Journal of Neurophysiology

Cerebellar Nodulectomy Impairs Spatial Memory of Vestibular and Optokinetic Stimulation in Rabbits

N. H. Barmack, P. Errico, A. Ferraresi, H. Fushiki, V. E. Pettorossi, V. Yakhnitsa

Abstract

Natural vestibular and optokinetic stimulation were used to investigate the possible role of the cerebellar nodulus in the regulation and modification of reflexive eye movements in rabbits. The nodulus and folium 9d of the uvula were destroyed by surgical aspiration. Before and after nodulectomy the vertical and horizontal vestibuloocular reflexes (VVOR, HVOR) were measured during sinusoidal vestibular stimulation about the longitudinal (roll) and vertical (yaw) axes. Although the gain of the HVOR (GHVOR = peak eye movement velocity/peak head velocity) was not affected by the nodulectomy, the gain of the VVOR (GVVOR) was reduced. The gains of the vertical and horizontal optokinetic reflexes (GVOKR, GHOKR) were measured during monocular, sinusoidal optokinetic stimulation (OKS) about the longitudinal and vertical axes. Following nodulectomy, there was no reduction in GVOKR or GHOKR. Long-term binocular OKS was used to generate optokinetic afternystagmus, OKAN II, that lasts for hours. After OKAN II was induced, rabbits were subjected to static pitch and roll, to determine how the plane and velocity of OKAN II is influenced by a changing vestibular environment. During static pitch, OKAN II slow phase remained aligned with earth-horizontal. This was true for normal and nodulectomized rabbits. During static roll, OKAN II remained aligned with earth-horizontal in normal rabbits. During static roll in nodulectomized rabbits, OKAN II slow phase developed a centripetal vertical drift. We examined the suppression and recovery of GVVOR following exposure to conflicting vertical OKS for 10–30 min. This vestibular-optokinetic conflict reduced GVVOR in both normal and nodulectomized rabbits. The time course of recovery of GVVOR after conflicting OKS was the same before and after nodulectomy. In normal rabbits, the head pitch angle, at which peak OKAN II velocity occurred, corresponded to the head pitch angle maintained during long-term OKS. If the head was maintained in a “pitched-up” or “pitched-down” orientation during long-term OKS, the subsequently measured OKAN II peak velocity occurred at the same orientation. This was not true for nodulectomized rabbits, who had OKAN II peak velocities at head pitch angles independent of those maintained during long-term OKS. We conclude that the nodulus participates in the regulation of compensatory reflexive movements. The nodulus also influences “remembered” head position in space derived from previous optokinetic and vestibular stimulation.

INTRODUCTION

The modulation of climbing fiber responses of nodular Purkinje cells by natural vestibular stimulation evoked by roll about the longitudinal axis suggests that the nodulus participates in spatial orientation, particularly with regard to the regulation of axial musculature (Barmack and Shojaku 1995). Nodular vestibular signals are not restricted to climbing fibers. Vestibular primary afferents bifurcate as they enter the brain stem, sending equal numbers of afferents to the vestibular complex and nodulus where they terminate as mossy fibers in the granule cell layer (Barmack et al. 1993a; Cajal 1911; Carleton and Carpenter 1984; Carpenter et al. 1972; Korte 1979). A vestibular secondary mossy fiber afferent projection to the nodulus originates from cholinergic neurons in the caudal apects of the medial and descending vestibular nuclei (Barmack et al. 1992; Epema et al. 1985, 1990; Magras and Voogd 1985; Sato et al. 1989; Thunnissen et al. 1989).

Vestibular climbing fiber projections originate from the β-nucleus and dorsomedial cell column (dmcc) in the contralateral inferior olive (Barmack 1996, 1999; Barmack et al. 1993b; Kaufman and Perachio 1994; Kaufman et al. 1996). These climbing fibers convey information from the utricular otoliths and vertical semicircular canals. Each vertical canal is uniquely represented by a climbing fiber projection to a different sagittal zone within the uvula-nodulus (Barmack and Shojaku 1995; Fushiki and Barmack 1997).

The vestibular climbing fiber-defined zones within the uvula-nodulus project onto discrete regions of the vestibular nuclei. Purkinje cells located in the lateral nodulus project to the caudal medial vestibular nucleus and those in the medial part of the nodulus project to the middle part of the medial vestibular nucleus (Shojaku et al. 1987). The nodulus also projects differentially onto the cerebellar nuclei (Barmack et al. 2000; Wylie et al. 1994).

Lesions of the nodulus in monkeys in conjunction with vestibular stimulation and eye movement recording have been used to investigate how the nodulus might contribute to postural orientation in space. In particular, the influence of nodulus lesions on the plane, velocity, and decay of postrotatory vestibular nystagmus (PRN) have been studied as a function of head pitch angle in space (Angelaki and Hess 1995b; Wearne et al. 1998). Normally, the plane of horizontal PRN remains horizontal with respect to gravity as a monkey is pitched about an inter-aural axis, suggesting that PRN is coded in a gravito-inertial coordinate system. However, in nodulectomized monkeys PRN no longer remains horizontal in space, rather it is executed in an orbital coordinate system.

The rapid decay of PRN could interfere with steady-state measurements of spatial orientation. In rabbits, a long-lasting afternystagmus can be generated following long-term optokinetic stimulation (OKS). When a rabbit is exposed to binocular optokinetic stimulation for about 1 h and then placed in the dark, optokinetic nystagmus (OKAN I) develops. The slow phase of OKAN I occurs in the same direction as the OKS used to evoke it and lasts for several minutes. However, OKS for 16–48 h evokes an optokinetic afternystagmus (OKAN II) whose slow phase has the opposite direction. OKAN II can last for several hours (Barmack and Nelson 1987; Pettorossi et al. 1999). Similar to PRN, OKAN II reflects a gravito-inertial coordinate system (Pettorossi et al. 1999). Because OKAN II is long-lasting, it is possible to use it to evaluate coordinate systems whose integrity may depend on the nodulus, without the disadvantage of the rapid decay that characterizes PRN.

