

JOURNALOFNEUROPHYSIOLOGY Vol. 74, No. 5, November 1995. Printed in U.S.A.

Phase Relations of Purkinje Cells in the Rabbit Flocculus During Compensatory Eye Movements

C. I. DE ZEEUW, D. R. WYLIE, J. S. STAHL, AND J. I. SIMPSON Department of Physiology and Neuroscience, New York University Medical Center, New York 1OOM; and Department of Anatomy, Erasmus University Rotterdam, 3000 DR Rotterdam, Postbus 1738, The Netherlands

SUMMARY AND CONCLUSIONS

1. Purkinje cells in the rabbit flocculus that respond best to rotation about the vertical axis (VA) project to flocculus-receiving neurons (FRNs) in the medial vestibular nucleus. During sinusoidal rotation, the phase of FRNs leads that of medial vestibular nucleus neurons not receiving floccular inhibition (non-FRNs) . If the FRN phase lead is produced by signals from the ~~OCCU~US, then the Purkinje cells should functionally lead the FRNs. In the present study we recorded from VA Purkinje cells in the flocculi of awake, pigmented rabbits during compensatory eye movements to determine whether Purkinje cells have the appropriate firing rate phases to explain the phase-leading characteristics of the FRNs.

2. Awake rabbits were sinusoidally rotated about the VA in the light and the dark at 0.05-0.8 Hz with different amplitudes. The phase of the simple spike ( SS ) modulation in reference to eye and head position was calculated by determining the eye position sensitivity and the eye velocity sensitivity using multivariate linear regression and Fourier analysis. The phase of the SS modulation in reference to head position was compared with the phase of the FRN modulation, which was obtained in prior experiments with the same stimulus paradigms.

3. The SS activity of nearly all of the 88 recorded floccular VA Purkinje cells increased with contralateral head rotation. During rotation in the light, the SS modulation showed a phase lead in reference to contralateral head position that increased with increasing frequency (median 56.9” at 0.05 Hz, 78.6’ at 0.8 Hz). The SS modulation led the FRN modulation significantly at all frequencies. The difference of medians was greatest (19.2”) at 0.05 Hz and progressively decreased with increasing frequency (all Ps < 0.005, Wilcoxon rank-sum test).

4. During rotation in the dark, the SS modulation had a greater phase lead in reference to head position than in the light (median 110.3’ at 0.05 Hz, 86.6’ at 0.8 Hz). The phase of the SS modulation in the dark led that of the FRNs significantly at all frequencies (difference of medians varied from 24.2O at 0.05 Hz to 9.1” at 0.8 Hz; all Ps < 0.005).

5. The complex spike (CS) activity of all VA Purkinje cells increased with ipsilateral head rotation in the light. Fourier analysis of the cross-correlogram of the CS and SS activity showed that the phase lag of the CS modulation in reference to the SS modulation at 0.05 Hz in the light was not significantly different from that at 0.8 Hz (median 199.7’ at 0.05 Hz, 198.3O at 0.8 Hz), even though the phases of the SS modulation at these two frequencies were significantly different (P < 0.001). These data indicate that the average temporal reciprocity between CS and SS modulation is fixed across the range of frequencies used in the present study.

6. The CS activity of most Purkinje cells did not modulate during rotation in the dark. Of 124 cases (each case consisting of the CS and SS data of a VA Purkinje cell obtained at 1 particular frequency) examined over the frequency range of 0.05-0.8 Hz,

17 cases (14%) showed CS modulation. In the majority (15 of 17) of these cases, the CS activity increased with contralateral head rotation; these modulations occurred predominantly at the higher stimulus velocities.

7. On the basis of the finding that FRNs of the medial vestibular nucleus lead non-FRNs, we predicted that floccular VA Purkinje cells would in turn lead FRNs. This prediction is confirmed in the present study. The data are therefore consistent with the hypothesis that the phase-leading characteristics of FRN modulation could come about by summation of VA Purkinje cell activity with that of cells whose phase would otherwise be identical to that of non- FRNs. The floccular SS output appears to increase the phase lead of the net preoculomotor signal, which is in part composed of the FRN and non-FRN signals.

8. The phase lead of the floccular output may be created by emphasizing the velocity component of mossy fiber signals origi- nating in the vestibular nuclei.

INTRODUCTION

The cerebellar flocculus is involved in the control of eye movements. In rabbits, the main oculomotor behaviors controlled by the flocculus are compensatory eye movements manifest as the vestibuloocular reflex (VOR) and the optokinetic reflex (OKR). The flocculus may contribute to the control of both the gain and phase of compensatory eye movements. The possible short- and long-term roles of the flocculus in gain control have been extensively studied (rabbit: Ito et al. 1982; Nagao 1983, 1989b; primate: Lisberger and Pavelko 1988; Lisberger et al. 1984; Takemori and Co- hen 1974; Waespe et al. 1983; Zee et al. 1981; chinchilla: Daniels et al. 1978; cat: Robinson 1976). Ito (Ito 1982; Ito et al. 1974, 1982; see also Waespe and Henn 1981) initially proposed that the flocculus augments OKR gain and medi- ates enhancement of VOR gain by vision. Lesions of the flocculus and the ventral paraflocculus can change the gain of the VOR and OKR (Daniels et al. 1978; Ito et al. 1982; Lisberger et al. 1984; Nagao 1983, 1989b; Robinson 1976; Takemori and Cohen 1974; Waespe et al. 1983; Zee et al. 198 1) , and gain changes have been correlated with changes in the simple spike (SS ) response of floccular Purkinje cells (Nagao 1988, 1989b).

The role of the flocculus in the control of the phase of compensatory eye movements is less well documented. Eye movements in flocculectomized animals lag those of normals during OKR and, to a lesser degree, during VOR (rabbit: Ito et al. 1982; Nagao 1983; cat: Robinson 1976). Nagao ( 1983) reported that optokinetic stimulation in phase with

0022-3077/95 $3.00 Copyright 0 1995 The American Physiological Society 2051

2052 C. I. DE ZEEUW, D. R. WYLIE, J. S. STAHL, AND J. I. SIMPSON

lead -

160 -

Light

JZIIl Nerve

FIG. 1. Summary of the phase relationships among different elements of the vestibuloocular reflex (VOR) circuit (From Stahl and Simpson 1992, 1995b). The flocculus-receiving neurons (FRNs) lead the non-FRNs at all frequencies in both the dark and the light. Phases (medians) are referenced to contralateral head position ( +, lead). For clarity of presentation, the phases of the medial vestibular

FRN nucleus neurons (FRNs and non-FRNs) and primary

nonFRN afferents (VIIIth nerve) have been shifted by 180”

JZI to reflect, respectively, their direct and indirect inhibitory action on the ipsilateral VIth nucleus neurons. VIth nucleus data were obtained in the light and the VIth nucleus phase values in the dark were obtained by referencing to eye position (Eye) in the dark.

Eye

0.05 0.1 0.5 1.0 0.05 0.1

Stimulus Frequency (Hz)

whole body rotation, but at twice the amplitude, induced a significant adaptive phase lead in the VOR that was abol- ished by bilateral lesion of floccular Purkinje cells with kai- nit acid. Injections of y-aminobutyric acid agonists into the rabbit flocculus, on the other hand, influenced the gain but not the phase of the VOR and OKR (Van Neerven et al. 1989). The issue of a floccular contribution to phase control was also raised by Stahl and Simpson (1992, 1995b), who demonstrated in awake, pigmented rabbits that flocculus- receiving neurons ( FRNs ) in the medial vestibular nucleus have a phase lead with respect to head position that is significantly greater than that of other oculomotor projecting neurons in the medial vestibular nucleus that do not receive an input from the flocculus (non-FRNs) . This phase difference was present throughout the tested frequency range (0.05-0.8 Hz) for sinusoidal rotation about the vertical axis (VA) in both dark and light (Fig. 1) as well as for optokinetic stimulation about the VA. This finding suggests that the phase lead of the FRNs over that of the non-FRNs could be produced by the signal contributed by floccular Purkinje cells. If so, the Purkinje cells should functionally lead the FRNs. The FRNs in the rabbit medial vestibular nucleus are innervated by those floccular Purkinje cells whose climbing fibers respond best to optokinetic stimulation about the VA, but not by those Purkinje cells whose climbing fibers respond best to stimulation about a horizontal axis (De Zeeuw et al. 1994). Therefore, the proposed phase-leading Purkinje cells should all be of the former (VA) type. To determine whether these Purkinje cells lead the FRNs, the phase of the SS modulation of floccular VA Purkinje cells was measured in awake, head-fixed, pigmented rabbits during sinusoidal head rotation in both the light and the dark. These phase measure- ments were compared with those previously obtained for FRNs under the same stimulus conditions (Stahl and Simp- son 1992, 1995b).

