VISUAL-VESTIBULAR INTERACTION DURING OFF-VERTICAL AXIS ROTATION

Transcription

1 ELSEVIER (95) Journal of Vestibular Research, Vol. 6, No.2, pp , 1996 Copyright 1996 Elsevier Science Inc. Printed in the USA. All rights reserved /96 $ Original Contribution VISUAL-VESTIBULAR INTERACTION DURING OFF-VERTICAL AXIS ROTATION Joseph M. Furman and Juan Carlos Mendoza Department of Otolaryngology, University of Pittsburgh School of Medicine, The Eye & Ear Institute Building, Pittsburgh, Pennsylvania, USA. Reprint address: Dr. Joseph M. Furman, Department of Otolaryngology, University of Pittsburgh School of Medicine, The Eye & Ear Institute Building, 203 Lothrop Street, Suite 500, Pittsburgh, PA 15213, USA 0 Abstract- The aim of this study was to further define the eye movement response to combined visual and vestibular stimulation, especially during linear acceleration. Subjects included 15 asymptomatic healthy individuals (8 females and 7 males) between the ages of 20 and 31 years. Vestibular stimulation consisted of earth-vertical axis rotation and off-vertical axis rotation (OVAR). Visual stimuli consisted of projected vertical stripes that were rotated for optokinetic trials and stationary for visual augmentation trials. A small laser target (0.5 mw, 0.5 arc) that rotated with the subject was used for fixation trials. Eye movements were measured with electro-oculography. Results showed that visual-vestibular interaction during sinusoidal rotation was not affected by a 15 off-vertical tilt. Constant velocity OVAR induced a continuous nystagmus whose slow component velocity contained a nonzero baseline, that is, a bias, and a periodic fluctuation at the rotation frequency, that is, a modulation component. The modulation component during visual fixation was reduced as compared with that seen during rotation ir.: the dark, but was not absent. Consta.nr OVAR hr the presence of earth-fixed stripes induced. a consistent sinusoidal modulation. Our results suggest that visual-vestibular interaction for otolith stimulation differs from visual-vestibular interaction for semicircular canal stimulation. The modulation component of the response to OVAR appears to be Presented at the Eighteenth Annual Meeting of the Association for Research in Otolaryngology, St. Petersburg Beach, Florida, USA, February modified by visual stimulation to a lesser extent than other vestibular-induced eye movements and thus may reflect a more "direct" vestibulo-ocular response. The bias component of the response to OVAR can be substantially influenced by vision and thus may depend upon more "indirect" pathways. 0 Keywords- vestibulo-ocular reflex; otolith; human. Introduction Integration of multisensory inputs is essential to maintain body orientation in space. In particular, current models of ocular motor control are based on the principle that visual and vestibular signals are combined by the central nervous system to elicit the necessary motor responses to control eye movement (1,2). Previous studies of visual-vestibular interaction in humans have concentrated on veemior. ger1erated by a moving visual surround (3); on the effect of rotational stimuli on optokinetic nystagmus ( 4); and on the ability of subjects to suppress the vestibula-ocular reflex (VOR) using vision (5-7). VVI studies that examine eye movement responses have mostly used angular acceleration to assess the interaction between semicircular canal stimulation and visual stimulation. These studies have, in large part, supported RECEIVED 26 June 1995; ACCEPTED 7 August

2 94 the basic premise set forth in current models of the ocular motor system (1,2,8) that visual and vestibular stimuli are combined in an additive fashion by the central nervous system. Recently, the interaction between linear acceleration and visual stimuli have yielded results that do not fully support models designed to explain results from angular acceleration studies. In particular, when a horizontal optoki :Jetic stin1ulu.s at r:onstant velocity with respect to the subject is combined with sinusoidal linear acceleration along the subject's interaural axis (9-11), the resultant eye movement response exhibits not only additive characteristics but also a component that suggests a nonlinear visual-vestibular interaction. Also, vertical sinusoidal optokinetic stimuli interact in a nonlinear manner with sinusoidal linear acceleration along the rostral-caudal body axis (12). When visual and vestibular stimuli are mutually orthogonal, a complex interaction between linear acceleration and optokinetic stimulation is evident (10). A study by Wall and Furman published