The Role of Gravity in Adaptation of the Vertical Angular Vestibulo-Ocular Reflex

Abstract

The gain of the vertical angular vestibulo-ocular reflex (aVOR) was adapted in side-down and prone positions in two monkeys and tested in four planes: left-/right-side down; forward/backward; and two intermediate planes that lie approximately in the planes of the vertical semicircular canal pairs, left anterior/right posterior (LA/RP) and right anterior/left posterior (RA/LP). Gain changes, expressed as a percent of preadapted values, were plotted as a function of head orientation in the planes of tilt, and fitted with sinusoids to obtain the gravity-dependent (amplitude) and gravity-independent (bias) components of adaptation. Gravity-dependent gain changes were always maximal when tested in a plane that included the head orientation in which the aVOR gain had been adapted. Changes were minimal when the head was tilted in a plane orthogonal to the plane of adaptation, and were smaller but still significant when tested in the two intermediate planes. Gravity-independent VOR gain changes were uniform over all planes of head tilt. Thus, the gravity-dependent and gravity-independent components could be separated experimentally. The aVOR gain changes from the head tilts in different directions were utilized to reconstruct the gain changes in three dimensions. They formed a continuous surface, which peaked in and around the position of adaptation. These studies support the postulate that gain adaptation has both gravity-independent and gravity-dependent components, and further show that these gain changes have a three-dimensional structure. These results are similar to those in humans, indicating that the gravity-dependent adaptation of the aVOR is likely to be a common phenomenon across species.

INTRODUCTION

The angular vestibulo-ocular reflex (aVOR) stabilizes gaze during head rotation by generating counterrotation of the eyes in the orbit. How well the aVOR performs this function can be characterized by the gain (eye velocity/head velocity). If eye counterrotation is inadequate, a retinal slip signal is generated, and consistent mismatch between head and eye movements stimulates changes in the aVOR gain to reduce retinal slip. In both monkeys and humans the first significant gain changes can be produced within 20 to 40 min,^1-6 and the gain changes can persist for many hours if subjects are kept in darkness with their heads fixed.⁷

Recently, it was demonstrated that aVOR gain adaptation depends on head orientation with regard to gravity for both the horizontal^8,9,10 and vertical aVOR.¹¹ By testing animals in several head orientations from left-side down (LSD) to right-side down (RSD), we demonstrated that aVOR gain changes are a continuous function of head orientation with regard to gravity. In these studies the peak gain changes occurred approximately in the position in which the aVOR gain was adapted and the gain changes declined as the head was moved away from this position. The evoked gain changes were closely approximated with a sinusoidal function and a bias. We inferred from these results that each gain change was composed of two separate components, a gravity-dependent gain change, which was represented by the sinusoid, and a gravity-independent component, which was the bias.^12-14 In a study of responses of the vertical aVOR of a number of monkeys, tilted from right-side down to left-side down positions, the magnitudes of both the gravity-dependent and gravity-independent components were comparable. This implied that the gravity-dependent component is of substantial magnitude. Furthermore, similar responses were found both in monkeys and in humans.^12-14

Several important questions arose from our previous results. In the previous studies, animals were tilted only in one plane. Therefore, it was not known whether head tilts in any direction had gravity-dependent and gravity-independent components. The spatial distribution of the gain changes also was not known. In the present study we tested vertical aVOR gains for a variety of head tilts after having adapted the gain in LSD, RSD, and prone head orientations, and reconstructed the gain changes in three-dimensional space.

METHODS

Two cynomolgus monkeys (Macaca fascicularis, M0102, M17115) were utilized in this study. The experiments conformed to the Guide for the Care and Use of Laboratory Animals,¹⁵ and were approved by the Institutional Animal Care and Use Committee. The methods used have been described in detail elsewhere^12,16 and will be summarized briefly. Surgical procedures were performed under anesthesia in sterile conditions in two stages. A head mount was implanted on the skull to provide painless head fixation in stereotaxic coordinates during testing,^16,17 and eye coils were implanted on the left eye. One coil measured the horizontal and vertical components of eye position,^18,19 and another, placed approximately orthogonal to the frontal coil,⁴ measured the torsional component of eye position.