METHODS

Animal preparation

Seven pigmented rabbits were prepared for chronic recording with the use of sterile surgical techniques. General anesthesia was induced with a combination of ketamine (32 mg/kg im), acepromazine (0.32 mg/ kg im) , and xylazine ( 5.0 mg/ kg om) , and sup- plemental doses (9 mg/kg ketamine, 0.09 mg/kg acepromazine, 2 mg/kg xylazine) were given every 30-45 min. An acrylic head fixation pedestal was formed and fixed to the skull by screws implanted in the calvarium. The pedestal was oriented so that the animal’s nasal bone made an angle of 57’ to the earth’s horizontal plane. With this orientation, the responses of the horizontal semicir- cular canals during rotation about the VA are close to their theoreti- cal optimum, whereas the responses of the anterior and posterior canals are essentially zero (Soodak and Simpson 1988). A craniotomy was made over the left paramedian lobule of the cerebellum, and a metal recording chamber to allow introduction of microelectrodes into the flocculus was fixed around the craniotomy by ex- tending the acrylic head fixation pedestal. This cylindrical chamber was oriented so that its axis was in a sagittal plane and made an angle of 27O to the vertical. The brain was covered by a silastic sheet and the chamber was closed by a screw top.

A search coil was implanted on the left eye to measure eye position. The coil, made of three turns of Teflon-coated stainless steel wire (Cooner #AS632), was wound parallel to the limbus around the sclera under the superior and inferior recti and the inferior oblique muscles. The leads of the eye coil terminated in a plug fixed to the pedestal.

Stimulation

After a recovery period of 1 wk, recordings were made during vestibular rotation. The animal was enclosed by a black drum decorated with randomly placed white graffiti; the drum extended from the zenith to 2%30’ below the horizon. Vestibular stimulation was provided by a servo-controlled turntable whose axis was vertical and passed through the midpoint of the interaural line.

PHASE DYNAMICS OF FLOCCULAR PURKINJE CELLS 2053

Animals were rotated sinusoidally at different frequencies with different amplitudes (0.05 Hz, 25.0”; 0.1 Hz, k5.0”; 0.2 Hz, 22.5”; 0.4 Hz, 52.5”; and 0.8 Hz , 2 1.3”) in light and dark. As pointed out by Stahl and Simpson ( 1995a,b), the selected stimulus amplitudes represent a compromise between the need to maximize slow-phase eye movement amplitude and the need to minimize fast phases. Because rabbits rarely maintain eye positions eccentric by more than a few degrees, the stimulus amplitudes were of necessity small.

Recording procedures

Table position, horizontal eye position, and Purkinje cell activity were recorded simultaneously. Table position was measured with a precision potentiometer coupled to the table shaft. Eye position was monitored using the implanted search coil (Robinson 1963). The animal was mounted with the implanted eye at the center of a magnetic field that alternated in spatial and temporal quadrature at 32 kHz. Calibration of the eye coil was achieved by rotating the field coils about the center of the eye while the animal maintained a stationary eye position.

Purkinje cell activity was recorded extracellularly with the use of glass microelectrodes filled with 2 M NaCl (2-4 MQ). The electrodes were driven by a hydraulic micropositioner through the paramedian lobule into the flocculus. Neuronal activity was filtered with a notch filter at the frequency of the magnetic fields used by the search coil system (32 kHz) and then amplified and filtered with a bandpass of 10 Hz-10 kHz. The activity from single Pur- kinje cells was identified by the presence of a pause in SS activity after the complex spike (CS ) (Eccles et al. 1966; Thach 1967). VA Purkinje cells were identified by determining the optimal axis for the CS response to visual world rotation produced by a planetarium projector. The planetarium rotated about the VA, which is the optimal axis for VA Purkinje cells, or about a horizontal axis oriented at 45’ contralateral azimuth, 135’ ipsilateral azimuth, which is the optimal axis for modulation of the other floccular Purkinje cells whose CSs respond to retinal image slip (for details see De Zeeuw et al. 1994; Graf et al. 1988; Simpson et al. 1988; Soodak and Simpson 1988; Van Der Steen et al. 1994). The planetarium rotated at a constant angular speed of lo/s, which is close to the optimal speed for modulating the floccular visual climbing fibers (Barmack and Hess 1980; Kusunoki et al. 1990; Simpson and Alley 1974). The direction of the planetarium rotation reversed every 5 s.

The SS and CS were discriminated with the use of window discriminators. SS and CS activity and eye and table position were recorded on-line with the use of a CED 1401 signal capture device and the Spike2 (Cambridge Electronics Design) program on a personal computer. The modulation of neuronal activity was pro- cessed as nonaveraged data or it was averaged and presented as a peristimulus histogram (for example see Fig. 2) compiled over a number of cycles (usually 10). In addition, all data were stored on tape. Recording sessions ran up to 4 h; they were terminated if the animals showed signs of agitation. Between recordings the brain was covered by a silastic sheet and the chamber was sealed.

Data analysis

To determine the phase and magnitude sensitivity, the SS response was analyzed by calculating apparent eye position and eye velocity sensitivities using linear multivariate regression. This approach has been used for abducens neurons and vestibular nucleus neurons, and the rationale has been discussed extensively (Stahl 1992; Stahl and Simpson 1995a,b). In brief, for each VA Purkinje cell up to 10 continuous cycles of data were analyzed for each stimulus paradigm. Cvcles during which saccades occurred were

6 1

CS PSTH

SS PSTH

2.8

v n 0

-2.8

5

1

Head Position

5 10 15 20

lime (set)

FIG. 2. Peristimulus time histograms (PSTHs) of the complex spike (CS ) and simple spike (SS ) responses of a representative vertical axis (VA) Purkinje cell in the rabbit flocculus to sinusoidal rotation at 0.05 Hz in the light. The total number of spikes per bin ( lOO-ms binwidth) was accumulated over 8 cycles. The SS modulation amplitude was 50% (amplitude: 29.3 spikes/s). Note the temporal reciprocity of the CS and SS modulations (see also Fig. 10). The eye position trace has been shifted 180” to facilitate comparison with head position.

not used for analysis. The data files collected in Spike2 were trans- ferred as ASCII files and analyzed with programs written in ASYST (Stahl 1992). Values for apparent eye position sensitivity (k) and apparent eye velocity sensitivity (r) were calculated by regressing firing rate (F) against eye position (E) and eye velocity (E’ )

F(t) = kE(t) + rE’(t) + c

The regression calculates the coefficients k. r. and c to minimize

2054 C. I. DE ZEEUW, D. R. WYLIE, J

k=-8.5spikesosec1/deg . tz Y sf

l W . L 40-

LL 30-

--- -2 4 0 1 2 3

Eye Position (c&g)

20 40 60 80 100 120 140 160

Time (set)

LL I 1 -

. .

-0.5 0 0.5 1

Eye Velocity (deg*sei?)

FIG. 3. Multivariate linear regression analysis of the SS modulation shown in Fig. 2. Actual and fitted firing rates are indicated in A by the dots and the solid line, respectively. B and C: partial regression plots illustrating how the eye position sensitivity (k) and eye velocity sensitivity (r) were obtained. In B, the eye velocity component ( rE’ ) is subtracted from the firing rate and the result is plotted against eye position; in C, the eye position component (kE) is subtracted from the firing rate and the result is plotted against eye velocity.

the sum of the squared residuals, which are the differences between the actual F(t) and F(t) calculated from the model equation (e.g., Fig. 3). The firing rate was compensated for drift (see Stahl 1992) and the significance of the fit curves was statistically evaluated; curves with P > 0.01 were discarded (3% of all analyzed cases). The firing rate phase in reference to eye position was obtained at each stimulus frequency (f ) from the relation

TABLE 1. Eye movement gains and phases

. S. STAHL, AND J. I. SIMPSON

phase = arctan [ 27$( r/k)]

This firing rate phase was converted to phase in reference to contralateral head position by adding the phase difference between eye position and head position; the phase difference between eye position and head position was determined by Fourier analysis (Stahl and Simpson 1995a,b). Phase lead was taken as positive and phase lag as negative.

The magnitude sensitivity, which is the ratio of the amplitudes of SS modulation and eye position modulation, was calculated from the relation

magnitude sensitivity = sqrt [ ( r27rf ) 2 + k2 ]

RESULTS

Eye movements

The gain and phase of the compensatory eye movement responses were comparable with those found for rabbits by Stahl (1992; see also Baarsma and Collewijn 1974; Nagao 1991) . In the light the average gain ranged from 0.63 to 0.8, depending on the stimulus frequency, whereas in the dark the gain ranged from 0.25 to 0.31 (Table 1) . The phase of the eye position in reference to head position varied with frequency (Table 1; see also Fig. 1) . In the light this variation was modest, from an average lead of 5.3’ at 0.05 Hz to a lag of 2.5O at 0.8 Hz. In the dark, the phase lead of the eye position decreased more substantially as frequency increased, from an average of 39.3O at 0.05 Hz to 7.8” at 0.8 Hz.