in 1990 (13) examined the influence of combining visual stimuli with constant velocity earthhorizontal axis rotation, a linear acceleration of constant magnitude and changing direction. That study also indicated that VVI for linear acceleration differs from that found with angular acceleration. Specifically, during earthhorizontal axis rotation at constant velocity while viewing a lighted visual surround, eye velocity displayed a significant sinusoidal variation, despite rotation at constant velocity. The present study assessed VVI during offvertical axis rotation (OVAR), a stimulus that is similar to earth-horizontal axis rotation in that it consists of a linear acceleration of constant magnitude and changing direction, but differs in that, operationally, the angle of offvertical tilt for OVAR is considerably less than 90. OVAR was selected as a vestibular stimulus to further our understanding of VVI because pure linear acceleration, pure angular acceleration, and combined linear and angular acceleration stimuli can be delivered using constant velocity OVAR, sinusoidal OVAR with no tilt (that is, earth-vertical axis rotation), and sinusoidal OVAR with a nonzero tilt J. M. Furman and J. C. Mendoza angle, respectively. The aim of our study was to further define the eye movement response during combined visual and vestibular stimuli, especially when the vestibular stimulation consisted in whole or in part of a linear acceleration, which necessarily stimulates the otolith organs. These studies were designed to compare directly VVI during angular acceleration with VVI during linear acceleration, using earth-fixed and head-fixed visual stimuli. T11is will serve as.a baseline for future studies of VVI wherein the visual stimuli are moving targets that evoke particular types of eye movement such as smooth pursuit or saccades. Methods All procedures were approved by the Biomedical Review Board of the University of Pittsburgh, and informed consent was received from all subjects. Subjects included 15 asymptomatic healthy individuals (8 females and 7 males) between the ages of 20 and 31 years (mean: 23.9 years). Exclusionary criteria for the study included a history of a neurologic or otologic disorder or abnormalities on any of the following tests according to laboratory norms: ocular motor battery, positional testing, caloric testing, earth-vertical axis rotation (EVAR) testing in darkness, and moving platform posturography (Equitest ). The test protocol used an OVAR chair, described elsewhere (14). Briefly, the OVAR chair consists of a rotation chair affixed to a computer-controlled 80 foot-pound (11 0 Nm) turntable attached to a tilt stand. A cylindrical enclosure with a radius of 1 m surrounded the test chair /turntable and was mounted to the tilt platform. Thus, any tilt of the axis of rotation was accompanied by a tilt of the enclosure such that the distance between the subject and the enclosure, onto which visual stimuli were projected, stayed constant. A head restraint device ensured minimal relative motion between the head and the chair. Subjects wore insert earphones that allowed communication and masked localizing auditory cues.

3 Visual-Vestibular Interaction during OVAR Optokinetic trials and trials that combined vestibular stimulation with an earth-fixed visual stimulus used light/dark vertical stripes (0.1 cycles per degree frequency) projected onto the enclosure from an overhead lamp that was affixed to a servomotor attached to the roof of the enclosure. For trials that combined a head-fixed target with vestibular stimulation, a small laser target (0.5 mw, 0.5 arc) was projected from the rotation chair onto the enclosure. Eye movements were measured with Decoupled electro-oculography using an optically isolated preamplifier (Ten1plin Engineering, Laytonville, CA) and an amplifier with a lowpass filter whose cutoff frequency was 40Hz. Eye movements were recorded on a chart recorder and sampled by a PDP 11/73 computer using a 12 bit analog-to-digital converter at a sampling rate of 100 Hz. The horizontal electro-oculography signal was calibrated by asking each subject to gaze at targets placed 10 to the left and right of center. Vertical eye movements were recorded only as ag aid in detecting blink artifacts in the horizontal traces. To ensure subject alertness, prerecorded trivia questions were presented through the earphones and subjects were instructed to answer them throughout each trial. All subjects were tested in each of 12 conditions by combining 4 visual conditions, and 3 rotational profiles (Table 1). Tilt Angles: In order to compare responses