During testing, the monkey’s head was fixed to a plastic frame, which held two sets of field coils. The axes of the field coils were along the interaural (pitch) and dorsoventral (yaw) axes of the head, establishing a head-fixed reference frame for measuring the orientation of the frontal and top search coils. Monkeys were positioned so that the eye with the search coils was at the center of the magnetic fields. A four-axis vestibular stimulator was used in this study.¹² The animals were surrounded by an optokinetic drum with subject-horizontal, 10-deg stripes that filled their field of vision.

All three components of eye movements were recorded, but only vertical components were relevant for this study. Vertical eye velocities were calibrated by rotating the animals in light at 30 deg/s about a spatial vertical axis while left-side down. It was assumed that vertical gain was close to unity in these conditions.²⁰ To decrease the aVOR gain, the primate and drum axes were rotated together in light with sinusoid of 0.2 Hz and peak velocity of ~42 deg/s (±14 deg amplitude). To increase the vertical aVOR gain, the animal and optokinetic drum were rotated sinusoidally at 0.2 Hz at peak velocity ~18 deg/s (±15 deg amplitude) in opposite directions. For the latter, the animals observed a vertical visual surround movement of 36 deg/s (18 deg/s + 18 deg/s). Vertical aVOR gains were adapted in one of three positions: left-side down (LSD), right-side down (RSD), or prone. Adaptation was carried out over a 4-h period in each instance. Animals were tested after each hour of gain adaptation, and in some cases one and two days after adaptation. At least seven days were allowed for aVOR gain to recover before the next gain adaptation was performed.

To determine changes in aVOR gain as a function of head position with regard to gravity, the animals were oscillated sinusoidally at 0.5 Hz 60 deg/s in darkness about a pitch (interaural) axis that was either upright or tilted in 10° increments up to 90°. The head was tilted in four different planes: sagittal (forward–backward, F-B), coronal (LSD/RSD), and the two intermediate planes that lie approximately in the planes of the two vertical semicircular canal pairs (left anterior/right posterior, LA/RP; and right anterior–left posterior, RA/LP). Because the animals were always oscillated about an interaural axis, canal activation was the same in every head orientation in each test, but the direction of the average static pitch that activated the otoliths varied as a function of the head tilt. Dynamic otolith activation varied for different directions of tilt, but as shown before, dynamic otolith activation over such small angles of tilt has no effect on the gain of the vertical aVOR.¹²

Desaccaded eye velocity was fitted with sinusoids to estimate the gain and phase of the response in each head orientation (temporal gain and phase). Changes in gain were expressed as a percentage relative to preadapted level and plotted as a function of head tilt. To analyze the magnitude of the gravity-dependent aVOR gain adaptation, we applied a sinusoidal approximation with an unknown bias (C) to this residual function:

$$y=A*\cos (x+B)+C$$

where C is a measure of the gravity-independent component, A is the magnitude of the gravity-dependent component, and B is the spatial phase. Depending on the type of experiment, standard t-tests and ANOVA with a Bonferroni post hoc approach were used to analyze pairs or sets of data, respectively.

RESULTS

A typical example of vertical eye velocities evoked by oscillation of an animal about an interaural axis in various head orientations relative to gravity is shown in Figure 1. When the animal was oscillated in any head orientation with regard to gravity before adaptation, the evoked eye velocities were approximately the same (Fig. 1, gray traces). After the gain of the vertical aVOR had been decreased with the animal RSD, the evoked eye velocities were much smaller when animal was tested in this head orientation (Fig. 1, black trace, 90°). For tilts of 60°, the reduction in gain was still substantial, but was smaller in 30° head orientations. When the animal was LSD (−90°), the gain was close to the preadapted value (LSD; Fig. 1, top trace).

FIGURE 1.

Vertical (pitch) eye velocities evoked by oscillation of the head about an interaural axis, which was spatially horizontal or tilted up to 90°. The cartoons on the left represent the head orientations at which the corresponding traces on the right were obtained. The evoked eye velocities were approximately the same in all head orientations when tested before adaptation (gray traces). When eye velocities were tested after four hours of adaptation, the smallest eye velocities were obtained when RSD, that is, at the head orientation in which the gain was adapted (second trace from the bottom). The changes in the amplitude of the eye velocities became progressively smaller as the head was oriented away from the position of adaptation. Upward deviations of the traces represent downward eye velocities. The bottom trace is stimulus velocity. The dashed gray vertical lines are saccades, which were eliminated from the analyses.