In preliminary experiments in rabbits with bilateral nodulectomies, we found that the plane of OKAN II remained in a gravito-inertial coordinate system when the rabbit was pitched about an inter-aural axis (Errico et al. 1996). In the present experiment we extend our analysis of the role of the nodulus in the orientation and spatial adaptation of OKAN II. We have examined the possible influence of the nodulus on vestibuloocular reflexes and optokinetic reflexes as well as its role in regulating OKAN II during pitch and roll vestibular stimulation. We conclude that the nodulus participates in the regulation of compensatory reflexive eye movements. This regulation is, in part, contingent on a “remembered” head position in space derived from previous optokinetic and vestibular stimulation.

METHODS

Subjects

Fourteen pigmented rabbits were anesthetized with ketamine hydrochloride (50 mg/kg), xylazine (6 mg/kg), and acepromazine maleate (1.2 mg/kg), to fix two, inverted, stainless steel screws to the calvarium aligned in a stereotaxic apparatus so that the lambda suture was 1.5 mm above the bregma suture (0 deg stereotaxic). These inverted stainless steel screws mated with a stiff head support bar on a vestibular rate table and with a flexible head support bar in the optokinetic stimulator. Both of these bars were inclined at a nose-down angle of 12 deg with respect to earth-horizontal. Consequently, when the rabbit was restrained in either apparatus, the plane of the horizontal semicircular canals was maintained horizontal in space. Rabbits were housed, handled, and stimulated according to the guidelines of the National Institutes of Health on use of experimental animals.

Surgical ablation of nodulus

In anesthetized rabbits (see Subjects), the occipital bone overlying folia 9b, 9c of the uvula was removed and the atlanto-occipital membrane was opened. The nodulus (folium 10) was removed by suction, usually with partial removal of folia 9c and 9d of the uvula (Fig. 1).

On completion of the experiment, the rabbit was deeply anesthetized (see Subjects) and perfused transcardially with 500 ml of saline at 100 cm pressure at room temperature. The rinse was followed by perfusion with 500 ml of 4% paraformaldehyde. The brain was removed and dehydrated in graded sucrose (10–30%) in 0.1 M PBS, pH 7.4, prior to being frozen and sectioned on a cryostat at 35 μM. Free-floating sections were mounted on slides for microscopic analysis and reconstruction of the nodulectomy.

Vestibular stimulation

The head of the rabbit was held rigidly in the center of rotation of a three-axis vestibular rate table, with the plane of the horizontal semicircular canals maintained in the earth-horizontal plane. The body of the rabbit was encased in foam rubber and fixed with elastic straps to a plastic tube aligned with the longitudinal axis of the rate table.

OKS

The vestibular rate table was located within a sphere having a diameter of 135 cm. A perforated aluminum globe, 4.4 cm in diameter, attached to the shaft of a galvanometer, was gimbal-mounted directly above the head of the rabbit (Leonard et al. 1988). A 20-W halogen light bulb with a rectangular filament was located at the center of the globe. The globe projected rectangular images of the filament on the interior surface of the large sphere that were 3 × 8.0 deg. These rectangular images were spaced every 15 deg with the long axis always orthogonal to the plane of rotation. The contrast of the projected images, K = (I Max +I Min)/(I MaxI Min) = 2.1, was enhanced by anodizing the interior and exterior surfaces of the perforated aluminum globe flat black.

For horizontal OKS, the rotational axis of the gimbal-mounted globe could be aligned with the vertical axis. For vertical OKS it was aligned with the longitudinal axis. Sinusoidal, monocular OKS was given at constant amplitude (±10 deg) over a frequency range of 0.005–0.200 Hz.

Long-term binocular OKS

Rabbits were placed in a restrainer at the center of a cylindrical drum (diameter, 110 cm; height, 115 cm). The interior wall of the drum comprised a contour-rich pattern. The head of the rabbit was restrained by a spring-loaded flexible coupling that attached to the rabbit restrainer and had holes mating with surgically implanted bolts on the calvarium. This flexible coupling permitted movement of the head in the sagittal plane, but prevented the potentially confounding influence of optokinetically evoked horizontal head movements. The flexible coupling maintained the head in a relatively constant orientation, aligning the plane of the horizontal semicircular canals with earth-horizontal (Barmack and Nelson 1987;Barmack and Young 1990). This method of restraint caused no pressure on any part of the body. It permitted partial movement of the head in the sagittal plane and vertical and lateral movement of the body. While restrained, rabbits were able to maintain normal posture with all four paws in contact with the support surface.

The optokinetic drum was rotated at a velocity of 5 deg/s, stimulating the left eye of the restrained rabbit in the posterior → anterior direction and the right eye in the anterior → posterior direction. During long-term OKS lasting 48 h, rabbits were removed from the optokinetic drum every 8 h. The rabbits were weighed and given food and water. Feeding was terminated only when rabbits withdrew to the rear of the feeding cage. This feeding regimen is sufficient for the rabbits to maintain body weight.

Rabbits easily adapted to this modified restraint and evinced no behavioral signs indicating a reluctance to resubmit to restraint after a hiatus for food and water. The heart rate of some rabbits was monitored during different periods of restraint. The heart rate remained within the normal range of 180–280 beats/min. When rabbits were first placed in the restraining apparatus, the heart rate was at the upper end of this range. After 1 h of restraint, the heart rate declined to 220–250 beats/min. After several hours of restraint, the heart rate declined further to 200–220 beats/min. During the evening, the heart rate declined to 180–220 beats/min. Any experimental intervention, such as increasing the level of ambient background auditory stimulation or touching the rabbit around the head, caused a temporary acceleration of heart rate.

One might expect that long-term OKS would induce nausea. However, rabbits lack an emetic reflex. Furthermore, the ingestion of ordinary rabbit food was undiminished even when long-term OKS caused obvious postural signs of spatial disorientation. These signs included optokinetic and head nystagmus as well as circling in a direction opposite to the drum rotation. These behaviors were suppressed by rabbits when they were feeding, but reappeared when they withdrew to the rear of the feeding cage.