generated by semicircular canal stimulation alone with responses generated in whole or in part by otolith organ stimulation, subjects were tested both with EVAR, that is, a tilt angle of zero, and with OVAR at a tilt of 15. Table 1. Test Conditions Visual conditions VOR VVOR VOR-Fix OKN Rotational profiles Sinusoidal EV AR Sinusoidal OV AR Constant Velocity OVAR A tilt angle of 15 o produces a maximum linear acceleration of 0.26 g. This angle was chosen since it was thought to be sufficient to generate substantial otolith organ stimulation without provoking unmanageable motion intolerance. Visual Conditions: Four visual conditions were used: 1) Rotation with eyes open in the dark (VOR); 2) Rotation while subjects viewed earth-fixed stripes projected against the enclosure (VVOR), 3) Rotation while subjects fixated a small laser target that rotated with them (VOR-Fix), and 4) No rotation while subjects were asked to "look" at moving projected stripes (OKN). Rotational Profiles: For EVAR trials, regardless of visual condition, the chair was rotated sinusoidally for 10 cycles at 0.05 Hz with a peak velocity of 50 /s. Similarly, for OKN trials while upright, the projected stripes were rotated for 10 cycles at 0.05 Hz with a peak velocity for 50 /s. For sinusoidal OVAR trials, regardless of visual condition, the chair was rotated for 10 cycles at 0.05 Hz with a peak velocity of 50 /s. Similarly, for OKN trials while the subject was tilted off-vertical in either the right-ear-down or the left-ear-down position, the projected stripes were rotated for 10 cycles at 0.05 Hz with a peak velocity of 50 /s. Sinusoidal OVAR trials and OKN trials while tilted were performed with the axis of rotation tilted throughout the trial For constant velocity OVAR trials, regardless of visual condition, chair rotation began with no off-vertical tilt. Once semicircular canal responses decayed, after about 60 s, the axis of rotation was tilted to produce OVAR. For each subject, constant velocity OVAR was performed ir both directions for VOR testing. ir:: the VVOR and VOR-Fix conditions, only one direction of constant veiocity (randomized) was used. To simplify further discussion, all constant velocity results will be presented as if they were clockwise. Horizontal eye movement data were analyzed using a nystagmus analysis program developed in our laboratory based on an algorithm described by Wall and colleagues in 1981 (15). After both automatic and manual 95

4 96 desaccading of the raw eye position data, the program generated slow component velocity (SCV) records that were fitted using a nonlinear iterative least-squares regression algorithm to generate parametric measures of the responses. For EVAR and sinusoidal OVAR trials, the first cycle of data was discarded to avoid any transient response. Remaining cycles were fit on a cycle-by-cycle basis using: scv(t) = A sin(21f/ + where A is the amplitude of the sinusoidal response, 1> is the phase, andf is the frequency of the sinusoidal rotation, which in this case is 0.05 Hz. A phase of oo describes a perfectly compensatory response. Analysis of constant velocity OVAR responses began approximately 2 s after tilt was completed. For constant velocity trials, the SCV during OVAR was assumed to be composed of a bias and a modulation component. Thus, on a cycle-by-cycle basis (that is, responses during each 360 rotation), we fit the following equation to each data cycle: scv(t) =bias+ modamp sin(27r/t + ). Where bias is the magnitude of the bias component during a particular cycle, modamp is the magnitude of the modulation component during that particular cycle, Tis 6 s because the rotational velocity was 60 /s, and 1> is the modulation component phase for a particular cycle (see below). Phase for constant velocity trials will be reported using a convention based on the equivalent purely linear acceleration along the interaural axis experienced during OVAR (see 16). We assumed that the per-rotatory sinusoidal horizontal responses during constant velocity OVAR could be interpreted as responses to interaural stimulation on a linear sled (Figure 1) where minimum acceleration (peak linear velocity) takes place at the center of the sled. Such an assumption was made so that our data could be compared with data of others who used linear acceleration of constant direction and changing magnitude, that is, a linear sled. Although OVAR consists of linear acceleration of constant Sinusoidal Linear Sled Motion Clockwise Otf-Yerta:al Axis Rotation A. J. M. Furman and J. C. Mendoza.B. ND 1 NU 1 Figure 1. Diagram of interaural linear acceleration experienced