In another experiment, the aVOR gain was decreased LSD. The gains were calculated in each head orientation about four axes and plotted as a function of the head tilt (Fig. 2). When the animal was oscillated about its interaural axis while positioned from LSD to RSD (coronal plane; Fig. 2A, insert above) before adaptation, the gain was uniform for most head orientations, although slightly decreased when LSD (Fig. 2A, open symbols). After the vertical aVOR gain had been decreased in LSD for four hours, the aVOR gain changes were substantial when tested LSD and gradually decreased as the animal was reoriented toward RSD (Fig. 2A, filled symbols). Gain changes for each head orientation in that plane were expressed as percentages of the preadapted values and a sinusoid with a bias could accurately fit the data (Fig. 2E). The gravity-dependent gain change was 15.7% (A), whereas the gravity-independent (C) was −20.6%.

FIGURE 2.

Vertical aVOR gain before (open symbols) and after four hours of gain decrease when LSD. The animals were tested during tilts along the LSD/RSD plane (A), the LA/RP plane (B), the F/B plane (C), and the RA/LP plane (D). The dashed lines in (A) to (D) represent the average values over all gain measurements obtained for a corresponding plane of head tilt before adaptation. The gains after adaptation, shown in (A) to (D), were expressed as a percent of the preadapted value for each plane of head tilt (E–H), and were fit by sinusoids with a bias. The bias was the gravity-independent change and the amplitude of the sinusoid was the gravity-dependent component. The dashed lines in E to H are the gravity-independent components. The cartoons on top represent the head orientations relative to the direction of tilt for experimental data shown below.

Gain changes were also determined along other directions of tilt. Before adaptation, the gains were generally uniform over all positions except when head was tilted toward LSD. After the gains had been adapted, there was a significant gain decrease when animal was tilted obliquely toward LSD, and the changes gradually decreased as the head was reoriented toward RSD (Fig. 2B and 2D). The gravity-independent changes were similar to those determined when tested during tilts in the coronal plane (−16.3% LA/RP; −19.8% RA/LP; Fig. 2B and 2D), but the gravity-dependent changes were smaller (9.9% LA/RP; 15.4% RA/LP). Gains for tilts in the sagittal plane after adaptation LSD were lower than before adaptation for all head orientations (Fig. 2C; compare open and filled symbols). There was essentially no gravity-dependent gain change (2.2%) for head positions along this plane, whereas the gravity-independent gain changes (−21.4%) were similar to those obtained in other directions of head tilt (Fig. 2G).

Gravity-dependent and gravity-independent gain changes for experiments in which the gains were increased and decreased in an on-side position are shown in Tables 1 and 2, and experiments in which the gains were decreased in the prone position are shown in Tables 3 and 4. There were significant differences between the three sets of data: coronal, sagittal, and the other two planes combined (P < .0001, ANOVA). When the gain was adapted on-side and then tested using tilts in the coronal plane, the average gravity-dependent gain changes were 16.6 ± 2.4 (Table 1). Changes were smaller and not statistically different from each other (t = 2.31, P = .2) when the head was tilted in the LA/RP (12.9 ± 2.1) or RA/LP planes (10.1 ± 1.9). The gravity-dependent component was negligible when tested during tilts in the sagittal plane (2.5 ± 1.0). Gravity-independent gain changes varied from one experiment to another, but were more uniform for all directions of head tilt (Table 2). There was no difference in the gravity-independent component when the animals were tested in these four planes (P = .45, ANOVA). Thus, the gravity-dependent component varied according to the head position relative to gravity, but the gravity-independent component was similar regardless of the plane of tilt.

TABLE 1.

Gravity-dependent gain change after on-side adaptation

Adaptation	Tilt LSD/RSD	Tilt LA/RP	Tilt RA/LP	Tilt F/B	Monkey ID
G down LSD	15.7	15.4	9.9	2.2	M0102
G down LSD	16.4	14.3	9.3	1.7	M17115
G down RSD	13.3	10.1	8.9	1.8	M0102
G down RSD	19.8	12.4	13.4	2.9	M17115
G up LSD	17.7	12.1	8.8	4.0	M0102
Average	16.6 ± 2.4	12.9 ± 2.1	10.1 + 1.9	2.5 + 1.0

TABLE 2.

Gravity-independent gain change (absolute value) after on-side adaptation

Adaptation	Tilt LSD/RSD	Tilt LA/RP	Tilt RA/LP	Tilt F/B	Monkey ID
G down LSD	20.6	19.8	16.3	21.4	M0102
G down LSD	24.0	26.9	18.4	25.0	M17115
G down RSD	29.9	32.1	24.3	26.6	M0102
G down RSD	35.5	31.2	22.4	24.3	M17115
G up LSD	12.1	15.9	10.1	7.4	M0102
Average	24.4 ± 8.9	25.2 ± 7.1	18.3 ± 5.6	20.9 ± 7.8

TABLE 3.

Gravity-dependent gain change after vertical aVOR gain decrease in prone position

Tilt LSD/RSD	Tilt LA/RP	Tilt RA/LP	Tilt F/B	Monkey ID
1.9	7.0	2.6	9.5	M0102
2.7	5.2	6.5	6.4	M17115
2.3	6.1	4.6	8.0	Average

TABLE 4.

Gravity-independent gain change (absolute values) after vertical aVOR gain decrease in prone position

Tilt LSD/RSD	Tilt LA/RP	Tilt RA/LP	Tilt F/B	Monkey ID
12.3	8.6	10.2	14.3	M0102
19.3	14.9	15.4	18.3	M17115
15.8	11.8	12.8	16.3	Average

The aVOR gain was also decreased in the prone position in two experiments. Although there is insufficient data to perform statistical analyses of gravity-dependent and gravity-independent gain changes observed for the different directions of tilt, the observed gain changes were similar to the changes described above (Tables 3 and 4). Gravity-dependent changes were maximal when the animal was tested using head tilts in the sagittal plane, which included the head orientation in which the aVOR gain had been adapted (8.0). The gain changes were smaller in the LA/RP and RA/LP planes (6.1 and 4.6, respectively), and negligible when tilted in the coronal plane (2.3). Thus, tilting the head in various directions again demonstrated that gravity-independent gain changes were uniform for all directions of head tilt (14.2 ± 2.2), whereas the gravity-dependent gain changes were maximal when the head was tilted along a plane that included the position of the head in which gain had been adapted, minimal for head positions in an orthogonal plane, and intermediate when the head was tilted in either of two intermediate planes.

Gain changes along the four vertical planes in which the animals were tested, which were 45° apart, were reconstructed in three dimensions using a spline interpolation.²¹ Examples of such gain reconstructions for two animals are shown in Figures 3, 4, and 5. In each instance, the gain changes formed a surface, with the maximal change occurring in the position of adaptation, and the changes were reduced as the animal was reoriented away from this position. When the head was tilted along the plane that was orthogonal to the plane that included the position of adaptation, there was no gravity-dependent component across the entire surface, although the gravity-independent component was present. These data indicate that the gravity-dependent gain change after adaptation in a particular head orientation is reflected over the whole range of head positions, regardless of the gravity-independent changes. This supports our hypothesis about the organization of the adaptive mechanisms of the aVOR.¹²

FIGURE 3.

Gain changes for head tilts in various directions, fitted by spline interpolations. The gain changes were determined after each hour of gain decrease in the LSD position in two animals (M0102: A,C,E,G; M17115: B,D,F,H). See text for details.

FIGURE 4.

Scheme as in Figure 3. The gain changes in animal M0102 were determined after each hour of gain increase in the LSD position (A,C,E,G), and after each hour of gain decrease while prone (B,D,F,H).

FIGURE 5.

Scheme as in Figure 3. Gain changes in animals M0102 (A,C,E) and M17115 (B,D,F) were determined after four hours of gain decrease in the RSD position (A,B), and one day (C,D) and two days (E,F) after the adaptation.

Development of the gain changes as a function of the time during which aVOR gain was adapted is shown in Figures 3 and 4. The animals M0102 (Fig. 3A, 3C, 3E, 3G, and Fig. 4) and M17115 (Fig. 3B, 3D, 3F, and 3H) were tested after each hour of adaptation over the 4-h period. Despite some variation, findings were the same for both animals. After one hour of gain decrease LSD, the maximal gain change was observed in the position of adaptation (Fig. 3A and 3B; Fig. 4A and 4B), and the changes were smaller as the animals were reoriented away from this position. The largest gain changes occurred in the first hour, were still substantial in the second hour, and were more limited in the third and fourth hours of adaptation (darkness of gray scale represents depth of the gain changes). In all instances, ~50% to 60% of the gain changes occurred in the first hour, ~25% after the second hour, and were progressively smaller after the third and fourth hours.

We have shown previously that the gain changes induced in humans by only one hour of adaptation¹³ and in monkeys by four hours of adaptation¹⁴ could be observed for 2 to 3 days when the subjects were tested in the plane that included the head position in which gain was adapted. After the aVOR gain was decreased for four hours RSD in M0102 (Fig. 5A) and in M17115 (Fig. 5B), the three-dimensional surface of the gain changes was similar to those obtained after gain decrease while LSD (Fig. 3G and 3H). Changes in gain were maximal RSD (13.3% in M0102 and 19.8% in M17115) and gradually decreased as the head was oriented LSD or toward RA and RP canals. When gains were tested one day after adaptation the gain change surface was bent toward RSD (7% in M0102 and 7.8% in M17115). When the vertical aVOR gain was tested two days after adaptation, the surface still had some decrease in the RSD direction, but the gravity-dependent changes in this plane were only 4.8% in M0102 and 5.5% in M17115. These gravity-dependent gain changes were still statistically significant when animals were tested only during tilts in the coronal plane, but there were substantial deviations in the individual values over the three-dimensional surface.

DISCUSSION

This study provides further evidence for our previous finding^12-14 that there are two components to every vertical aVOR gain adaptation by demonstrating that these two components were present when tested over a wide variety of head orientations. We further demonstrate here that vertical aVOR gains can also be adaptively decreased in prone positions. Although the gravity-dependent components were somewhat smaller than when adaptation was done on-side, the gain changes were significant and had similar three-dimensional characteristics.

In previous studies we demonstrated that gravity-dependent gain changes are a continuous function for head orientations in the coronal plane.^12-14 Zee and colleagues^8,9 were the first to demonstrate that horizontal aVOR adapted in one head orientation had a different gain when tested in different head orientations. We extended this work by showing that there are two components to adaptation in the vertical aVOR, a gravity-dependent and a gravity-independent component. Here we show that the gravity-dependent changes are distributed over the three-dimensional space that includes every head position in which the aVOR gain was tested. After the aVOR gain was adapted in a particular head orientation and then tested with head tilt in planes that included the position in which gain was adapted, both gravity-dependent and gravity-independent components were present. Only gravity-independent components were present when the head was tilted in orthogonal planes, and the gravity-dependent components became proportionally smaller when the animals were tested in intermediate planes. This implies that a gravity-independent component is present in all components of the aVOR whenever the aVOR gain in any direction is adapted. Although the aVOR has been studied extensively, the existence of two components for every gain change has not been obvious, probably because aVOR gains have not been tested outside of the position of adaptation.

That the gravity-dependent and gravity-independent components are separate was also suggested by dual-state adaptation of the vertical aVOR in opposing positions.¹² When the vertical aVOR gain was increased with the animal one side down and decreased in the other side-down position, the gravity-independent component was zero for tilts in the coronal plane, and only the gravity-dependent component was present. Whether the gravity-independent component would reappear for tests using tilts in the sagittal plane is not known, but from our data we would predict that this would not occur.

Another important finding was that the gravity-dependent gain changes persisted for long periods after four hours of adaptation, with some evidence of these changes being present after three days.¹² The animals were not restrained in any way except during testing over this period and moved freely. Similar long-duration changes in aVOR gains were produced by one-hour periods of adaptation in humans.¹³ Thus, the gravity-dependent gain changes are not only universal across all head positions and probably occur in all components of the aVOR across a wide variety of species, they are also long-lasting.

The sites in the vestibular system at which the gravity-dependent components of adaptation are processed are still unknown. Initial predictions that the nodulus and uvula, which receive direct otolith organ input, would be critical sites for gravity-dependent gain changes proved incorrect when ablation of these structures had no effect on this component.¹⁴ The flocculus is known to be critical for gain adaptation²² (see also Ref. 23 for review), but as yet there is no indication that otolith information is processed through the flocculus. Therefore, it is likely that the gravity-dependent gain changes occur in the vestibular nuclei on neurons that receive convergent canal and otolith inputs.^24,25

It is striking that the magnitude of the gravity-dependent gain changes could be as large as the gravity-independent gain changes. This implies that those changes in aVOR sensitivity that are dependent on head position with regard to gravity have an important function. As yet, this function is still unclear, but because the aVOR exists to stabilize the eyes and reduce retinal slip to enable maximal visual acuity during head movement, it seems likely that these gain changes are spatially tuned to improve vision during head movements with the head in positions close to the position in which the aVOR gain was adapted.¹⁰