SS responses of VA Purkinje cells

The SS activity of nearly all 88 identified VA Purkinje cells increased with contralateral (in reference to the recording side) head rotation in the light and the dark. Only one VA Purkinje cell increased its SS activity with ipsilateral head rotation at all stimulus frequencies; this neuron was not included in the tables, figures, and comparisons with the FRNs presented below. The SS activity of another VA Purkinje cell increased with ipsilateral rotation at 0.05 Hz, but with contralateral rotation at higher stimulus frequencies; this cell was included in the analysis. Cells that were excited by ipsilateral rotation were frequently encountered, but they could not be identified as VA Purkinje cells. Either they did not show a pause in SS activity after the CS (i.e., the recordings were either from inter-neurons or from several si-

0.05 Hz 0.1 Hz 0.2 Hz 0.4 Hz 0.8 Hz

Light n 35 29 43 35 43 Gain 0.67 5 0.02 0.75 t 0.03 0.8 t 0.03 0.72 5 0.03 0.63 + 0.03 Phase 5.3 5 0.37 3.1 + 0.54 1.6 5 0.38 0.6 5 0.4 -2.5 + 0.7

Dark n 24 23 31 33 26 Gain 0.25 + 0.02 0.27 + 0.02 0.27 5 0.03 0.31 5 0.03 0.29 5 0.02 Phase 39.3 t 1.7 13.8 5 1.5 5.2 5 0.8 6.5 5 1.1 7.8 t 1.1

Gain and phase values are averages t SE; positive values indicate phase lead. Peak stimulus velocities are 1.6, 3.1, 3.1, 6.3, and 6.3”/s for 0.05, 0.1, 0.2, 0.4, and 0.8 Hz, respectively. Tabulation of eye movement performance during rotation in the light and the dark at different frequencies. Phase of eye position in reference to head position (“) was obtained from Fourier analysis. In the convention used, an eye movement in perfect phase compensation for head movement would have a phase of 0”. n, number of cells recorded at each frequency in the light and dark.

PHASE DYNAMICS OF FLOCCULAR PURIUNJE CELLS 2055

TABLE 2. Amplitude, r/k, phase, and magnitude sensitivity of SS response

0.5 Hz 0.1 Hz 0.2 Hz 0.4 Hz 0.8 Hz

Light n Average DC, spikes/s Average amplitude/DC, % Average rlk, s Average phase, deg Median phase, deg Average magnitude sensitivity, spikes l s-’ l deg-’

Dark n Average DC, spikes/s Average amplitude/DC, % Average rlk, s Average phase, deg Median phase, deg Average magnitude sensitivity, spikes l s-’ l deg-’

35 29 55.1 2 4.5 51.4 5 5.6 45.3 t 2.5 46.4 t 3.3

4.99 t 0.26 3.19 t 0.15 62.7 !I 4.7 66.6 5 5.4

56.9 57.6 7.6 + 0.7 6.7 5 1.1

24 61 of: 4.6 20.3 + 1.6

9.45 5 0.28 111 k 5.6

110.3 10.9 5 0.8

23 63 5 6.1 33.6 + 2

6.82 + 0.1 90.7 IT 3.8

87.9 15 5 1.5

43 35 49.1 5 3.8 48.9 5 4.3 42.9 _+ 2.1 57.7 5 3.7

2.52 5 0.06 1.57 t 0.03 74 2 4.5 76.4 + 3.8

63.7 72.4 13.1 5 1.4 20.9 2 2.8

31 33 59.2 2 5.0 60.3 2 4.7 32.5 + 1.7 48.7 + 2.2

4.46 5 0.03 3.47 2 0.02 85.2 5 2.3 89.9 + 2.7

88.8 89.6 31.8 t 3.0 51.4 + 5.8

43 62 + 4.8 53.9 + 2.7

1.43 + 0.01 79.6 + 2.4

78.6 83.5 Ir 8.7

26 67.7 + 5.7 46.9 t 2.7

1.16 t 0.01 88.1 + 2.1

86.6 136 + 16.1

Values, except for y1 and median phase values, are averages + SE. Tabulation of parameters of Purkinje cell simple spike (SS) responses to rotation in the light and the dark at different frequencies. r/k values are obtained from average phase values. Phase is stated in reference to head position. Magnitude sensitivity is ratio of the amplitudes of firing rate and eye position modulation. r, apparent eye velocity sensitivity; k, apparent eye position sensitivity; DC, mean discharge rate.

multaneously recorded Purkinje cells) or their CS activity was optimally modulated by visual world rotation about the ipsilateral 135” axis in the horizontal plane (Graf et al. 1988).

Because it took -30 min to complete all paradigms for one cell, not all cells were tested with all paradigms. Re- sponses to rotation in the light were obtained from a total of 81 cells, whereas responses in the dark were obtained from 43 cells. The number of VA Purkinje cells analyzed for each paradigm varied from 23 to 43. In the light, data were obtained at all stimulus frequencies from 17 cells, whereas in the dark, data were obtained at all frequencies from 16 cells; We were able to collect data at all frequencies in both light and dark from 8 cells.

SS modulation amplitude

For the purpose of comparison with previously reported recordings of rabbit floccular Purkinje cells (Nagao 1988- 199 1) , we determined the percent modulation amplitude [ (amplitude/mean discharge rate) X 100% ] using Fourier

A 65

60 T k 55

?J 50

5 45 E

a 40 g 35 Q iI 25 30

s 20

B

analysis. The mean discharge rate of the VA Purkinje cell SS response was greater during rotation in the dark than in the light at all frequencies (Table 2). Both in the light and the dark, the percent modulation amplitude (Fig. 4A) was highest at 0.4 Hz (light: 57.7%; dark: 48.7%). At each frequency the percent modulation amplitude was higher in the light than in the dark (P < 0.05, t-test). The average percent modulation amplitude over the different frequencies was close to linearly related to both maximum eye velocity (correlation coefficients for the dark and light were 0.98 and 0.87, respectively; Fig. 4B) and maximum head velocity (correlation coefficients for the dark and light were 0.98 and 0.90, respectively; Fig. 4C).

SS phase in the light

The phase of the SS modulation was calculated from the eye position and eye velocity sensitivities (for comparable r and k values during 0.1 Hz, constant-speed optokinetic stimulation see Leonard 1986). During rotation in the light, the average r/k progressively decreased from 4.99 s at 0.05

0.1 1 0 1 2 3 4 5 Stimulus Frequency (Hz) Maximum Eye Velocity (degkec)

C Light

Dark

; 1 i ; ; ; f Maximum Head Velocity (degkc)

FIG. 4. Average amplitude of modulation ( % ) as a function of stimulus frequency (A ) , maximum eye velocity (B) , and maximum head velocity (C) in the dark and the light. Error bars: rt 1 SE.

2056

A c. I. DE ZEEUW, D. R. WYLIE, J. S. STAHL, AND J. I. SIbfPSON

Dark B Light lo- lo-

9 \

8- \ 8- \ b

6-

0 P-cell

l FRN

o nonFRN

C 120-

Ei 2i g loo- .- -5

5 80- tz I E 60- 0 8 ; 40- <D

it

v. I


Dark

u. I


nonFRN

Light

P-cell FRN

20: ’ “ “ I I 1 , , , , l , f 201 ““‘I r 1 1 I , , , ,

0.1 1 0.1 1 Stimulus Frequency (Hz) Stimulus Frequency (Hz)

FIG. 5. Variation of r/k (A and B) and phase ( C and D) for SS modulation of Purkinje cells as a function of stimulus frequency in the dark (A and C) and the light (B and D). A and B: average r/k of the Purkinje cells is larger than that of both the FRNs and non-FRNs at all frequencies in both the dark and the light. C and D: median phase lead of the SS response was significantly larger than that of the FRNs and non-FRNs at all frequencies in both the dark and the light. For clarity of presentation, the phase of the medial vestibular nucleus neurons was shifted 180°, as described in Fig. 1.

Hz to 1.43 s at 0.8 Hz (Table 2; Fig. 5). The average r/k values of the Purkinje cells were higher than those of the FRNs and non-FRNs in the medial vestibular nucleus.

The median phase lead of the Purkinje cell SS modulation in reference to contralateral head position increased progressively from 56.9’ at 0.05 Hz to 78.6” at 0.8 Hz (Fig. 50). The median SS phase lead was significantly greater than that of the FRN modulation at all frequencies (Figs. 5 D and 6; for FRN data see Table 3). The difference was greatest at 0.05 Hz (difference of medians = 19.2”, P < 0.001; Wil- coxon rank-sum test) and progressively decreased with increasing frequency ( 14.7” at 0.1 Hz, P < 0.001; 11.8’ at 0.2 Hz, P < 0.001; 8.8’ at 0.4 Hz, P < 0.005; 6.8’ at 0.8 Hz, P < 0.005).