during constant velocity off-vertical axis rotation (OVAR). During constant velocity OV AR, the projection of the rotating gravity vector onto the interaural axis describes a sinusoidally varying linear acceleration. This concept can be understood by comparing the orientations of a subject during a clockwise (CW} constant velocity OVAR revolution to an equivalent cycle of sinusoidal linear sled motion, that is, a pure linear acceleration. Assume that the equivalent cycle of sled motion starts with the sled at the right end of track, which is equivalent to the right ear down (RED} position (A}. Since the OVAR is CW, the subject then moves toward the nose down (NO} position, which is equivalent to reaching peak velocity to the left in the center of the sled (B). The time at which the sled reaches the left end of the track (C) and starts traveling to the right is equivalent to the left ear down (LED) position in the OVAR chair. Finally, the time at which the sled reaches the center of the track traveling at maximum velocity to the right is equivalent to the OVAR chair reaching the nose up (NU) position (0}. magnitude and changing direction, the projection of the linear acceleration vector on the interaural axis has a sinusoidally changing magnitude for constant velocity OVAR. Thus, a direct comparison with responses to pure interaurallinear acceleration can be performed. A phase of 0 for a response to pure interaural linear acceleration corresponds to perfect compensation (that is, when the subject moves to the left, the eyes move to the right) (Figure 2.) Similarly, because during OVAR subjects reach minimum equivalent interaural linear acceleration (peak equivalent linear velocity) in the nose up and nose down positions, perfect compensation requires that the velocity of the sinusoidal modulation component peaks at the nose up and nose down positions, with the direction of rotation defining the direction (sign) of the peak: for clockwise rota-

5 Visual-Vestibular Interaction during OVAR 97 R Equivalent Interaural Stimulus Velocity Slow Component Velocity L R L R Slow Component Velocity L R Slow Component Velocity L 30 Phase Lead RED ND LED NU RED ND OVAR Position Figure 2. Diagram of computation of modulation component phase. Modulation component phase is reported with respect to the equivalent interaural stimulus velocity (A) (see Figure 1 ). The response shown in (B) is perfectly compensatory, that is, the modulation component velocity is the inverse of the equivalent stimulus velocity. By definition, this response has a phase of 0. For example, for a clockwise constant velocity OVAR rotation, perfect compensation (phase = 0 ) refers to a modulation component that peaks to the left at the Nose Up (NU) position, which is equivalent to the peak rightward velocity on a linear sled at the center of the track (see Figure 1 ). Panels C and D illustrate responses that lag or lead a perfectly compensatory response and are thus assigned phase values of -30 or + 30, respectively. tions, eye velocity should peak to the left at nose up and peak to the right at nose down; for counterclockwise rotations, eye velocity should peak to the right at nose up and peak to the left at nose down. Results Sinusoidal Response Eye movement responses during sinusoidal stimuli were not affected by a 15 o tilt of the

6 98 chair's axis of rotation (Figures 3 and 4). The combination of an earth-fixed visual stimulus and a vestibular stimulus (VVOR) increased the gain of the response to about 1. 0, whereas purely visual (OKN) and purely vestibular (VOR) gains remained close to 0.5 (Figure 4). All subjects were able to effectively suppress their VOR using a head-fixed visual target, that is, the gain of responses during the VOR Fix condition was negligible. Constant Velocity Responses As expected, constant velocity OVAR induced a continuous nystagmus whose SCV contained a nonzero baseline, that is, the bias, and a periodic fluctuation at the rotation frequency, that is, the modulation component. Figure 5 shows the eye movement response during clockwise constant velocity OVAR trials. Figure 6 presents the mean values of response parameters for all subjects, with the exception of the mean value for modulation component phase for VOR-Fix. For this parameter, we averaged values from 11 subjects because 4 subjects had responses that were so small ( < 2 o Is peak-to-peak) that phase values could not be computed. The bias component for the VOR was small and variable. Seven subjects showed a consistently compensatory response (bias to the left during CW rotations and bias to the right during CCW