Definitions of head and stimulus orientation during long-term OKS

During long-term OKS, the head-body of the rabbit was positioned at different pitch angles with respect to earth-horizontal. The axis of rotation of the optokinetic drum also could be independently aligned. The head orientation at which the horizontal semicircular canals were horizontal in space was defined as head pitch angle = 0 deg. The angle of the axis of rotation of the optokinetic drum when the axis was co-axial with the gravitational vector was defined asOKS angle = 0 deg. The axis of rotation of the optokinetic drum could be tilted forward by 20 or 35 deg (OKS angle = −20 or −35 deg) or backward (OKS angle = +20 or +35 deg) in the sagittal plane of the rabbit.

Four rabbits received long-term OKS with the head pitched at 0 deg and the drum rotating in the horizontal plane (head pitch angle = 0 deg and OKS angle = 0 deg). Three rabbits were stimulated at head pitch angle = +20 deg (nose-up) and OKS angle = 0 deg. In subsequent tests they were also stimulated at head pitch angle = −20 deg (nose-down) and OKS angle = 0 deg. Three rabbits were stimulated at head pitch angle = +35 deg (nose-up) and the OKS angle = +35 deg.

Vestibular stimulation during OKAN II

Static changes in head-body position about the inter-aural (pitch) and longitudinal (roll) axes were systematically varied after OKAN II was induced. The range of positions examined was ±180 deg for pitch and ±60 deg for roll.

Eye movement recording

Eye movements were measured by an infrared light projection technique (Barmack and Nelson 1987). The right eye was topically anesthetized with proparacaine hydrochloride. A small suction cup (3 mm diameter), on which a light-emitting diode (LED) was mounted, was affixed to the right eye. The entire suction cup–LED assembly weighed 135 mg. The narrow projection beam of the LED was aligned with the visual axis and detected with a photosensitivex-y position detector (SC 50, UDT), fixed relative to the head. The detector had a diameter of 3.8 cm and provided continuousX-Y voltages proportional to the horizontal and vertical position of the incident centroid of infrared light. This eye movement recording system was calibrated by moving the LED on a model of the rabbit eye through a known angular displacement. The system had a sensitivity of 0.2 min of arc and was linear to within 5% for eye deviations of 15 deg and to 8% for deviations of 30 deg (Barmack 1981).

In two rabbits eye movements were recorded with a 50 Hz infrared video camera to evaluate the gain of torsional vestibuloocular reflex (GTVOR) evoked by static steps about the inter-aural axis. Two LEDs were mounted on a suction contact lens. These LEDs were imaged by a video camera connected to a computer. The relative angle of the two LEDs on the eye during torsion was determined trigonometrically.

Eye position was expressed as a rotation vector E = tan (ρ/2)u (where u is the unity vector and ρ is the angle of rotation about u at a head pitch angle = 0 deg) (Haustein 1989). The angle of OKAN II with respect to earth-horizontal was reconstructed from its vertical and horizontal slow-phase eye velocity components and plotted as a function of head pitch angle. Horizontal and vertical components of the eye position vector were differentiated using Savitzky and Golay smoothing method to obtain coordinate velocity (Savitzky and Golay 1964). Angular eye velocity vector was calculated as Ω =2/(dE/dt + ExdE/dt)/(1 + ‖E2), where Ω is the angular eye velocity vector, E is the eye position, and xis the cross vector product. The derivative was taken by averaging the slopes of two adjacent position data points (200 points/s) of the horizontal and vertical components. Average slow-phase eye velocities of horizontal and vertical eye movements were computed from multiple inter-saccadic samples taken at 5-ms intervals. Averaged values were used because, under some stimulus conditions, slow phases were curvilinear rather than constant velocity. This was especially true following large changes in head pitch angle. Approximately 20–30 consecutive slow-phase velocities were averaged at different pitch or roll angles. The accuracy of this measurement was limited by the 5% linear error in eye position measurement.

Individual slow-phase velocity values were used only for describing the time course of changes in slow-phase velocity immediately following a step-change in head position. A first-order exponential decay was fitted to the eye responses to determine the time constants of the changes in nystagmus orientation and velocity: y =y 0 +Ae -t, wheret is time and τ is the time constant.

Short-term conflicting vestibular and optokinetic stimulation

Conjoint vestibular-optokinetic stimulation was used to test the possible contribution of the uvula-nodulus to optokinetic reduction of the gain of the vertical vestibuloocular reflexes (VVOR) (GVVOR). We used two different stimulus frequencies (0.02 and 0.10 Hz) to test these interactions. The VVOR was evoked during sinusoidal stimulation about the longitudinal axis for 2 min to obtain “control” records. Then, the gimbal-mounted planetarium projector was used to present an optokinetic stimulus of equal amplitude and direction as the vestibular stimulus. The “conflicting” optokinetic stimulus was initially presented monocularly for 4 min. Then, after the suction contact lens–LED was removed from the right eye, the rabbit received vestibular stimulation coupled with binocular OKS for an additional 20 min. Following this period of conjoint vestibular and optokinetic stimulation, the contact lens was re-attached to the right eye and vestibular stimulation coupled with monocular OKS was delivered for another 2–3 min. After this, the OKS was stopped and the recovery of the GVVOR was followed for the next 10–15 min.

Statistical analyses

A “mixed” analysis of variance (ANOVA) was the primary analytic model approach used to analyze the “before-and-after” Bode plots in Figs. 3 and 4 and the “OKAN II angle versus pitch angle” plots in Figs. 6 and 8. This approach used random and fixed class variables (or factors). Individual rabbits comprised random effects, while reflex gain, before and after, and frequency or pitch angle werefixed effects. Three different covariance types (unstructured, compound symmetry, and Huynh-Feldt) were evaluated. The optimal covariance type was selected as the one that minimized the Akaike's information criteria (Littell et al. 1996).