SS phase in the dark

During rotation in the dark, the average ,rlk values were larger (P < 0.01, t-test) than those in the light at all fre-

quencies except at 0.8 Hz (compare Fig. 5, A and B). The average r/k progressively decreased from 9.45 s at 0.05 Hz to 1.16 s at 0.8 Hz (Table 2). As during rotation in the light, all Purkinje cell r/k values were greater than those of the FRNs and non-FRNs (Fig. 5A). The phase lead of the SS modulation in reference to head position was greatest at 0.05 Hz and flattened over the higher frequencies (Fig. 5C). Figure 7 shows the SS phase for all individual Purkinje cells recorded during rotation at 0.05 and 0.8 Hz. The SS phase lead was greater during rotation in the dark than in the light at all the tested frequencies, a difference that reflects the fact that the primary vestibular afferents lead head velocity in the tested frequency range (see Fig. 1) . The difference between the phase of the SS modulation of the VA Purkinje cells and the phase of the FRNs varied from 24.2” at 0.05 Hz to 9.1” at 0.8 Hz (Fig. 5C). As during rotation in the light, the Purkinje cells led the FRNs significantly at all frequencies during rotation in the dark (P < 0.001 at 0.05 Hz; P < 0.001 at 0.1 Hz; P < 0.001


Eye Velocity

P-cells FRN-MPN

IPN

Eye Posit ion

4 spikesesec?deg \x \ \

‘1 FIG. 6. Polar plot of phase and magnitude sensitivity of Purkinje cells

( P-cells), FRNs, non-FRNs, and abducens nucleus neurons for 0.2-Hz rotation in the light ( 22.5”). Note that the Purkinje cells lead all the other neuronal types. The FRNs are further classified according to whether they do or do not project to the ipsilateral oculomotor complex in the midbrain [MPN: midbrain projecting neuron as identified with the use of antidromic stimulation (see Stahl and Simpson 199Sb) J . The FRN-MPNs presumably have an excitatory projection to the ipsilateral medial rectus motoneurons, whereas the FRN-nonMPNs presumably have an inhibitory projection to the ipsilateral abducens nucleus. The FRN-MPNs and the FRN-nonMPNs do not differ significantly with regard to phase or magnitude sensitivity, and elsewhere in the present study they are collectively referred to as FRNs. Orientation of sector gives the average firing rate phase in reference to ipsilateral eye position. Length of sector indicates the average magnitude sensitivity (width of sector: ?2 SE).

at 0.2 Hz; P < 0.001 at 0.4 Hz; P < 0.005 at 0.8 Hz; Wilcoxon rank-sum test),

SS magnitude sensit-ivity I

The magnitude sensitivity of the SS response (see METH- ODS) increased with increasing frequency during rotation in both the light and the dark (Fig. 8). In the light, the magnitude sensitivity increased from 7.6 spikes l s -’ l deg -’ at 0.05 Hz to 83.5 spikes l s-’ l deg -I at 0.8 Hz. In the dark, the

magnitude sensitivity increased from 10.9 spikes l s -I 4 deg ’ at 0.05 Hz to 136 spikes l s-l 4 deg -’ at 0.8 Hz. At each frequency the magnitude sensitivity was higher in the dark than in the light (P < 0.05, f-test). The difference may be explained, in part, by an amplitude nonlinearity similar to that observed in abducens and vestibular nucleus neurons in rabbits by Stahl and Simpson ( 1995b). Stahl and Simpson found that magnitude sensitivity increased as eye movement amplitude decreased at a fixed stimulus frequency. In the present experiments, smaller eye movement amplitudes were produced in the dark, so this nonlinear effect would contribute to the separation of the dark and light curves in Fig. 8.

Contribution of visunl signals to SS phase I

The Purkinje cell modulation during rotation in the light reflects synergism between vestibular and visual inputs, In Fig. 9, the phase contribution of the visual system is demonstrated by plotting the difference in VA Purkinje cell phase for rotation in light and dark over the 0.05- to O&Hz frequency range. At all frequencies, the visual system contri- butes a phase lag, which is greatest at lower frequencies. Figure 9 also indicates similar data for rabbit medial vestibular nucleus FRNs and non-FRNs (from Stahl and Simpson 1995b). At all frequencies, the light versus dark phase difference is greatest in VA Purkinje cells, intermediate in FRNs, and least in non-FRNs. This hierarchy is expected if VA Purkinje cells are an important conduit through which the visual signals pass.

CS response in the light

In the flocculus, the CS and SS modulations are generally temporally reciprocal in the presence of vision (Graf et al. 1988; Kano et al. 199 1 b; Nagao 1989a; 199 I), but previous studies only examined one frequency and usually the temporal reciprocity was not quantified. To evaluate this reciprocal relation under different stimulus paradigms, the phase of the CS modulation in reference to the SS modulation during rotation in the light was determined at both 0.05 and 0.8 Hz, which are the frequencies for which the phases of the SS modulation had their greatest significant difference (P < 0.001, Wilcoxon rank-sum test). If only those cells for which the data at both stimulus frequencies were available (n = 19) were included, the average phase differ-

TABLE 3, r/k und phase of FRiV response

0.05 Hz 0.1 Hz 0.2 Hz 0.4 Hz 0.8 Hz

Light n Average r/k, s Average phase, deg Median phase, deg

Dark n Average rlk, s Average phase, deg Median phase, deg

26 26 26 26 26 2.4 t 0.15 1 .s 5 0.08 I. 1 % 0.07 0.9 k 0.07 0.9 -+ 0.17

38.4 5 1 .s 43.2 -+ 1.4 51.7 2 1.7 63.2 f- 1.6 70.2 f 1.4 37.7 42.9 51.9 63.6 7 I .8

22 22 22 22 22 3.1 % 0.26 2.2 k 0.27 1.53 2 0.2 1.22 2 0.3 0.7 -+ 0.16

83.3 -+ 4.4 71.8 i 2.4 69.9 2 2.1 71.8 2 2.3 79.2 t 1.9 x5.1 69.7 69 73 77.5

Values, except for n and median phase values, are averages 2 SE. FRN, flocculus-receiving neuron; for other abbreviations see Table 2. Tabulation of parameters of FRN response to rotation in the light and dark at different frequencies (from Stahl 1992).


Dark

180’ D

I Position \

V

Contra ( 0’ Head

i

Position

\I L-l 270’ FIG. 7. Polar plots of the SS responses of individual VA Purkinje cells during rotation in the dark (left) and in the light

(right) at the lowest (0.05 Hz) and highest (0.8 Hz) frequencies used in this study. The plots show SS modulation amplitude (amplitude/mean discharge rate x 100%) and SS phase in reference to contralateral head position; the modulation amplitude of a particular cell corresponds to the distance from the dot to the center of the plot, whereas the phase is indicated by the angle with the horizontal axis. Note that the SS modulations of virtually all Purkinje cells lie in the top 2 quadrants, signifying

Light

that firing rate increased during contralateral head rotation.

ence between the SS phase lead at 0.05 and 0.8 Hz was 26.9’. The phase of the CS modulation in reference to the SS modulation was estimated with the use of a fast Fourier transform of the CS-SS cross-correlogram (Fig. 10; for details on cross-correlations see Gerstein and Kiang 1960). The phase lag of the CS modulation in reference to the SS modulation at 0.05 Hz was not significantly different from that obtained for the same cells at 0.8 Hz (median of 199.7O at 0.05 Hz; 198.3’ at 0.8 Hz: average of 193.2 t lo”, mean t SE, at 0.05 Hz; 194.9 t 6.8’ at 0.8 Hz). Thus the temporally reciprocal relation between the CS and SS modulation at the lowest stimulus frequency was not differ-

~160- aI

7; 140- iii

l 120- ii Y ‘5 loo- - c 80- .- .- .cI i! 60- ii

t 40-

-E 20- z) z 0 II III I I 1 I,,,,

0.1 1 Stimulus Frequency (Hz)

FIG. 8. Variation of magnitude sensitivity of the SS responses of Pur- kinje cells with frequency. Note that the magnitude sensitivity increased with increasing frequency in both the light and the dark, and that at all frequencies the average magnitude sensitivity was higher in the dark than in the light. Error bars: tl SE.

ent from that at the highest frequency even though the ent from that at the highest frequency even though the phases of the SS modulations were significantly different phases of the SS modulations were significantly different at the two frequencies. at the two frequencies.

C’S response in the dark C’S response in the dark

Although the CS activity of the large majority of Purkinje cells was not modulated in the dark, some instances of modulation were observed. To quantitate the presence of CS mod-

Although the CS activity of the large majority of Purkinje cells was not modulated in the dark, some instances of modulation were observed. To quantitate the presence of CS mod-

O- O- S? S? z -lO- z -lO- t;i t;i n n I I -2o- -2o- E E P, P, ; -3o- ; -3o-

tii tii g -4o- g -4o- q nonFRN a a n FRN

nonFRN FRN

Q -5()- 0 P-cell a’ d

-60 I III, I I I IIIII 0.1 1 Stimulus Frequency (Hz)

FIG. 9. Contribution of vision to neuronal phase. Firing rate phase for VOR in the dark was subtracted from the phase for VOR in the light and plotted vs. stimulus frequency. The presence of vision had large effects on the phase of all 3 cell groups, particularly at the lowest frequencies. How- ever, the Purkinje cells were most affected, indicating that the stronger visual signal of the FRNs (in comparison with the signal of non-FRNs) is probably attributable to the VA Purkinje cell input to the FRNs. All phases are in reference to contralateral head position. Data for FRNs and non- FRNs are derived from Stahl and Simpson (1995b).


A 8000

6000

Full Cycle Crosscorrelogram

"0.0 2.5 I 5.0 715 10.0 ' 12.5 ' 15.0 17.5 I 20.0

B I I 400 ms Crosscorrelogram

-0.10 -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30

Time(sec)

FIG. 10. Example of a CS-SS cross-correlation over the full stimulus cycle (A, 625ms bins) and over 400 ms (B, I-ms bins) for a VA Purkinje cell during turntable rotation at 0.05 Hz in the light. In both correlograms, time 0 indicates the occurence of CSs. The PSTH for the same set of cycles is shown in Fig. 2. To determine the temporal reciprocal relationship between the CS and SS modulation, the full-cycle cross-correlogram was fast Fourier transformed and the fundamental component was used to determine the phase difference between the CS and SS modulations. In this case, the phase difference was 202”; the difference corresponds approximately to the time between time 0 and the peak divided by the time of the full cycle multiplied by 360”. Note that the pause in B lasts - 13 ms; the pause is not resolved in the full-cycle correlogram because a larger binwidth was used.

ulation in the dark, the CS activity was considered to be modulated when the number of spikes during rotation in one direction was at least twice the number of spikes during rotation in the opposite direction. Of 124 cases (each case consisting of the CS and SS data obtained at 1 particular frequency) evaluated throughout the tested frequency range (n = 23 at 0.05 Hz; n = 21 at 0.1 Hz; n = 3 1 at 0.2 Hz; n = 27 at 0.4 Hz; n = 22 at 0.8 Hz), 17 cases ( 14%) showed CS modulation. Of those 17 cases, 15 ( 88%) increased their CS activity with contralateral head rotation so that the CS modulation was approximately in phase with the SS modulation. Figure 11 shows the response of one of these cells. CS modulation in the dark occurred predominantly at the higher frequencies (n = 0 at 0.05 Hz; n = 1 at 0.1 Hz; n = 5 at 0.2 Hz; n = 6 at 0.4 Hz; n = 3 at 0.8 Hz). The 15 cases for which the CS activity increased with contralateral rotation were obtained from 11 of the 43 Purkinje cells tested in the dark. From 6 of these 11 cells we also recorded the CS activity for one or more frequencies during rotation in the light; all these cells increased their CS activity with ipsilateral head rotation in the light.

DISCUSSION

In the present study we demonstrate that the SS responses of floccular Purkinje cells in awake rabbits are activated in a

consistent way during horizontal vestibular stimulation and that these VA Purkinje cells lead the neurons that they innervate in the medial vestibular nuclei. In addition, we demonstrate that the CS modulation of VA Purkinje cells is generally temporally reciprocal to the SS modulation in the presence of vision, and that the CS activity of a small proportion of these cells is modulated during vestibular stimulation in the dark.

A I CS PSTH

0.0 0.5 1.0 1.5 2.0 2.5

I SS PSTH

0.0 0.5 1.0 1.5 2.0 2.5

0.0 0.5 1.0 1.5 2.0 2.5

0.0 0.5 1.0 1.5 2.0 2.5

Time (set)

FIG. 11. PSTHs of a Purkinje cell with CS modulation during rotation in the dark. Stimulus frequency was 0.4 Hz and binwidth for both CS and SS was 100 ms. Total number of spikes per bin was accumulated over 8 cycles. Note that the SS and CS modulations are approximately in phase and that both increase when the head rotates to the contralateral side. The eye position trace has been shifted 180” to facilitate comparison with head position.


SS modulation

The SS activity of nearly all floccular VA Purkinje cells increased with contralateral head rotation in the light and the dark. Of 88 VA Purkinje cells recorded at random from the floccular VA zones, the SS activity of only two cells increased with ipsilateral rotation at one or more of the tested frequencies. These data agree with those of Leonard ( 1986), who reported that >93% of the VA Purkinje cells in the flocculus increased their SS activity for ipsilaterally directed eye movements. Because only the floccular VA Purkinje cells and not the Purkinje cells involved in the “vertical” components of compensatory eye movements project to the medial vestibular nucleus (De Zeeuw et al. 1994), the present results are also in line with the observation that all FRNs in the rabbit medial vestibular nucleus are excited during rotation of the head toward the ipsilateral side and movement of the eyes toward the contralateral side (Kawaguchi 1985; Stahl 1992; Stahl and Simpson 1995b).

Several previous studies have reported much higher percentages of floccular Purkinje cells that showed increased SS activity during ipsilateral head rotation. Ghelarducci et al. ( 1975 ) found in rabbits that 49% (42 of 86) of floccular Purkinje cells increased their SS activity during ipsilateral head rotation in the dark, ’ whereas Miles et al. ( 1980) observed in the monkey that 6 of 21 “eye-movement-only” Purkinje cells in the [ventral para] flocculus were excited for contralateral eye movements. These latter cells may be equivalent to floccular Purkinje cells of rabbits (Stahl 1992; see also below). In other studies of the rabbit flocculus (Dufosse et al. 1978; Nagao 1989- 1991), the percentage of purported VA Purkinje cells whose SS activity increased during rotation of the head toward the ipsilateral side in the dark varied from 18% to 24%.2 The discrepancy between the present and previous results probably cannot be explained by our missing a type of Purkinje cell clustered in specific locations, because the “VA inphase and VA outphase Pur- kinje cells” have been reported to be randomly distributed within the VA floccular zones (Nagao 199 1) . A possible explanation may be found in the procedure of selecting the Purkinje cells. In contrast to the present study, those by Ghelarducci et al. (1975) and Miles et al. (1980) made no distinction between Purkinje cells whose preferred axis of CS modulation was the VA and those whose preferred axis was in the horizontal plane (HA cells). The studies by Du- fosse et al. ( 1978) and Nagao ( 199 1) claimed to identify the VA Purkinje cells by the presence of horizontal eye movements evoked by microstimulation at the recording site. However, because the threshold of myelinated fibers is usually lower than the threshold of a cell body (Ranck 1975)) these investigators may have evoked in a number of cases horizontal eye movements by stimulating passing VA axons, even though the recorded Purkinje cell was not a VA Pur- kinje cell. Because the percentage of HA Purkinje cells in the studies by Dufosse et al. and Nagao that showed an increase of SS activity during ipsilateral head rotation was

’ This percentage was calculated from Fig. 1 D of Ghelarducci et al. (1975).

* These percentages were calculated from the polar diagram figures in Dufossk et al. (1978) and Nagao (1989-1991).

substantially higher than that of their VA cells (63% vs. 18- 24%; Dufosse et al. 1978; Nagao 1989ab, 1991), it is quite possible that some of the cells they identified as VA Purkinje cells excited by ipsilateral head rotation were HA cells. In the present study, the percentage of VA Purkinje cells excited by ipsilateral head rotation was not overestimated because the VA Purkinje cells were identified by determining the optimal axis for CS modulation to visual world rotation. In fact, when we observed Purkinje cells whose SS activity increased with ipsilateral head rotation, determination of the optimal axis for the CS modulation often revealed that these cells were HA Purkinje cells. Similarly, most of the Purkinje cells in the ventral paraflocculus of the monkey that increased their SS activity for contralateral eye movement actually had their dominant response during vertical, not horizontal, pursuit (Lisberger et al. 1994; see also Stone and Lisberger 1990).

SS phase

The major difference between FRNs and non-FRNs in the medial vestibular nucleus of the rabbit is that FRNs possess a greater phase lead in reference to eye position (Stahl 1992; Stahl and Simpson 1992, 1995b). The difference is present during VOR in the light and the dark over the entire tested frequency range of 0.05-0.8 Hz, and is also present during OKR (only tested at 0.1 Hz). The present recordings demonstrate that the VA Purkinje cells, which project to the medial vestibular nucleus (De Zeeuw et al. 1994), functionally lead the FRNs at all frequencies during VOR in the light and the dark, indicating that the floccular signal could produce the phase difference between FRNs and non-FRNs.

The FRNs and non-FRNs contribute to the net premotor signal, which is defined as the summed activity of all cells synapsing on extraocular motoneuron pools. If the floccular signals account for the phase advance of the FRNs and thereby the net premotor signal, one would expect that eye movements in flocculectomized rabbits would lag those of normals. In albino rabbits, Ito et al. ( 1982) found that unilateral flocculectomy produced a significant phase lag ( 13- 20”) of the ipsilateral eye during rotation in the dark (at 0.1 and 0.5 Hz) and during OKR (at 0.033 and 0.05 Hz). In pigmented rabbits, bilateral flocculectomy produced a significant phase lag during VOR in the dark at 0.16-0.5 Hz, but not at 0.1 Hz (Nagao 1983, 1989b); during OKR the phase lag increased significantly at 0.1 and 0.18 Hz. For VOR in the light at 0.01-0.5 Hz, flocculectomy in pigmented rabbits induced a phase lag of 8- 15” (Kimura et al. 199 1) . In cat, flocculectomy decreased the phase lead by 7” during VOR in the dark at 0.05 Hz, but the best significance level reached was 0.1 (t-test) (Robinson 1976; cf. Keller and Precht 1979). Thus the effect of flocculectomy on the phase of eye movements in rabbits and cats generally agrees with the proposition that the flocculus advances the phase of the net premotor signal.

In the monkey, lesions of the ventral paraflocculus and/ or flocculus do not produce consistent phase shifts of compensatory eye movements (Takemori and Cohen 1974; Zee et al. 1980, 1981). However, the flocculus and ventral paraflocculus of monkeys cannot be directly compared with the


flocculus of rabbits because the majority of recorded Pur- kinje cells in these areas of the primate cerebellum consisted of gaze velocity cells (Miles et al. 1980), which are absent in rabbits (Leonard 1986). The primate Purkinje cells comparable with those of the rabbit may be the eye-movement- only group, which made up about one fourth of the recorded population in the monkey (Miles et al. 1980). It is unknown whether these eye-movement-only Purkinje cells lead their medial vestibular targets (FTNs, Lisberger and Pavelko 1984; Lisberger et al. 1994; Miles et al. 1980), as has been demonstrated in the current study for the rabbit VA Purkinje cells and FRNs.

Phase-lagged premotor cell group

Because the FRNs in the magnocellular and parvicellular medial vestibular nuclei lead the non-FRNs, the FRN signals can be imagined to be synthesized by imposing the VA Purkinje cell activity on a neuronal group whose properties are, in the absence of the floccular input, identical to non- FRNs (Stahl and Simpson 1995b). A consequence of this synthesis would be an increase in the phase lead of the net premotor signal. The phase-leading property of the floccular signal may indicate that one role of the rabbit flocculus is to compensate for excessive integration by the brain stem neural circuitry. If the brain stem integrator network has limitations in the precision of the phase lag it produces, a separate structure may be required for fine tuning. Stahl ( 1992) proposed that the flocculus could perform this role by taking the overintegrated premotor signal from the brain stem, emphasizing the component in phase with eye velocity relative to the component in phase with eye position, and then injecting that signal into the vestibular nuclei. As such, the flocculus creates a signal that leads the original input signal. This scenario predicts the existence of a cell group producing an overintegrated signal whose phase lies between the motoneurons and eye position (see Fig. 1) . The weighted sum of the FRNs, non-FRNs, and the phase-lagged premotor cell group would be in phase with the abducens neurons. The proposed lagging cell group could be located in the prepositus hypoglossi, which projects to the ~~OCCU~US, the oculomotor nucleus, and the abducens nucleus (for review see McCrea et al. 1979; cat: McCrea and Baker 1985; McCrea et al. 1980; rabbit: Barrnack et al. 1992a,b), and which contains some cells with a phase close to eye position (cat: Escudero et al. 1992; Lopez-Barneo et al. 1982; monkey: McFarland and Fuchs 1992).

Another possible location for the phase-lagged premotor cell group is the caudal medial vestibular nucleus, which was not explored by Stahl ( 1992). This nucleus also projects to the flocculus (Kotchabhakdi and Walberg 1978; Sato et al. 1983) and the oculomotor complex (Thunnissen 1990), and it receives a prominent input from nodular (Wylie et al. 1994) but not floccular VA Purkinje cells (De Zeeuw et al. 1994). In contrast to flocculectomy, nodulectomy in rabbits advances the VOR and OKR phase (Ito et al. 1982; Nagao 1983), suggesting that the Purkinje cells in the nodulus exert a lagging effect. This organization raises the interesting pos- sibility that two cerebellar structures, the flocculus and the nodulus, both receiving their visual climbing fibers from

olivary neurons in the dorsal cap and ventrolateral outgrowth (Alley et al. 1975; Kano et al. 1991a; Takeda and Maekawa 1989; Tan et al. 1995), exert opposing effects that together finely regulate the phase of compensatory eye movements.

Impact of oculomotor plant

The relationship between abducens neuron firing rate and eye movement can be approximated by a transfer function with two poles and one zero (Stahl 1992; Stahl and Simpson 1995a). This relationship is largely dictated by the cascaded viscoelastic properties of the oculomotor plant. In such a system, the various summary measures of firing rate modulation (e.g., k, r, r/k, magnitude sensitivity, and phase) can be shown to depend on stimulus frequency in a predictable fashion (Fuchs et al. 1988; Stahl and Simspon 1988). Simi- lar relationships between the firing rate parameters and stimulus frequency would be expected for the premotor cells controlling the abducens neurons, provided that there are no significant frequency-dependent transfer function elements (such as poles or OS) between the premotor neurons and the abducens neurons. In fact, vestibular nucleus neurons with projections to motoneurons do show the same frequency- dependent variations in k, r, r/k, magnitude sensitivity, and phase seen in abducens neurons. The current results show that Purkinje cells also show these variations, indicating that the mechanics of the oculomotor plant form a major influence on the relationship between Purkinje cell SS activity and eye movement.

Contribution of visual signals to SS phase

In addition to the prominent phase difference between FRNs and non-FRNs, the FRNs may have a stronger visual signal than the non-FRNs (Stahl and Simpson 1995b). As shown in Fig. 9, the difference between phase in the light and dark is greater for FRNs than non FRNs at all frequencies, indicating that vision makes a larger contribution to FRN phase. If this difference between FRNs and non-FRNs is attributable to the VA Purkinje cell input to the FRNs, then the VA Purkinje cells should exhibit an even larger dark versus light phase difference than FRNs. This prediction is upheld (Fig. 9). The phase lag contributed by the visual system compensates for the increasing phase lead of the primary vestibular afferents with decreasing stimulus frequency, i.e., vision compensates for the low-frequency defi- cits of the primary vestibular afferents (see Fig. 1).

Because the gain of the eye movements is smaller in the dark than in the light, one cannot categorically exclude a contribution of extraocular muscle proprioception to the differences in firing rate phases in light and dark. Extraocular proprioceptive inputs to the rabbit flocculus have been demonstrated (Kimura et al. 1991; Maekawa and Kimura 1980). However, because the visual system is ultimately responsible for the difference in eye movement phase in light and dark, it is parsimonious to attribute the corresponding difference in neuronal phase to the visual system.

CS response in the light

During VOR in the light, the SS phase lead in reference to contralateral head position at 0.05 Hz was significantly

2062 C. I. DE ZEEUW, D. R. WYLIE, J

different from that at 0.8 Hz. The phase of the CS modulation in reference to the SS modulation at 0.05 Hz was not significantly different from that at 0.8 Hz, suggesting that the CS modulation of the VA Purkinje cells is temporally reciprocal to the SS modulation during rotation in the light at the frequencies used in the present study. Although this phase relation varied considerably from cell to cell, the average phase lag of CS modulation in reference to SS modulation during rotation in the light was 193.2” at 0.05 Hz and 194.9” at 0.8 Hz. Analogous results for optokinetic stimulation at 0.33 Hz ( t 1.3”) have been reported by Nagao ( 1988), who found an average phase difference of 201.6O between CS and SS modulation.

In the present study, the peak velocity of the head varied from 1.6”/s (0.05 Hz) to 6.3”/s (0.8 Hz). Considering the average gain at these frequencies (0.67 at 0.05 Hz; 0.63 at 0.8 Hz), the maximum retinal slip velocity varied from 0.53 to 2.3”/s. During OKR, the CS modulation in the rabbit flocculus is greatest for retinal image velocities of 0.2-2.0°/ s (Barmack and Hess 1980; Kusunoki et al. 1990; Simpson and Alley 1974). Therefore it is reasonable to conclude that the reciprocal temporal relation between CS and SS modulations during VOR in the light is present at least for retinal image velocities for which the CS activity is best modulated.

CS response in the dark

During rotation in the dark, a minority ( 14%) of our cases of Purkinje cell recordings showed CS modulation. In the majority (88%) of these cases, the modulation increased with contralateral head rotation, which is opposite to the CS behavior in the light. This reversed response generally occurred at the higher peak velocities of 3.1 or 6.3”/s. These findings agree with those of Ghelarducci et al. ( 1975 ) , who observed that the CS activity of floccular Purkinje cells in rabbits could be modulated during rotation in the dark (peak velocity 4.7”/s) and that most of these cells showed increased activity with contralateral head rotation. In addition, the present data agree with the observation of Nagao ( 1989a, 1991) that CS activity is not modulated during rotation in the dark at low peak velocity ( 1.6”/s).

Because the polarity of the CS modulation during rotation in the dark was generally opposite to that in the light, the CS modulation in the dark cannot be due to an incomplete darkness. This CS signal must be conveyed by an input to the caudal dorsal cap of the inferior olive, which is the source of the climbing fibers of the VA Purkinje cells of the rabbit flocculus (Alley et al. 1975; Tan et al. 1995). The most likely candidate for this input is the nucleus prepositus hypoglossi, which contains cells that project to the caudal dorsal cap (cat: McCrea and Baker 1985; rabbit and rat: Barmack et al. 1993; De Zeeuw et al. 1993) as well as cells that are modulated during horizontal VOR (cat: Escudero et al. 1992; Lopez-Barneo et al. 1982; monkey: McFarland and Fuchs 1992).

We thank P. L. DiGiorgi for technical assistance. This research was supported by grants from NW0 (R 95 - 260)) KNAW,

and the National Institute of Neurological Disorders and Stroke (NS-

. S. STAHL, AND J. I. SIMPSON

13742). D. R. Wylie was supported by a postdoctoral fellowship from the National Sciences and Engineering Research Council (Canada).

Present addresses: J. Stahl, Dept. of Neurology, University Hospitals of Cleveland, 2074 Abington Rd., Cleveland, OH; D. R. Wylie, Dept. of Psychology, University of Alberta, Edmonton, T6G QE9, Canada.

Address for reprint requests: C. I. De Zeeuw, Dept. of Anatomy, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Nether- lands.

Received 28 October 1994; accepted in final form 25 June 1995.

REFERENCES

ALLEY, K., BAKER, R., AND SIMPSON, J. I. Afferents to the vestibulo-cerebellum and the origin of the visual climbing fibers in the rabbit. Brain Res. 98: 582-589, 1975.

BAARSMA, E. A. AND COLLEWIJN, H. Vestibulo-ocular and optokinetic reac- tions to rotation and their interaction in the rabbit. J. Physiol. Land. 238: 603-625, 1974.

BARMACK, N. H., BAUGHMAN, R. Mr., AND ECKENSTEIN, F. P. Cholinergic innervation of the cerebellum of rat, rabbit, cat, and monkey as revealed by choline acetyltransferase activity and immunohistochemistry. J. Comp. Neural. 317: 233-249, 1992a.

BARMACK, N. H., BAUGHMAN, R. W., ECKENSTEIN, F. P., AND SHOJAKU, H. Secondary vestibular cholinergic projection to the cerebellum of rabbit and rat as revealed by choline acetyltransferase immunohistochernistry, retrograde and orthograde tracers. J. Comp. Neural. 317: 250-270, 1992b.

BARMACK, N. H., FAGERSON, M., AND ERRICO, P. Cholinergic projection to the dorsal cap of the inferior olive of the rat, rabbit, and monkey. J. Comp. Neural. 328: 263-281, 1993.

BARMACK, N. H. AND HESS, D. T. Multiple-unit activity evoked in the dorsal cap of inferior olive in the rabbit. J. Neurophysiol. 43: 15 1 - 163, 1980.

DANIELS, P. D., HASSUL, M., AND KIMM, J. Dynamic analysis of the vestibulo-ocular reflex in the normal and flocculectomized chinchilla. Exp. Neural. 58: 32-45, 1978.

DE ZEEUW, C. I., WENTZEL, P. R., AND MUGNAINI, E. Fine structure of the dorsal cap of the inferior olive and its GABAergic and non-GABAergic input from the nucleus prepositus hypoglossi in rat and rabbit. J. Comp. Neural. 327: 63-82, 1993.

DE ZEEUW, C. I., WYLIE, D. R., DIGIORGI, P. L., AND SIMPSON, J. I. Projec- tions of individual Purkinje cells of identified zones in the flocculus to the vestibular and cerebellar nuclei in the rabbit. J. Comp. Neural. 349: 428-448, 1994.

DUFOSSE, M., ITO, M., JASTREBOFF, P. J., AND MIYASHITA, Y. A neuronal correlate in rabbit’s cerebellum to adaptive modification of the vestibulo- ocular reflex. Brain Res. 150: 61 l-616, 1978.

ECCLES, J. C., LLINAS, R., AND SASAKI, K. The excitatory synaptic action of climbing fibers on the Purkinje cells of the cerebellum. J. Physiol. Land. 182: 268-296, 1966.

ESCUDERO, M., DE LA CRUZ, R. R., AND DELGADO-GARCIA, J. M. A physiological study of vestibular and prepositus hypoglossi neurones projecting to the abducens nucleus in the alert cat. J. Physiol. Land. 458: 539-560, 1992.

FUCHS, A. F., SCUDDER, C. A., AND KANEDO, C. R. S. Discharge patterns and recruitment order of identified motoneurons and internuclear neurons in the monkey abducens nucleus. J. Neurophysiol. 60: 1874- 1895, 1988.

GERSTEIN, G. L. AND KIANG, W. Y. An approach to the quantitative analysis of equations of electrophysiological data from single neurons. Biophys. J. 1: 15-28, 1960.

GHELARDUCCI, B., ITO, M., AND YAGI, N. Impulse discharges from flocculus Purkinje cells of alert rabbits during visual stimulation combined with horizontal head rotation. Brain Res. 87: 66-72, 1975.

GRAF, W., SIMPSON, J. I., AND LEONARD, C. S. Spatial organization of visual messages of the rabbit’s cerebellar flocculus. II. Complex and simple spike responses of Purkinje cells. J. Neurophysiol. 60: 2091-2121, 1988.

ITO, M. Cerebellar control of the vestibulo-ocular reflex-around the flocculus hypothesis. Annu. Rev. Neurosci. 301: 275-296, 1982.

ITO, M., JASTREBOFF, P. J., AND MIYASHITA, Y. Specific effects of unilateral lesions in the flocculus upon eye movements in albino rabbits. Exp. Brain Rex 45: 233-242. 1982.

” \ ,


ITO, M., SHIDA, T., YAGI, N., AND YAMAMOTO, M. The cerebellar modification of rabbit’s horizontal vestibulo-ocular reflex induced by sustained head rotation combined with visual stimulation. Proc. Jpn. Acad. 50: 8% 89, 1974.

KANO, M., KANO, M. S., AND MAEKAWA, K. Binocular interaction and signal components of optoknetic responses of climbing fiber afferents in the cerebellar flocculus and nodulus of the pigmented rabbit. Neurosci. Res. 88: 455-458, 1991a.

KANO, M., KANO, M. S., AND MAEKAWA, K. Optokinetic response of simple spikes of Purkinje cells in the cerebellar flocculus and nodulus of the pigmented rabbit. Exp. Brain Res. 87: 484-496, 1991b.

KAWAGUCHI, Y. Two groups of secondary vestibular neurons mediating horizontal canal signals, probably to the ipsilateral medial rectus muscle, under inhibitory influences from the cerebellar flocculus in rabbits. Neu- rosci. Res. 2: 434-446, 1985.

KELLER, E. L. AND PRECHT, W. Visual-vestibular responses in vestibular nuclear neurons in the intact and cerebellectomized, alert cat. Neurosci- ence 4: 1599-1613, 1979.

KIMURA, M., TAKEDA, T., AND MAEKAWA, K. Contribution of eye muscle proprioception to velocity-response characteristics of eye movements: involvement of the cerebellar flocculus. Neurosci. Res. 12: 160- 168, 1991.

KOTCHABHAKDI, N. AND WALBERG, F. Cerebellar afferent projections from the vestibular nuclei in the cat: an experimental study with the method of retrograde axonal transport of horseradish peroxidase. Exp. Brain Res. 31: 591-604, 1978.

KUSUNOKI, M., KANO, M., KANO, M. S., AND MAEKAWA, K. Nature of optokinetic response and zonal organization of climbing fiber afferents in the vestibulocerebellum of the pigmented rabbit. I. The flocculus. Exp. Brain Res. 80: 225-237, 1990.

LEONARD, C. S. Signal Characteristics of Cerebellar Purkinje Cells in the Rabbit Flocculus during Compensatory Eye Movements (PhD thesis). New York: New York Univ., 1986.

LISBERGER, S. G., MILES F. A., AND ZEE, D. S. Signals used to compute errors in monkey vestibuloocular reflex: possible role of flocculus. J. Neurophysiol. 52: 1140- 1153, 1984.

LISBERGER, S. G. AND PAVELKO, T. A. Functional properties of brainstem cells inhibited from the cerebellar flocculus in monkey. Sot. Neurosci. Abstr. 10: 290, 1984.

LISBERGER, S. G. AND PAVELKO, T. A. Brain stem neurons in modified pathways for motor learning in the primate vestibulo-ocular reflex. Sci- ence Wash. DC 242: 771-773, 1988.

LISBERGER, S. G., PAVELKO, T. A., AND BROUSSARD, D. M. Responses during eye movements of brainstem neurons that receive monosynaptic inhibition from the flocculus and ventral paraflocculus in monkeys. J. Neurophysiol. 72: 909-927, 1994.

LOPEZ-BARNEO, J., DARLOT, C., BERTHOZ, A., AND BAKER, R. Neuronal activity in prepositus nucleus correlated with eye movement in the alert cat. J. Neurophysiol. 47: 329-352, 1982.

MAEKAWA, K. AND UMURA, M. Mossy fiber projection to the cerebellar flocculus from the extraocular muscle afferents. Brain Res. 19 1: 3 13 - 325, 1980.

MCCREA, R. A. AND BAKER, R. Anatomical connections of the nucleus prepositus of the cat. J. Comp. Neurol. 237: 377-407, 1985.

MCCREA, R. A., BAKER, R., AND DELGADO-GARCIA, J. Afferent and efferent organization of the prepositus hypoglossi nucleus. Prog. Brain Res. 50: 653-665, 1979.

MCCREA, R. A., YOSHIDA, K., BERTHOZ, A., AND BAKER, R. Eye movement related activity and morphology of second order vestibular neurons termi- nating in the cat abducens nucleus. Exp. Brain Res. 40: 468-473, 1980.

MCFARLAND, J. L. AND FUCHS, A. F. Discharge patterns in nucleus prepositus hypoglossi and adjacent medial vestibular nucleus during horizontal eye movement in behaving macaques. J. Neurophysiol. 68: 319-332, 1992.

MILES, F. A., FULLER, J. H., BRAITMAN, D. J., AND Dow, B. M. Long-term adaptive changes in primate vestibuloocular reflex. III. Electrophysiologi- cal observations in flocculus of normal monkeys. J. NeurophysioZ. 43: 1437- 1476, 1980.

NAGAO, S. Effects of vestibulocerebellar lesions upon dynamic characteristics and adaptation of vestibulo-ocular and optokinetic responses in pigmented rabbits. Exp. Brain Res. 53: 36-46, 1983.

NAGAO, S. Behaviour of floccular Purkinje cells correlated with adaptation

of optokinetic eye movement response in pigmented rabbits. Exp. Brain Res. 73: 489-497, 1988.

NAGAO, S. Behavior of floccular Purkinje cells correlated with adaptation of vestibulo-ocular reflex in pigmented rabbits. Exp. Brain Res. 77: 53 l- 540, 1989a.

NAGAO, S. Role of cerebellar flocculus in adaptive interaction between optokinetic eye movement response and vestibulo-ocular reflex in pigmented rabbits. Exp. Brain Res. 77: 541-55 1, 1989b.

NAGAO, S. Eye velocity is not the major factor that determines mossy fiber responses of rabbit floccular Purkinje cells to head and screen oscillation. Exp. Brain Res. 80: 221-224, 1990.

NAGAO, S. Contribution of oculomotor signals to the behaviour of rabbit floccular Purkinje cells during reflex eye movements. Neurosci. Res. 12: 169-184, 1991.

RANCK, J. B. Which elements are excited in electrical stimulation of mam- malian central nervous system: a review. Brain Res. 98: 417-440, 1975.

ROBINSON, D. A. A method of measuring eye movement using a scleral search coil in a magnetic field. IEEE Trans. Biomed. Electron. 10: 137- 145, 1963.

ROBINSON, D. A. Adaptive gain control of vestibuloocular reflex by the cerebellum. J. Neurophysiol. 39: 954-968, 1976.

SATO, Y., KAWASAKI, T., AND IKARASHI, K. Afferent projections from the brainstem to the floccular three zones in cats. II. Mossy fiber projections. Brain Res. 272: 37-48, 1983.

SIMPSON, J. I. AND ALLEY, K. E. Visual climbing fiber input to rabbit vestibulo-cerebellum: a source of direction-specific information. Brain Res. 82: 302-308, 1974.

SIMPSON, J. I., LEONARD, C. S., AND SOODAK, R. E. The accessory optic system of rabbit. II. Spatial organization of direction selectivity. J. Neuro- physiol. 60: 2055-2072, 1988.

SOODAK, R. E. AND SIMPSON, J. I. The accessory optic system of rabbit. I. Basic visual response properties. J. Neurophysiol. 60: 2037-2054, 1988.

STAHL, J. S. Signal Processing in the Vestibulo-Ocular Reflex Circuitry of the Rabbit (PhD thesis). New York: New York Univ., 1992.

STAHL, J. S. AND SIMPSON, J. I. Responses of abducens nucleus neurons to vestibular stimulation in awake rabbits. Sot. Neurosci. Abstr. 14: 955, 1988.

STAHL, J. S. AND SIMPSON, J. I. Floccular contribution to signal processing in the rabbit vestibular nucleus. In: Sensing and Controlling Motion, Vestibular and Sensorimotor Function, edited by B. Cohen, D. L. Tomko, and F. Guedry. New York: NY Acad Sci., 1992, p. 181-189.

STAHL, J. S. AND SIMPSON, J. I. Dynamics of rabbit abducens nucleus neurons in the awake rabbit. J. Neurophysiol. 73: 1383- 1395, 1995a.

STAHL, J. S. AND SIMPSON, J. I. Dynamics of rabbit vestibular nucleus neurons and the influence of the flocculus. J. NeurophysioZ. 73: 1396- 1413, 1995b.

STONE, L. S. AND LISBERGER, S. G. Visual responses of Purkinje cells in the cerebellar flocculus during smooth-pursuit eye movements in monkeys. I. Simple spikes. J. Neurophysiol. 63: 1241- 1261, 1990.

TAKEDA, T. AND MAEKAWA, K. Olivary branching projections to the flocculus, nodulus and uvula in the rabbit. I. An electrophysiological study. Exp. Brain Res. 76: 323-332, 1989.

TAKEMORI, S. AND COHEN, B. Loss of visual suppression of vestibular nystagmus after flocculus lesions. Brain Res. 72: 213 -224, 1974.

TAN, J., GERRITS, N. M., NANHOE, R. S., SIMPSON, J. I., AND VOOGD, J. Zonal organization of the climbing fiber projection to the flocculus and nodulus of the rabbit. A combined axonal tracing and acetylcholin- esterase histochemical study. J. Comp. Neurol. 356: l-22, 1995.

THACH, W. T. Somatosensory receptive fields of single units in the cat cerebellar cortex. J. Neurophysiol. 30: 675 -696, 1967.

THUNNISSEN, I. E. The Vestibulo-Cerebellar and Vestibulo-Oculomotor Pro- jections in the Rabbit (PhD thesis). Rotterdam, The Netherlands: Eras- mus Univ., 1990.

VAN DER STEEN, J., SIMPSON, J. I., AND TAN, J. Functional and anatomic organization of three-dimensional eye movements in rabbit cerebellar Aocculus. J. Neurophysiol. 72: 31-46, 1994.

VAN NEERVEN, J., POMEPEIANO, O., AND COLLEWIJN, H. Depression of the vestibulo-ocular and optokinetic responses by intrafloccular microinjec- tion of GABA-A and GABA-B agonists in the rabbit. Arch. Ital. Biol. 127: 243-263, 1989.

WAESPE, W., COHEN, B., AND RAPHAN, T. Role of the flocculus and para-


flocculus in optokinetic nystagmus and visual-vestibular interactions: ef- to the vestibular and cerebellar nuclei in the rabbit. J. Comp. Neural. fects of lesions. Exp. Brain Res. 50: 9-33, 1983. 349: 448-464, 1994.

WAESPE, W. AND HENN, V. Visual-vestibular interaction in the flocculus ZEE, D. S., LEIGH, R. J., AND MATHIEU-MILLAIREZ, F. Cerebellar control of of the alert monkey. II. Purkinje cell activity. Exp. Brain Res. 43: 349- ocular gaze stability. Ann. Neural. 7: 37-40, 1980. 360, 1981. ZEE, D. S., YAMAZAKI, A., BUTLER, P. H., AND GUCER, G. Effects of

WYLIE, D. R., DE ZEEUW, C. I., DIGIORGI, P. L., AND SIMPSON, J. I. Projec- ablation of flocculus and paraflocculus on eye movements in primate. J. tions of individual Purkinje cells of identified zones in the ventral nodulus Neurophysiol. 46: 878-899, 1981.

phase relations of purkinje cells in the rabbit flocculus

Documents