rotations), 5 subjects had anticompensatory responses to one direction of rotation, 1 subject had anticompensatory responses in both directions, and 2 subjects had essentially no bias. Mean velocity during both VVOR and OKN trials was variable, but generally had a value of about half the stimulus velocity. No significant bias velocity was observed during VOR-Fix trials. In all 3 protocols that involved constant velocity off-vertical axis rotation (VOR, VVOR, VOR-Fix), all subjects showed a consistent sinusoidal modulation with an amplitude of about 1 to 5 /s (see Figure 5 and 6). Eye velocity peaked close to the left-ear down position for clockwise rotations. Surprisingly, the amplitude of the modulation component J. M. Furman and J. C. Mendoza tended to be larger during VVOR than during rotation in the dark. That is, despite constant velocity rotation with a lighted earth-fixed visual surround, each of the subjects had a nonconstant slow component velocity. Also, although the modulation component during VOR-Fix was reduced as compared to that seen during rotation in the dark (VOR), it was still present. While our analysis concentrated on the s1ow responses, we analyzed the eye position response during VOR Fix to establish a baseline for future studies in which we will use a moving target to generate saccadic and smooth pursuit eye movements during vestibular stimulation. As seen in Figure 5E, a small but consistent sinusoidal modulation of eye position was induced by constant velocity OVAR while the subject viewed a head-fixed target. A nonnegligible, that is> 2 peak-to-peak, position modulation was observed in 11 of 15 subjects, suggesting that they were unable to suppress completely their otolith-ocular reflex with vision. We were concerned that the observed position modulation was artifactual, that is, a result of lateral head movements with respect to the target during constant velocity OVAR. Thus, we re-tested 6 subjects using a biteboard custommade for each subject to eliminate relative movement of the head with respect to the visual target. The visual target for these re-tests consisted of a 2.5-mm light-emitting-diode placed at the end of a m-long rod attached to a bite board. Three of the 6 re-tested subjects had previously shown a nonnegligible position modulation during VOR-Fix for constant velocity OVAR (peak-to-peak modulation: 3.6 ± 1.0 ). These 3 subjects again showed a peak-to-peak modulation of eye position during VOR-Fix larger than 2 (peakto-peak modulation: 4.2 ± 1.0 ). The other 3 subjects, who had not shown a consistent modulation in the VOR-Fix condition using our standard head restraint, again showed no consistent position modulation. This re-test of 6 subjects confirmed that the eye position modulation seen during constant velocity OV AR in the VOR-Fix condition was not artifactual. However, since electrooculography does not

7 Figure 3. Slow component eye velodty during sinusoidal stimulus protocols. Panels A through D illustrate the slow component velocity (SCV) responses for one subject during sinusoidal earth-vertical axis rotation (EVAR) in 4 visual conditions. Panels F through I illustrate SCV responses during sinusoidal off-vertical axis rotation (OVAR). Panels E and J show the stimulus velocity. EVAR OVAR 0 <1.) > <!.);;-. ;>-, UJ > &' CD CD :E, z <!);;-. ;.-,'--' 0 UJ z- E Time(s)

8 100 J. M. Furman and J. C. Mendoza EVAR :i: A cs ti J...J o :j = B:l..c:: I D l o... Q , VOR VVOR OKN VOR-Fix OVAR ! d , c:: "' o W D i I I I I :::::::::::::::::::::::::::::::o::::::::::::--::::9:::::::::::::::::::::::::::::: i -I 0 -'----.r------r r-----' VOR VVOR OKN VOR-Fix D Figure 4. Gain and phase data for responses to sinusoidal stimuli. Data are pooled across all 15 subjects. Error bars indicate one standard deviation. Phase values are not reported for VOR-Fix trials because the magnitudes of the responses in these conditions were negligible. Note that gain and phase responses to earth-vertical axis rotation and off-vertical axis rotation are virtually identical. provide absolute eye position, we could not exclude the possibility that some of the subjects were not gazing directly at the fixation target. Discussion Our results suggest that visual-vestibular interaction during otolith stimulation differs from that during semicircular canal stimulation. This is only evident for constant velocity OVAR, a pure otolithic stimulus, wherein many subjects were unable to entirely suppress otolith-induced eye movements, and a sinusoidal modulation of eye velocity was observed during constant velocity rotation. In contrast, during sinusoidal rotations