A full factorial repeated measures design was fit using reflex gain, frequency, and before-and-after, and all interactions of these effects. In addition to the ANOVA tables, this design also estimated adjusted means. These least-squares means were adjusted forother terms in the model. Differences between the least-squares means were estimated for the before-and-after factor but not for frequency or pitch angle. All analyses were run using SAS System, Version 8.1 (Littell et al. 1996). A difference was considered statistically significant at P < 0.05.

Statistical evaluations for the data illustrated in Figs. 7, 10, and 11and Table 1 were performed using two-tailed, paired Student's t tests. A difference was considered statistically significant at P < 0.05. Goodness of fit was established by χ2 for sine functions and by correlation coefficient, R, for exponential decay of first- and second-order polynomials. Fit was obtained by minimizing the mean square error between the data and the curve (Levenberg-Marquardt algorithm).

View this table:
Table 1.

Nodulectomy changes head pitch angle at peak OKAN II velocity

RESULTS

Nodulectomy included folia 9c and 9d of the uvula

The nodulectomy included most of the nodulus and usually parts of folia 9c and 9d. The extent of the damage to individual folia is documented schematically for 10 rabbits in Fig.1. Folia 9b was damaged in only two rabbits (R13 and R497). It is important to emphasize what was not damaged. Cerebellar damage did not extend beyond the lateral bounds of the posterior vermis or rostrally to the anterior lobe vermis. Nor was there damage to the brain stem or cerebellar nuclei, particularly the fastigial nucleus.

Fig. 1.

Bilateral nodulectomies also included damage to uvula, but not to cerebellar nuclei. A: schematic of the rabbit cerebellum in the sagittal plane, 0.2 mm from the midline. The nodulus consists of folium 10; the uvula consists of folia 9a–d. The rectangle corresponds to the area encompassed by the photomicrograph in B. The photomicrograph shows that the dorsal leaflet of folium 10 in R496 remained partially intact. The nodulectomy also partially damaged folium 9d. Note that there was neither damage to folia 9a–c, nor to the fastigial nucleus. C: schematic illustrations of the nodulectomy in each of the rabbits used in this experiment. N. Fast., fastigial nucleus.

Nodulectomy reduces GVVOR, but not horizontal vestibuloocular reflexes (GHVOR) or GTVOR

The effects of nodulectomy on eye movements were both time- and frequency-dependent. Within hours after recovery from the operation, spontaneous nystagmus was often detected (Fig.2). Probably this was related to the slight asymmetries of the nodulectomies. Spontaneous nystagmus was often accompanied by an increased GVVOR, especially at frequencies below 0.10 Hz. Higher stimulus frequencies (0.50–0.80 Hz) masked the nystagmus. After several days, the spontaneous nystagmus evoked at lower stimulus frequencies dissipated. There was a large cross-coupling effect of vertical and horizontal eye movements at lower stimulus frequencies (0.005–0.010 Hz) during vestibular roll about the longitudinal axis in both normal and nodulectomized rabbits. At higher stimulus frequencies (0.200–0.800 Hz) this cross coupling effect was minimal (Fig.2 C).

Fig. 2.

Multiple effects of nodulectomy on eye movements evoked by sinusoidal roll vestibular stimulation about the longitudinal axis. Horizontal and vertical eye positions of the right eye of R165 were measured during sinusoidal roll at several frequencies (0.005–0.800 Hz).A: prior to the nodulectomy, there was significant cross-coupling of vertical and horizontal eye position during roll about the longitudinal axis. B: 1 day after the nodulectomy the vertical vestibuloocular reflexes (VVOR) was accompanied by a nystagmus with an upward slow phase. This nystagmus was not evident at stimulus frequencies above 0.50 Hz.C: 16 days after the nodulectomy the nystagmus was gone, but the gain of the vertical vestibuloocular reflex (GVVOR) was reduced relative to the preoperative control. Arrowheads indicate transitions in stimulus frequency. VE, vertical eye position; HE, horizontal eye position; VT, vertical table position.

Because of the variability in spontaneous eye movements immediately after the nodulectomy, we waited 2–3 wk after the operation to evaluate the GHVOR and GVVOR. At this time there was a clear reduction in GVVOR (P < 0.01) and no reduction in GHVOR (P > 0.15) (Fig. 3).

Fig. 3.

Nodulectomy reduces the GVVOR. Bode plots of the horizontal vestibuloocular reflexes (HVOR) (A) and VVOR (B) were measured for 5 rabbits before and 2–3 wk after nodulectomy. There was a reduction in GVVOR(P < 0.01), but not in GHVOR(P < 0.15). Error bars indicate SE.

We measured the gain of the torsional vestibuloocular reflex (GTVOR) in two rabbits before and after nodulectomy using steps about the inter-aural axis. Preoperatively, for a step of 40 deg, GTVOR = 0.32 ± 0.08. Postoperatively, GTVOR = 0.30 ± 0.06. These measurements agree with the more systematic measurements of torsional eye movements in freely moving rabbits (Van der Steen and Collewijn 1984).

Nodulectomy does not reduce the gains of vertical or horizontal optokinetic reflexes

Rabbits were placed in the center of the optokinetic sphere. As the rabbit was oscillated sinusoidally, the gimbal-mounted globe projected contrast-rich images on the interior of the sphere. This optokinetic stimulus provided a monocular, unobstructed visual field of 190 deg horizontally and 130 deg vertically. The HOKR and VOKR were measured over a frequency range of 0.005–0.200 Hz, using stimulus amplitudes of ±10 deg. Four rabbits were tested before and at least 2 wk after nodulectomy. There was no change in GVOKR or GHOKR (Fig.4).

Fig. 4.

Nodulectomy does not reduce the gain of horizontal or vertical optokinetic reflexes (GHOKR, GVOKR). Monocular horizontal and vertical optokinetic reflexes were measured in 4 rabbits before and 2–3 wk after nodulectomy. Nodulectomy caused no change the gain of either reflex. Error bars indicate SE.