at 0.05 Hz, visualvestibular interaction did not appear to be affected by the presence of dynamic otolith stimulation; both gain and phase during sinusoidal OVAR remained nearly identical to that measured during sinusoidal EVAR. Our results regarding visual-vestibular interaction support the idea that the bias component and the modulation component of the eye movement response to constant velocity OVAR are mediated by different mechanisms (17). The bias component of the response to constant velocity OVAR in a lighted visual surround (VVOR) was increased significantly from its small value during rotation in the dark (VOR) and was decreased essentially to zero when rotation was combined with a head-fixed target (VOR-Fix). On the contrary, the modulation component of the response to constant velocity OVAR was consistently present throughout all conditions. Thus, it appears that the visual system has less of an influence on the modulation component than on either the bias component or semicircular-canalinduced eye movements. The basis for this difference in visual-vestibular interaction may be that the modulation component is a more "direct" vestibula-ocular response whereas the bias component and low frequency angular VOR-induced eye movements depend upon more "indirect" pathways and are thus subject to a more significant influence from vision. Responses to constant velocity OVAR have been conceptualized as equivalent to the response to combined (but out of phase) sinusoidal interaural and naso-occipitallinear acceleration (16). In this formulation, horizontal eye movements, such as those measured in this study, would be expected to have dynamics (phase relationships) similar to those of the eye movements induced by pure interaural acceleration, such as that produced by a linear sled. Indeed, we report constant velocity OVAR modulation component phase data based upon such a formulation (see Figures 1 and 2). Modulation component magnitude is a more difficult measure to compare across stimuli because of its dependence upon

9 Visual-Vestibular Interaction during OVAR 101.., -... ::..J A...J B IJ..JI I c; _:0 N D ll...l, 30> F ' l' G z 0 VJ ;e "".c U c 0 ;;.., IJ.l l \: iii. : k; :/; J:d< f\ :iai; t;'t.; ;;;. 1" I. I It ' Nose Down Nose Up Nose Down Time(s) Figure 5. Eye movement during constant velocity off-vertical axis rotation. Panels A through H illustrate eye position and slow component eye velocity responses for one subject during clockwise constant velocity offvertical axis rotation (OVAR) in 4 visual conditions. Note the large sinusoidal (modulation) component during VVOR and the clearly discernible position fluctuation during VOR-Fix. Panel I indicates OVAR chair position. Note that eye position is shown graphically as arbitrarily having an average value of 0 because absolute eye position cannot be determined using electro-oculography. point of regard and vergence angle. Our phase data, which agree with previous data obtained during earth-horizontal axis rotation (18,19), show a phase lead that is advanced when compared to the phase observed in responses to pure linear acceleration without a rotating gravity vector (10,20). Possibly, this difference in phase is related to the equivalent frequency

10 B ] I 1 -; ± I:: - l &..... j r mm lm - ml CW CCW VOR-Fix VVOR OKN VOR Figure 6. Bias and modulation component data for responses to constant velocity off-vertical axis rotation. Data are pooled across all 15 subjects, except that for VOR-Fix, modulation component phase could only be estimated in 11 of 15 subjects because of negligible modulations in the other 4 subjects. Response parameters for clockwise (CW) and counterclockwise (CCW) rotation in the dark (VOR) are shown separately. range of most OVAR studies (including the present study) versus most linear sled studies. The equivalent frequency in the present study was 0.16 Hz, below the frequencies used in most linear sled experiments-about 0.2 to 5.0 Hz. Although this study focused upon visualvestibular interaction, the sinusoidal rotations in the dark allowed a comparison of semicircular canal-ocular responses with and without combined otolithic stimulation. Our data indicate that at the frequency (0.05 Hz) and J. M. Furman and J. C. Mendoza amplitude (50 /s peak velocity) tested, a dynamic otolith stimulus affected neither the gain or phase of the angular VOR nor the gain or phase of responses to combined visual-vestibular stimuli (VVOR, VOR-Fix). Had changes in gain or phase been observed, we would have considered them a result of the influence of semicircular canal-otolith interaction on the velocity