Nodulectomy does not alter the gravito-inertial reference frame of OKAN II during static pitch

Five rabbits were exposed to long-term OKS at an optokinetic drum velocity of 5 deg/s. The head position of the rabbit was maintained at head pitch angle = 0 deg, while the plane of the optokinetic drum rotation was also horizontal (OKS angle = 0 deg). Subsequently the rabbits were transferred from the optokinetic drum to the rate table and prepared for eye movement recording. In the dark, OKAN II developed over a period of 5–10 min, reaching relatively constant velocities of 15–50 deg/s. After constant velocity was developed at head pitch angle= 0 deg, the rabbit was statically pitched about the inter-aural axis in 5–15 deg increments. The rabbit remained at each head pitch angle for 20–30 s while the angle of OKAN II slow phase relative to orbit-horizontal was recorded (Fig.5). Measurements of the angle of OKAN II were made using a series of both ascending and descending pitch angles and OKAN II peak velocities at the same head pitch angles were averaged. In both pre- and postoperative testing OKAN II remained horizontal in space, independent of the head pitch angle (Fig.6, B and C).

Fig. 5.

Optokinetic afternystagmus (OKAN II) remains horizontal in space when rabbit is pitched about an inter-aural axis. Long-term binocular optokinetic stimulation (OKS) (48 h at 5 deg/s) was used to evoke long-lasting OKAN II. A 1: OKAN II is illustrated in a rabbit maintained at head pitch angle = 0 deg. A 2: the same animal was pitched “nose-down” by 60 deg. During changes in head pitch, the plane of OKAN II changed with respect to an orbital coordinate system. B: horizontal and vertical eye positions are plotted against each other to illustrate change in the angle of OKAN II relative to orbit-horizontal at different head pitch angles.

Fig. 6.

Nodulectomy does not disrupt the linear relation between head pitch angle and plane of OKAN II when the head is pitched about the inter-aural axis. A: OKAN II was evoked by 48 h of long-term binocular OKS at 5 deg/s. B: the influence of static pitch on the axis of eye rotation was measured in a group of 5 rabbits before (open circles) and after (filled circles) nodulectomy. The linear relationship between the axis of rotation and head pitch angle was unaffected by nodulectomy. Error bars indicate standard error of the mean. C: OKAN II angle was measured in 3 rabbits over a more extended range of head pitch angles (±180 deg) than the group data illustrated in B, before (open symbols) and after (filled symbols) nodulectomy. Data points from these 3 rabbits were fitted with sinusoids (prenodulectomy, light gray; postnodulectomy, dark gray) whose widths comprise a 95% confidence interval (χ2 = 134.5). The diagonal dashed line indicates theoretical perfect compensation of OKAN II angle for changes in head pitch angle. D: peak OKAN II velocity is plotted as a function of head pitch angle before (open symbols) and after (filled symbols) nodulectomy. The peak velocities of the preoperative data are normalized and fitted with a second-order polynomial (dashed line) within ±90° of head pitch angle. The postoperative data are normalized with respect to the preoperative data and fitted with a second-order polynomial (solid line). Note that there is a large postoperative increase in OKAN II velocity and that there is a phase shift of approximately 45 deg in the nose-down direction of the peak OKAN II velocity with respect to the preoperative data.

Nodulectomy shifts the head pitch angle at which maximal OKAN II velocity is evoked

Although the OKAN II remained horizontal in space during pitch after a nodulectomy, the postnodulectomy characteristics of OKAN II velocity were changed in two respects. First, there was a general increase in the velocity of OKAN II at head position angles ranging from −135 to +45 deg (Fig. 6 D). In five rabbits measured 2 wk after the nodulectomy, OKAN II velocity was elevated with respect to the preoperative values. Second, after nodulectomy, the head pitch angle at which peak OKAN II velocity was evoked was shifted from 0 to −45 deg (nose-down). This shift in the angle at which OKAN II peak velocity was achieved following nodulectomy was independent of the head pitch angle maintained during long-term OKS. The same postnodulectomy shift was observed following long-term OKS in which the head pitch angle was varied from −20 deg (nose-down) to +20 deg (nose-up) (Table1).

Nodulectomy does not alter time constants for changes in OKAN II angle and velocity following quasi-step changes in head pitch angle

We measured time constants for changes in OKAN II velocity and for changes in the angle of OKAN II following quasi-step changes in head pitch angle in three rabbits before and after nodulectomy. Quasi-steps of 60 deg about the inter-aural axis were completed within 1 s. In both normal and nodulectomized rabbits, these quasi-steps evoked comparable changes in OKAN II velocity and OKAN II rotation angle. In normal rabbits, nose-down steps, from an initial head pitch angle = 0 deg, were characterized by time constants of 1.9 s. Changes in OKAN II velocity were characterized by time constants of 1.5 s (Fig. 7). These time constants were shortened nominally, without statistical significance, to 1.8 and 1.4 s, respectively, following nodulectomy (Student'st test, P > 0.20).

Fig. 7.

Nodulectomy does not alter the time constant for change in OKAN II axis following quasi-step changes in head position about the inter-aural axis. The time constants for change in OKAN II axis and change in OKAN II velocity following quasi-step changes in head pitch angle were measured in 3 rabbits before and after nodulectomy.A 1,2: before nodulectomy, quasi-step changes in head pitch angle evoked changes in the rotation angle of OKAN II relative to orbit-horizontal as well as changes in OKAN II slow-phase velocity. A 60 deg nose-down quasi-step, from an initial head pitch angle = 0 deg, was executed within 1 s, as indicated by the light gray bars. This step evoked a compensatory change in the axis of rotation of OKAN II with a time constant of 1.9 s. Such a quasi-step also evoked a decrease in slow-phase velocity with a time constant of 1.5 s. B 1,2: after nodulectomy there was only a nominal change in these time constants without statistical significance. Solid lines indicate the averaged responses and dashed lines indicate exponential curves fitted to the averaged responses.

Nodulectomy disrupts horizontal trajectory of OKAN II during static roll

OKAN II was induced in three rabbits as described above before and after nodulectomy. The trajectories of OKAN II slow phases were examined during static roll about the longitudinal axis. Before nodulectomy, the trajectory of the OKAN II slow phase remained horizontal, independent of the vertical position of the eye within the orbit (Fig. 8 A, open circles). After the nodulectomy OKAN II slow phases started from an initial vertical position that was dependent on the amplitude of the static roll. However, OKAN II slow phases did not remain horizontal, but acquired a centripetal vertical drift. This drift was larger for larger amplitude rolls, in which case the initial vertical deviation of the eyes was larger. For static rolls of 60 deg, the trajectories of the OKAN II slow phases deviated from horizontal by 35–45 deg. These changes in slow-phase trajectories are illustrated diagrammatically in Fig. 8 B.

Fig. 8.

Nodulectomy distorts plane of OKAN II during static roll.A: in normal rabbits, static roll about the longitudinal axis caused a change in vertical eye position while OKAN II is simultaneously expressed. The plane of OKAN II remains horizontal while the eye changes its vertical position. Following nodulectomy, the plane of OKAN II is distorted. The slow phase of OKAN II is associated with a centripetal vertical drift. The angle of this centripetal drift increases for vertical eye positions of greater eccentricity. This is illustrated diagrammatically in B 1,2. Error bars indicate SE.

Nodulectomy does not alter the depression and recovery of GVVOR by vertical OKS

The nodulus receives vestibular climbing fiber signals that encode roll about the longitudinal axis. The activity of a subset of these vestibularly modulated climbing fibers also can be modulated by OKS about the longitudinal axis (Barmack and Shojaku 1995). In both normal and nodulectomized rabbits we tested whether conflicting vertical OKS could depress GVVOR and whether prolonged vestibular-optokinetic conflict (30 min) influenced the time course of recovery of GVVOR when the conflict was stopped. We used two different stimulus frequencies, 0.02 and 0.10 Hz, to test these interactions. The responses of both intact and nodulectomized rabbits to conflicting vestibular-OKS were the same (Fig. 9). At 0.02 Hz the optokinetic conflict caused a 55% reduction in GVVOR within two cycles of conflicting OKS. GVVOR recovered incompletely after the optokinetic conflict was stopped. The optokinetically induced depression of GVVOR at 0.10 Hz developed more slowly, presumably because of the lower gain of the optokinetic reflex at this frequency. However, after the binocular period of conflicting stimulation, the GVVOR was reduced by 60%. The recovery of GVVOR at 0.10 Hz was also incomplete. We conclude that the nodulus does not play a critical role in the reduction of GVVOR during vestibular-optokinetic conflict.

Fig. 9.

Nodulectomy does not influence OKS-induced reduction in GVVOR or the time course of GVVOR recovery. Rabbits received sinusoidal vestibular stimulation about the longitudinal axis at either 0.02 Hz (A) or 0.10 Hz (B). After 2 min, conflicting sinusoidal OKS of equal amplitude and frequency was given for 30 min, depressing GVVOR. When the conflicting OKS was stopped, GVVOR recovered with a time course of minutes. The OKS-induced reduction in GVVOR and the subsequent GVVOR recovery were equivalent before and after nodulectomy. Error bars indicate SE.

Nodulectomy impairs memory of head pitch angle as reflected in OKAN II velocity

Three rabbits received long-term OKS with the head pitch angle = +20 deg (nose-up, Fig.10 B 1) or −20 deg (nose-down, Fig. 10 B 2). These head pitch angles oriented the head obliquely in space while the OKS angle = 0 deg. Preoperatively, OKAN II remained horizontal in space while the rabbits were pitched about the inter-aural axis. The peak velocity of OKAN II was evoked at a head pitch angle corresponding to that maintained during the long-term OKS (Fig.10 A 1,2, Table 1). OKAN II peak velocity occurred at a head pitch angle = +28.9 deg for rabbits maintained at a head pitch angle = +20 deg (nose-up). Similarly, when the head pitch angle = −20 deg (nose-down) during long-term OKS, peak OKAN II velocity occurred at a head pitch angle = −29.5 deg.

Fig. 10.

Nodulectomy abolishes “memory” of head position–dependent peak OKAN II velocity. In normal rabbits peak OKAN II velocity is associated with the head pitch angle during long-term OKS. On either side of this optimal pitch angle, OKAN II is evoked at lower velocities. To test the influence of “remembered” head pitch angle on OKAN II velocity, 3 rabbits received horizontal long-term OKS at two different head pitch angles: B 1: head pitch angle = +20 deg, OKS angle = 0 deg, and B 2: head pitch angle = −20 deg, OKS angle = 0 deg. The data from one such rabbit, R36, are plotted here. Following long-term OKS, the plane of OKAN II (A 1) and OKAN II (A 2) peak velocity were measured at different head pitch angles. Each data point represents the mean of 20 consecutive OKAN II slow-phase values. Note that the pitch angle at which peak OKAN II slow-phase velocity occurred was dependent on the pitch angle maintained during long-term OKS.A 1: data points were fitted with sinusoids (light gray for +20 deg; dark gray for −20 deg) whose widths comprise a 95% confidence interval (χ2 = 134.5).A 2: data points representing OKAN II velocity following long-term OKS at head pitch angle = +20 deg (open circles) were normalized and fitted with a second-order polynomial (dashed line) within ±90° of the head pitch angle at which peak OKAN II velocity occurred. Data points representing OKAN II velocity following long-term OKS at head pitch angles = −20 deg (filled circles) were normalized with respect to the +20 deg fitted function and fitted with a second-order polynomial (solid line). These measurements were repeated 2 wk after nodulectomy.C 1: as in the prenodulectomy measurements, the plane of OKAN II remained aligned with earth-horizontal at different head pitch angles. The diagonal dashed lines inA 1 and C 1indicate theoretical perfect compensation of OKAN II angle for changes in head pitch angle. C 2: after nodulectomy, peak OKAN II velocities increased, but the head pitch angle at which OKAN II velocity was maximal no longer depended on the head pitch angle at which the head was maintained during long-term OKS. The fitted second-order polynomials were normalized with respect to the +20 deg preoperative measurements.

Postoperatively, following long-term OKS, OKAN II remained horizontal in space while the head was pitched about the inter-aural axis (Fig.10 C 1). However, the head pitch angle at which peak OKAN II velocity occurred was independent of the head pitch angle maintained during long-term OKS. Peak OKAN II velocities occurred at head pitch angle = −45 deg (nose-down) for stimulus conditions in which the rabbit was maintained either at head pitch angles = +20 deg (nose-up) or −20 deg (nose-down) (Fig.10 C 1,2, Table 1).

Head pitch angle rather than retinal stimulation angle determines peak OKAN II velocity

Changes in head pitch angle not only reflect changes of the head within a vestibular space, but also reflect changes in the OKS angle on the retina. Our previous measurements of the TVOR indicated that a step-pitch vestibular stimulus of 20 deg would evoke a torsional eye movement with GTVOR = 0.3. Assuming that this torsional gain remained constant during vestibular and concomitant optokinetic stimulation, a torsional change in eye position during a maintained head pitch angle = 20 deg could reduce the OKS angle on the retina by as much as 6 deg, resulting in a net 14 deg OKS angle on the retina.

To rule out the possibly confounding effects of head pitch angle and retinal stimulation angle on OKAN II peak velocity, we examined OKAN II in three rabbits under two different stimulus conditions before and after nodulectomy. In the first condition, rabbits received long-term OKS at a head pitch angle = 0 deg and OKS angle = 0 deg. In the second condition rabbits received long-term OKS at a head pitch angle = +35 deg and OKS angle = +35 deg (nose-up). Again, assuming a GTVOR = 0.3, this stimulus condition would result in an OKS angle of +24.5 deg on the retina.

Prior to the nodulectomy when both the head pitch angle and the OKS angle were maintained at +35 deg, peak OKAN II velocity was phase shifted by +30.7 deg (Fig.11 A 1,2, Table 1). In other words, the OKAN II peak velocity reflected a “remembered” head pitch angle, more than OKS angle on the retina. Following long-term OKS in nodulectomized rabbits, OKAN II peak velocity occurred at a head pitch angle = −45.6 deg (nose-down) (Fig. 11 C 1,2, Table 1).

Fig. 11.

Nodulectomy reduces influence of head position maintained during long-term OKS on subsequently evoked OKAN II slow-phase velocity.Rabbit 36 received long-term OKS under two conditions.B 1: first, the head pitch angle = 0 deg and OKS angle = 0 deg (open circles).B 2: second, the head pitch angle = +35 deg and the OKS angle = +35 deg (filled circles).A 1: before nodulectomy, OKAN II remained horizontal in space, when the head was pitched about the inter-aural axis. Data points were fitted with a sinusoid whose width comprises a 95% confidence interval. A 2: the preoperative velocity of OKAN II peaked at +30 deg for the OKAN II evoked by long-term OKS with the head pitch angle = +35 deg (B 2). Data points inA 2 were fitted with a second-order polynomial, normalized with respect to measurements of OKAN II velocity following long-term OKS at a head pitch angle = 0 deg (B 1). C 1: after nodulectomy, OKAN II remained horizontal in space, when the head was pitched about the inter-aural axis. The diagonal dashed lines inA 1 and C 1indicate theoretical perfect compensation of OKAN II angle for changes in head pitch angle. C 2: peak OKAN II velocity showed no dependence on head pitch angle maintained during long-term OKS. For both stimulus conditions, peak OKAN II velocity was recorded with a head pitch angle = −45 deg (nose-down).

DISCUSSION

Regulation of OKAN II orientation and velocity during static changes in head pitch angle

There are four major findings in this report. First, nodulectomy reduces the gain of the vertical (GVVOR), but not horizontal (GHVOR), vestibuloocular reflex. Second, both normal and nodulectomized rabbits were capable of maintaining OKAN II horizontal in space when pitched about the inter-aural axis. Third, the OKAN II was characterized by a centripetal drift when the eye was vertically displaced by static roll tilt. Fourth, the nodulus contributed to plastic regulation of OKAN II velocity. In normal rabbits the head pitch angle at which OKAN II attained peak velocity was dependent on the head pitch angle maintained during long-term OKS. If the head was pitched nose-down during long-term OKS, then OKAN II peak velocity occurred when the rabbit was subsequently tested in a nose-down orientation. This dependence of OKAN II velocity on head pitch angle maintained during long-term OKS was absent in nodulectomized rabbits.

It should not be surprising that a nodulectomy does not reduce the gain of the HVOR. Electrophysiological studies of nodular Purkinje neurons have shown a total absence of climbing fiber-mediated responses evoked by horizontal vestibular stimulation (Barmack and Shojaku 1995; Fushiki and Barmack 1997). The lack of horizontally evoked CFRs does not preclude the possibility that primary afferents project to the nodulus as mossy fibers. While there is anatomical evidence for such a projection, with the exception of an earlier physiological study (Precht et al. 1976), there is little evidence that mossy fiber signals contribute significantly to the modulation of simple spikes recorded from nodular Purkinje cells (Barmack et al. 1993a; Fushiki and Barmack 1997).

Our data concerning the contribution of the nodulus to regulation of reflexive eye movements during vestibular stimulation about the roll axis are more problematical. Although nodulectomy caused a small, but statistically significant reduction in GVVOR, this reduction may have been a secondary consequence of some other role for the nodulus that has gone untested. For example, if the nodulectomy altered the normally maintained posture of the head, then the plane at which GVVOR would be optimized might have been shifted, as it was in the case of optokinetic modification of peak OKAN II velocity orientation discussed in the following text.

Absence of changes in orientation of OKAN II in nodulectomized rabbits compared with changes in orientation of PRN in nodulectomized primates

The nodulus plays no role in the orientation of OKAN II in rabbits. However, in nodulectomized monkeys, PRN and OKAN I do not remain horizontal in space. Rather, PRN appears to be executed in an orbital coordinate system (Angelaki and Hess 1995a;Wearne et al. 1998). While these conflicting results might be attributed to differences in lateral-eyed and frontal-eyed experimental species, they might also reflect other possibilities such as the following: 1) extent of damage to nonnodular tissue, or 2) different types of eye movements analyzed; PRN and OKAN I in the monkey versus OKAN II in the rabbit. With regard to the question of species differences, a similar zonal organization of the nodulus and uvula in monkeys and rabbits as well as projections from discrete subnuclei of the inferior olive and efferent connection with the vestibular nuclei has been found (Voogd et al. 1996). Differences between PRN and OKAN II may be important. Both PRN and OKAN II reflect an imbalance in preoculomotor circuitry. Superimposed on this imbalance is an otolithic signal indicating head position in space. During OKAN II, this otolithic signal may not be superimposed on the same preoculomotor circuitry engaged by PRN and OKAN I. In PRN the otolith signal may interact with the impulsive and decaying signals from the semicircular canals generated by deceleration following rotation at constant velocity. Depending on conditions of vestibular system stimulation, PRN may be expressed in a coordinate system defined by orientation of the semicircular canals (Fushiki et al. 2001).

In rabbit, the nodulus appears necessary to maintain the horizontal orientation of OKAN II during roll tilt. In its absence, there is a vertical centripetal drift of the eye toward a more central orbital position. This drift could be coupled to a change in GVVOR, a possibility not tested in the present investigation.

Orientation of OKAN II during static changes about the roll axis

The advantage of using OKAN II is that, unlike PRN, it is long-lasting and has a relatively constant velocity. OKAN I and PRN are transient phenomena. In monkeys, PRN has been used to model spatial orientation by three independent processes, as follows: 1) control of the horizontal time constant, 2) control of vertical and torsional time constants, and 3) generation of cross-coupled vertical and torsional velocity component as the gravito-inertial axis tilts relative to the head (Wearne et al. 1998). In nodulectomized monkeys, the time constant for horizontal velocity storage, determined from OKAN I and PRN, increased from 12 to 23.6 s, while the vertical and roll time constants decreased from 10.1 to 3.3 s (vertical) and 13.4 to 2.7 s (roll) (Wearne et al. 1998). If these time constants were differentially modified by nodulectomy, then the time course PRN reorientation might reflect different orientations, depending on the time course of PRN decay. Conversely, in rabbits, OKAN II was essentially time-invariant. Consequently, it was easy to determine that nodulectomy did not interfere with the earth-horizontal orientation of OKAN II or its rate of adjustment following quasi-step changes in head pitch angle.

Nodulectomy abolishes influence of head pitch angle maintained during long-term OKS on head pitch angle at which OKAN II peak velocity is evoked

In normal and nodulectomized rabbits head pitch angle and OKS angle maintained during long-term OKS did not influence the plane of subsequently evoked OKAN II. However, in normal rabbits the head pitch angle at which OKAN II peak velocity was recorded wasinfluenced by the head pitch angle and OKS angle maintained during long-term OKS. This influence of head pitch angle and OKS angle on OKAN II peak velocity was lost in nodulectomized rabbits. Similar conclusions were reached for nodulectomized monkeys (Waespe et al. 1985; Wearne et al. 1998). However, in monkeys, the distinction between PRN orientation and velocity was attributed to the sparing of lateral regions of the nodulus and uvula, regions to which velocity regulation was ascribed.

These data raise the interesting possibility that the nodulus may regulate the slow-phase velocity of reflexive eye movements contingent on head position. This regulation is modifiable when the nodulus is intact. In nodulectomized monkeys, the orientation of PRN was no longer constant in space. In nodulectomized rabbits, although the plane of OKAN II remained horizontal in space, OKAN II peak velocity no longer reflected the head pitch angle maintained during long-term OKS. In nodulectomized rabbits, maximal OKAN II velocity occurred at a head position angle ∼ −45 deg. This angle may represent a range limit of downward head pitch normally experienced by the rabbit. Within this range, the nodulus modulates nystagmus velocity. Without the nodulus, head position information (otolithic input) tends toward a default setting corresponding to a downward head pitch limit.

Direct or indirect associative role of nodulus in regulation of OKAN II velocity?

The influence of head position angle on OKAN II velocity may be mediated either directly or indirectly through the nodulus. A direct influence could be sustained by the projections of the nodulus to regions within the vestibular complex (Walberg and Dietrichs 1988). Several anatomical studies have demonstrated that Purkinje cells located within sagittal zones in the uvula, as well as the nodulus, project to different regions within the vestibular complex (Bernard 1987; Shojaku et al. 1987;Tabuchi et al. 1989; Wylie et al. 1994).

However, the modulation of OKAN II velocity may be an indirect consequence of nodular influence on autonomic mechanisms, such as blood pressure. Changes in sympathetic outflow and blood pressure can be evoked by electrical stimulation of the uvula and nodulus (Bradley et al. 1987, 1990). Reflexive eye movements can also be influenced by changes in head position (Ray 2000; Yates 1992). A change in OKAN II velocity could indirectly reflect altered orthostatic functions normally influenced by the nodulus. Regardless of the exact mechanism by which OKAN II velocity is regulated, our results suggest that the nodulus plays an important associative role in the refinement of movement and posture based on sensory experience.

Acknowledgments

We thank M. Westcott and D. Bambagioni for expert histological and technical assistance.

This research was supported by National Eye Institute Grant EY-04778 to N. H. Barmack and grants from Consiglio Nazionale delle Ricerche, Ministero dell'Universitá e della Ricerca Scientifica e Tecnologica, and the Italian Space Agency (ASI I/R/101/00) to V. E. Pettorossi.

Footnotes

  • Address for reprint requests: N. H. Barmack, Neurological Sciences Institute, Oregon Health and Science University West Campus, 505 NW185th Ave., Beaverton, OR 97006 (E-mail:barmackn{at}ohsu.edu).

REFERENCES

View Abstract