storage system, a hypothetical neural circuit that perseverates the response to vestibular or optokinetic stimulation by lengthening the dominant time constant of the ocular response. This integrator is thought to receive inputs fr01n the semicircular canals and the visual system (1,2) as well as from both static and dynamic otolith signals (21,22). Otolithic signals during constant velocity OVAR are thought to "charge" velocity storage, leading to a longer time constant of per-rotatory responses when using a tilt-then-rotate paradigm (14,21). In contrast, static otolith inputs, such as those encountered when stopped following constant velocity OVAR, are thought to "discharge" or "dump" velocity storage, leading to a shorter post-rotatory time constant. Our negative, results regarding semicircular canalotolithic interactions can be explained by assuming that at a frequency in the vicinity of 0.05 Hz and a tilt angle of 15, velocity storage charging and dumping either are inactive or cancel one another. Acknowledgments-The authors wish to thank Anita Lieb and Georgina Peters for technical assistance. This work was supported in part by the Air Force Office of Scientific Research, Grant No. F l-0261, and by grant #DC01791 from the National Institutes of Health. REFERENCES 1. Raphan T, Cohen B, Matsuo V. A velocity storage mechanism responsible for optokinetic nystagmus (OKN), optokinetic after-nystagmus (OKAN) and vestibular nystagmus. In: Baker R, Berthoz A, eds. Control of gaze by brain stem neurons. New York: Elsevier /North-Holland Biomedical Press; 1977: Robinson DA. Vestibular and optokinetic symbiosis: an example of explaining by modelling. In: Baker R, Berthoz A, eds. Control of gaze by brain stem neurons. New York: Elsevier/North-Holland Biomedical Press; 1977: Dichgans J, Brandt T. Visual-vestibular interaction and motion perception. In: Dichgans J, Bizzi E, eds. Cerebral control of eye movements and motion perception. Basel: Karger; 1972; Koening E, Allum JHJ, Dichgans J. Visual-vestibular interaction upon nystagmus slow phase velocity in man. Acta Otolaryngol. 1978;85: Guedry FE. Relations between vestibular nystagmus

11 Visual-Vestibular Interaction during OVAR 103 and visual performance. Aerospace Med. 1968;39: Schmid R, Zambarbieri D, Magenes G. Modifications of vestibular nystagmus produced by fixation of visual and nonvisual targets. Ann N Y Acad Sci. 1981;374: Baloh RW, Jenkins H, Honrubia V, Yee R, Lau C. Visual-vestibular interaction and cerebellar atrophy. Neurology. 1979;29(1): Lau C, Honrubia V, Jenkins H, Yee R. Linear model for visual-vestibular interaction. Aviation Space Environ Med. 1978;49: Buizza, A, Leger A, Droulez J, Berthoz A, Schmid R. Influence of otolithic stimulation by horizontal linear acceleration on optokinetic nystagmus and visual motion perception. Exp Brain Res. 1980;39: Mendoza JC, Merfeld DM. The interaction of constant-velocity optokinetic nystagmus and the linear vestibula-ocular reflex in humans. Society for Neurosciences, 1993; Abstract 143.3, 19(1): Merfeld DM, Christie JRI, Young LR. Perceptual and eye movement responses elicited by linear acceleration following spaceflight. Aviation Space Environ Med. 1994;65: Lathan CE, Wall C, Harris LR. Human eye movements response to z-axis linear acceleration: the effect of varying the phase relationships between visual and vestibular inputs. Exp Brain Res. 1995; 103(2): Wall C, Furman JM. Visual-vestibular interaction in humans during earth-horizontal axis rotations. Acta Otolaryngol. 1990; 109: Furman JM, SchorR, Schumann T. Off-vertical axis rotation: a test of the otolith-ocular reflex. Ann Otol Rhinal Laryngol. 1992; 101: Wall C, Black F. Algorithms for the clinical analysis of nystagmus eye movements. IEEE ;BME-28: Paige G, Tompko D. Responses to linear head motion in the squirrel monkey; 1: Basic characteristics. J Neurophysiol. 1991;65: Hain T. A model of the nystagmus induced by offvertical axis rotation. Bioi Cybern. 1986;54: Benson AJ, Bodin MA. Interaction of linear and angular accelerations on vestibular receptors in man. Aerospace Med. 1966;37: Wall C, Furman JM. Nystagmus responses in a group of normal humans during earth-horizontal axis rotation. Acta Otolaryngol. 1989;108: Skipper JJ, Barnes GR. Eye movements induced by linear acceleration are modified by visualization of imaginary targets. Acta Otolaryngol (Suppl). 1989; 468: Raphan T, Cohen B. Effects of gravity on rotatory nystagmus in monkeys. Ann N Y Acad Sci. 1981; 374: Waespe W, Cohen B, Raphan T. Dynamic modification of the vestibula-ocular reflex by the nodulus and uvula. Science. 1985;288: