- Email: [email protected]
Vision and vestibular adaptation
Vision and vestibular adaptation JOSEPH L. DEMER, MD, PhD, and BENJAMIN T. CRANE, BS, Los Angeles, California
This article summarizes six recent degree-of-freedom studies of visual-vestibular interaction during natural activities and relates the findings to canalotolith interactions evaluated during eccentric axis rotations. Magnetic search coils were used to measure angular eye and head movements of young and elderly subjects. A flux gate magnetometer was used to measure three-dimensional head translation. Three activities were studied: standing quietly, walking in place, and running in place. Each activity was evaluated with three viewing conditions: a visible target viewed normally, a remembered target in darkness, and a visible target viewed with ´2 binocular telescopic spectacles. Canal-otolith interaction was assessed with passive, whole-body, transient, and steady-state rotations in pitch and yaw at multiple frequencies about axes that were either oculocentric or eccentric to the eyes. For each rotational axis, subjects regarded visible and remembered targets located at various distances. Horizontal and vertical angular vestibuloocular reflexes were demonstrable in all subjects during standing, walking, and running. When only angular gains were considered, gains in both darkness and during normal vision were less than 1.0 and were generally lower in elderly than in young subjects. Magnified vision with ´2 telescopic spectacles produced only small gain increases as compared with normal vision. During walking and running all subjects exhibited significant mediolateral and dorsoventral head translations that were antiphase locked to yaw and pitch head movements, respectively. These head translations and rotations have mutually compensating effects on gaze in a target plane for typical viewing distances From the Jules Stein Eye Institute, Departments of Ophthalmology and Neurology, University of California, Los Angeles. Supported by grants from the National Eye Institute (DC-01404, DC02952, EY-08656, and core grant EY-00331) and Medical Scientist Training Program (GM-0804). Joseph L. Demer was the recipient of a Research to Prevent Blindness Lew R. Wasserman Merit Award. Presented at the UCLA Conference on Vestibular Adaptation, Santa Monica, Calif., May 23-25, 1996. Reprint requests: Joseph L. Demer, MD, PhD, Jules Stein Eye Institute, 100 Stein Plaza, UCLA, Los Angeles, CA 90095-7002. Copyright © 1998 by the American Academy of Otolaryngology– Head and Neck Surgery Foundation, Inc. 0194-5998/98/$5.00 + 0 23/1/87913 78
and allow angular vestibulo-ocular reflex gains of less than 1.0 to be optimal for gaze stabilization during natural activities. During passive, whole-body eccentric pitch and yaw head rotations, vestibuloocular reflex gain was modulated as appropriate to stabilize gaze on targets at the distances used. This modulation was evident within the first 80 msec of onset of head movement, too early to be caused by immediate visual tracking. Modeling suggests a linear interaction between canal signals and otolith signals scaled by the inverse of target distance. Vestibulo-ocular reflex performance appears to be adapted to stabilize gaze during translational and rotational perturbations that occur during natural activities, as is appropriate for relevant target distances. Although immediate visual tracking contributes little to gaze stabilization during natural activities, visual requirements determine the performance of vestibulo-ocular reflexes arising from both canals and otoliths. (Otolaryngol Head Neck Surg 1998;119:78-88.)
T
he human head undergoes incessant motion. Angular head movements occur constantly in daily life because of transmitted heartbeat, tremor, sway, ambulation, and vehicular travel.1 Even when a person is attempting to stand completely at rest, involuntary angular head movements occur about the vertical axis (yaw), the interaural axis (pitch), and the anteroposterior axis (roll).2 Such head movements during quiet standing have root-mean-square (RMS) velocities of around 1 degree/second and commonly occur at frequencies above 1.0 Hz, where smooth visual-tracking mechanisms have significant performance limitations.3,4 During quiet standing these movements typically have peak components at frequencies exceeding 5 Hz. Larger head velocities are encountered in all three rotational axes during gentle walking in place,2,5 with predominant frequencies sometimes exceeding 2.5 Hz for yaw and 8 Hz for pitch.6,7 Natural activities are also characterized by complex patterns of head rotation and translation during the viewing of objects at a variety of distances from the eyes. Vigorous activities such as running or hopping produce still greater head velocities. During these head movements the vestibulo-ocular reflex (VOR) serves vision by generating reflex eye movements that stabilize images on the retina. This
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
image stabilization is critical to vision because angular image motion exceeding about 2 to 3 degrees/second degrades visual acuity.8 Whereas the human angular VOR has been studied extensively for passive head rotations and self-generated head rotations about a single axis, the response of the VOR to head translations has been examined less extensively. For pure head rotations about the center of the eye, an ideal VOR would always generate an eye movement of equal magnitude and opposite direction to the head movement. This is the situation most commonly studied in the laboratory. Consideration of translation adds complexity to the issue because, unlike the case of pure rotation, target distance is a significant determinant of effective gaze stabilization. Further, there are two eyes, neither of which is generally colocated with the axis of rotation of the head, so that natural head movements normally consist of combinations of head rotation and translation. This article addresses the operation of the VOR in gaze stabilization during the combined rotational head movements that occur during standing and ambulation and relates this to the interactions of the semicircular canals and otoliths during off-axis rotations. ANGULAR VOR DURING NATURAL ACTIVITIES
Refinements of the magnetic search coil technique7,9 permit the recording of angular VORs during natural activities not associated with perceptible head movements. Reference magnetic field coils 2 M in diameter create a sufficiently homogeneous central field that the signal representing the angular position of the search coil is not appreciably dependent on translation within a substantial central region.9 Measurements during standing and walking were made in 9 normal young (average age 29 years) and 11 normal elderly (average age 78 years) subjects. Written informed consent was secured after approval of all protocols by the institutional review board for the protection of human subjects. Subjects had normal corrected vision and had no known vestibular pathology. Sense coils were placed on one of the patient’s eyes while he or she was under topical anesthesia (Skalar Medical, Delft, The Netherlands) and at the front of a snug headband for measurement of angular gaze and head positions, respectively. Each subject stood in the center of the magnetic field reference cube. Testing was performed for all combinations of five activities, two viewing conditions, and interleaving two conditions of illumination. Activities consisted of (1) standing motionless with heels and toes together; (2) walking in place so that each footfall was synchronized with a metronome beating at 2 Hz; (3) running in place so that each footfall was synchronized with a metronome beating at 3 Hz (young subjects only); (4)
DEMER and CRANE
79
Fig. 1. Head velocity in pitch and yaw during quiet standing and walking in place, for young and elderly subjects. Elderly subjects have higher angular head velocity than young subjects in both axes.
active, single-axis sinusoidal head movement at 0.8 Hz in pitch paced by frequency modulation of an audible tone; and (5) active, single-axis sinusoidal head movement at 0.8 Hz in yaw paced by frequency modulation of an audible tone. For walking and running in place we used the same gait frequencies chosen by Takahashi et al.,10 which corresponded well to the rates of uncoached gait reported by Grossman et al.7 For each activity subjects regarded a high-contrast target located at standing eye level on the wall 6 to 9 M distant. Viewing conditions consisted of (1) unmagnified viewing with refractive correction if necessary and (2) ´2 binocular telescopic spectacles with a magnified field of 16.8 degrees. The unmagnified visual field peripheral to the telescopes was masked. For each combination of activity and viewing condition, measurements were interleaved first under normal indoor illumination (visually enhanced VOR [VVOR]) and then again in complete darkness with instructions to look at the remembered target location (VOR). For each of the test conditions, values of RMS angular head velocity, frequency of the peak Fourier component of angular head velocity, and VOR or VVOR gain were each obtained in duplicate trials. Duplicate measurements were averaged for subsequent analysis, which was performed with analysis of variance and covariance.11 Gaze and head positions were low-pass filtered over a bandwidth of 0 to 100 Hz before digital computer sampling with 16-bit precision at 400 Hz. Eye position in the head was obtained by subtraction of head position from gaze position. After automated removal of saccades,3 VOR and VVOR gains were taken to be the least squares slope of the plot of slow-phase eye veloc-
80
DEMER and CRANE
Otolaryngology– Head and Neck Surgery July 1998
Fig. 2. Eye and head velocities illustrating angular vertical VOR of normal young subject during standing, walking, and running in place. Upward velocities are positive. VOR gain and RMS head velocity (0 to 20 Hz) are indicated for each activity. Note 10-fold greater velocity sensitivity during standing than during walking and running. Top row, Gains are defined by linear regression of eye velocity on head velocity, with analysis of points lying within noise-defined limits of the fit (parallel lines). Saccades were truncated at around 2, 25, and 35 degrees/second for the three activity conditions, respectively, and these data points are visible outside the acceptance limits of the fits. Spontaneous 1 degree/second downward bias present for all activities reflects a drift common in normal subjects in darkness.
ity against head velocity. The RMS value of head velocity was computed for pitch and yaw. Head velocity was also subjected to Fourier transformation to obtain the frequency of the peak component in pitch and yaw. Understanding of the VOR during natural activities understanding of the characteristics of head motion that serve as the stimulus to the reflex. Neither the presence of a visible target nor the presence of spectacle magnification significantly influenced rotational or translational head motion when data were averaged across young or elderly subjects, so it was appropriate to pool data across viewing conditions. Predominant frequency components on Fourier analysis of head velocity were related to harmonics of the rate of footfalls. The observed RMS (0 to 20 Hz) velocities for angular head motion in roll, pitch, and yaw are summarized in Fig. 1. Angular head velocity in young standing subjects for both pitch and yaw exceeded by sixfold to eightfold the instrumentation noise in our search coil system (approximately 0.1 degree/second, 0 to 20 Hz) and was greater in the elderly. As expected, angular head velocity was greater for walking than standing (Fig. 1).
Interestingly, head velocity was 15% to 20% greater in the elderly than the young subjects for both pitch and yaw during both standing and walking in place (p < 0.01). Young subjects had even greater angular head velocities during running than during walking in place, but because few of the elderly subjects were able to run in place, quantitative comparisons of the two groups were not performed for this condition. Physiologic angular head instability thus provides a sufficient stimulus for reliable measurement of VOR gain even in subjects attempting to stand motionless. Using these techniques, we were able, under both lit and darkened conditions, to observe eye movements that were directed in an opposite direction from angular head movements during all activities studied, as illustrated for pitch in Fig. 2. Saccades and quick phases were occasionally present and were removed with velocity and acceleration criteria. For the remaining points eye velocity values were plotted against head velocity values, and the slope of the linear regression was taken to represent angular VOR or VVOR gain. Phase data are of course unavailable with this tech-
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
DEMER and CRANE
81
Fig. 3. VVOR gains with normal vision and with ´2 telescopic spectacles in young and elderly subjects during quiet standing, walking in place, and active sinusoidal head rotation at 0.8 Hz. Horizontal broken line indicates a gain of 1.0.
nique, but they can be obtained over a range of frequencies with Fourier analysis. In this typical example VOR gain as measured in darkness during standing was less than the conventionally assumed ideal value of 1.0 during both standing and walking. Averaged across both activities, VOR gains for the young were 0.91 ± 0.01 (mean ± SEM) and 0.84 ± 0.02 in pitch and yaw, respectively. For the elderly, corresponding gains were 0.87 ± 0.02 and 0.72 ± 0.03. Whereas VVOR gain in light was modestly higher than VOR gain, VVOR gains during normal vision were similar in both subject groups and remained in the same sense suboptimal for both pitch and yaw. Mean VVOR gain data are summarized for the yaw axis in Fig. 3; for yaw, VVOR gain during standing and walking did not exceed 0.9 in either the young or the elderly subjects. Magnified vision with ´2 telescopic spectacles would conventionally be expected to require an angular VVOR gain of 2.0. In both subject groups observed pitch and yaw gains with telescopic spectacles were considerably less than 2.0 during standing and walking, reflecting very little of the potential visualvestibular interaction (Fig. 3). In fact, VVOR gain with ´2 telescopic spectacles exceeded 1.0 in only the young subjects, and even in them it did not exceed 1.2. During running, VVOR gain in the young subjects averaged less than 1.0.12 The above data suggest that during natural activities vision has at most a modest effect in enhancing VOR gain. The influence of vision was strikingly greater during volitionally generated head rotations, as illustrated
Fig. 4. Records of VOR,VVOR with normal vision, and VVOR with ´2 telescopic spectacles in yaw axis for a young normal subject during actively generated head rotation. Left column displays plots of slow-phase eye and head velocity against time, with rightward velocities depicted as positive. Each cycle of eye velocity was automatically fit with a sinusoid by least squares, except that individual cycles whose gains were statistical outliers for low gain were instead marked by horizontal lines. Right column displays plots of slow-phase eye velocity against head velocity; distribution is equal to gain. Note that magnified vision markedly increases gain to a value considerably greater than that observed for VOR in either yaw or pitch during ambulation (Fig. 2).
for representative active yaw head rotations in Fig. 4. Mean VVOR gain values during active yaw head rotations are shown in Fig. 3. Similar results were observed for pitch (not shown). As reported previously gain enhancement by magnified vision was greater in young than elderly subjects for both pitch and yaw.13,14 Gains measured during active head rotation at 0.8 Hz were higher than during natural activities in both age groups,
82
DEMER and CRANE
Fig. 5. Peak ocular translation in three degrees of freedom during quiet standing and walking in place for young and elderly subjects. Note significantly greater translation in elderly than in young subjects.
reaching values of as great as 1.5 for both pitch and yaw in the young, and 1.3 in the elderly. Other experiments have indicated that this potent visual-vestibular interaction persists at even the highest frequencies attainable by subjects during self-generated head rotations.15,16 Thus, during self-generated head rotation at a moderate frequency, visual-vestibular interaction apparently has a much more efficacy than it has during natural activity. We speculate that the greater VVOR gain enhancement during self-generated head rotation is not a direct visual effect at all but might instead reflect contributions from motor preprogramming, efference copy of skeletal muscle commands, or prediction.3 Head Translation During Natural Activities
Of course, rotational measurements do not fully characterize head motion. Head translation was also measured with a flux gate magnetometer array (Flock of Birds; Ascension Technologies, Burlington, Vt.), with the receiver affixed to the vertex of the headband above the interaural line. The flux gate magnetometer system determined linear and angular head position in three dimensions at 13.9 Hz, operating simultaneously but not synchronously with more rapidly sampled angular position recording with search coils. Flux gate magnetometer data were used primarily to determine the limits of translation during trials, to estimate the velocity of translation in each degree of freedom, and to define the approximate phase relationship between translation and rotation. It should be noted that the head is a rigid body with nonnegligible size compared with the eye, so that head rotations about an axis distant from the eye could cause the eye to translate. To account for
Otolaryngology– Head and Neck Surgery July 1998
this effect, we made a geometric correction on the sampled data to obtain eye translation from head translation and head rotation, taking into account the eccentricity of the eye from the flux gate sensor on the head. Experiments with the head clamped to a support showed that physiologic linear head velocity during quiet standing place exceeded the velocity noise of the flux gate sensor by a factor exceeding 2 in young subjects and by a factor of 4 in the normal elderly. Measurable eye translations were present during standing, walking, and running in both young and elderly subjects, with data on peak displacement summarized in Fig. 5. Peak mediolateral (ML) displacement was similar in both groups during standing (possibly due to a floor effect from system noise), but was about 50% greater in the elderly during walking (p < 0.05). Peak dorsoventral eye displacement was about twice as great in the elderly as the young during both standing and walking (p < 0.05). Both of these differences were mainly attributable to greater head translation in the elderly. The most striking finding arising from study of translation during walking and running was the universal occurrence in both subject groups of a tight phase relationship between eye translation and head rotation. Upward translation was tightly phaselocked to downward pitch rotation, and rightward translation was tightly phase-locked to leftward yaw (Fig. 6). Ocular translation arises from two sources during head motion: first, from the direct translation of the head, and second, from the eccentric location of the eyes relative to the axis of rotation of the head (Fig. 7).12 It may be seen that for moderately remote target distances upward head translation has the effect of offsetting the gaze disturbance produced by downward head rotation; a similar effect occurs for the relationship between leftward head translation and rightward yaw. It is possible to use geometric relationships to compute the parameters of the line of sight (gaze line) from measurements of head rotation and translation. The general kinematic effect of translation on required ocular rotation was explored computationally with the assumption that the mean observed ratio of dorsoventral translation to pitch rotation remains constant regardless of target distance.12 Results for yaw would be similar. The computations have been plotted in Fig. 8. For remote targets the required compensatory ocular rotation would of course be opposite to head rotation, but a surprising finding is that for sufficiently near targets (i.e., 10 cm), the compensatory ocular rotation must be suppressed or even be in the same direction as the head. Telescopic spectacles would magnify both the rotational and translational effects of head movements,
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
DEMER and CRANE
83
Fig. 6. Horizontal and vertical head, eye, and gaze rotation, and eye translation in normal subject during running in place in darkness with remembered distant target, as measured with search coils and flux gate magnetometer sensor. Note that translation was consistently out of phase with rotation, which reduces angular VVOR gain required to stabilize the retinal image of a target located 6 M distant. Thus angular gaze in the head shows small perturbations with gait.
but translation would still reduce the VVOR gain required for image stabilization on the retina. The preceding findings indicate that, in addition to angular head motion during natural activities, the VOR faces a visual demand from head translation. Comparison of angular VOR gains between groups such as the elderly and young may not fairly evaluate their ability to stabilize gaze because differences in phase-locked head translation may alter the angular VOR gain required for a particular target distance. The gain of the angular VOR may be chronically adapted to the value most commonly required in daily life, perhaps as dictated by habitual relationships between head translation and rotation. Finally, visual-vestibular interaction mediated by magnified vision has a smaller effect during head motion incidental to natural activities than is the case during repetitive, volitional head movements. Passive Eccentric Rotations
As noted above, rotation of the head about an axis offset from that of the eyes produces simultaneous angular and linear motion that must be compensated by the oculomotor system to stabilize gaze. With geometry it is possible to compute the angular eye movement required to maintain foveation of proximate or distant targets.17 For example, during rotation about an axis 20
cm posterior to the eyes, a gain of around 3.0 is required to stabilize the retinal image of a target 20 cm distant. The effects of passive eccentric rotation in pitch and yaw were studied in young, normal subjects ranging in age from 19 to 44 years. Magnetic search coils were used to directly measure eye and head movements, avoiding potential confounding by mechanical decoupling of subjects’ heads from the rotators. Pitch rotations were either about an oculocentric axis or about an axis 15 cm posterior to the eyes. Yaw rotations were about one of five axes: 10 cm anterior to the eyes (oculocentric), 4 cm posterior to the eyes, 7 cm posterior to the eyes (otolith centric), or 20 cm posterior to the eyes. Real and remembered targets were located 600 or 300 cm from the eyes, 25 cm from the eyes, and 10 cm from the eyes. Trials with visible targets in the light (VVOR) were interleaved with VOR trials with instructions to the subjects to visualize in darkness the remembered targets at the distances previously viewed. Two motion profiles were used: sinusoids at a constant peak velocity of 30 degrees/second and frequencies ranging between 0.8 and 2.0 Hz, and unpredictable, pseudorandom step changes in head position having typical head accelerations in pitch of 900 to 1100 degrees/second2 (146 N-M = 108 ft-lb rotator), and in yaw of 200 to 300 degrees/second2 (27 N-M rotator).
84
DEMER and CRANE
Otolaryngology– Head and Neck Surgery July 1998
Fig. 7. Geometric illustration of the interrelationship between translation and rotation in stabilization of gaze.
Fig. 8. Calculated ideal VOR gain during walking in place as a function of target distance with an assumed average observed relationship between pitch head rotation and dorsoventral translation.12 For remote targets, ideal VOR gain approaches the conventionally assumed value of 1.0. For near targets the compensatory value of VOR gain may be zero or even reversed.
These studies indicate orderly effects of both target distance and axis of head rotation on VOR and VVOR gains for both pitch17 and yaw.18 Mean values of gain obtained during sinusoidal rotation at frequencies at 1.2 Hz are illustrated here for yaw in Fig. 9. Target proximity reduced both VVOR and VOR gain for very anterior axes of rotation, and increased gain for axes posterior to the eyes (p < 0.001). Gains never achieved the geometrically ideal values depicted by the solid and broken lines, and gains were more modified by rotational axis and target distance for the VVOR than for the VOR. In darkness, rotation about an axis between the otoliths was associated with VOR gain independent of target distance. Note that for this axis the otoliths experienced no net linear acceleration. However, in light, the axis where
Fig. 9. Effect of axis of rotation and target distance on mean VVOR and VOR gains of 15 subjects undergoing sinusoidal rotation at 1.2 Hz about multiple axis eccentricities while regarding targets at various distances (D, in centimeters). Differences in gain caused by variations in target distance were statistically significant (p < 0.01) at all eccentricities and frequencies except those marked with asterisks. Solid lines represent theoretic gains required to stabilize the image of the target on the retina. Note that the theoretic lines intersect at the eyes (0 degrees eccentricity). Error bars are drawn to represent the ±1 SEM, although the errors are often smaller than the plot symbols. Upper panel, VVOR gain measured in the light with a visible target. VVOR gain was progressively increased by target proximity for axis eccentricity posterior to eyes and progressively decremented by target proximity for eccentricity anterior to eyes. Effect of target distance on VVOR gain was least at an eccentricity 4 cm posterior to eyes. For VVOR data at frequencies greater than 1.2 Hz, variations in target distance were not statistically significant (p > 0.01) at eccentricity of 7 cm. Lower panel, VOR gain measured in darkness with remembered targets. VOR gain was progressively increased by target proximity for axis eccentricity posterior to eyes and progressively decremented by target proximity for eccentricity anterior to eyes. Gain changes were smaller than during rotations with a visible target (VVOR). Effect of target distance on VOR gain was least at an eccentricity 7 cm posterior to the eyes, corresponding to location of otoliths.
VVOR gain became independent of target distance was between the eyes and the otoliths. We interpret these findings to indicate that otolithic input interacts with canal input to provide gain enhancement during eccen-
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
tric rotations and that this effect may be further modified by vision during steady-state rotation in light. For both pitch and yaw rotations very posterior to the eyes (i.e., 15 to 20 cm), VVOR gain with 10 cm targets did substantially exceed unity at relatively low frequencies (i.e., 0.8 Hz), but gain declined with increasing frequency up to 2.0 Hz, the highest frequency tested. This VVOR gain enhancement is similar to the behavior of pursuit tracking, which has a declining gain with increasing stimulus frequency. During sinusoidal eccentric rotation in both pitch and yaw, the VOR exhibited a phase change systematically related to frequency of head rotation. With the axis of rotation well behind the otoliths, near target viewing was associated with a gain increase and phase lag, both increasing with frequency. With the axis of rotation anterior to the otoliths, near target viewing was associated with a gain decrement and with phase lead increasing with frequency. We modeled this behavior as a linear interaction between canal input, and otolith input scaled according to the inverse of target distance.18,19 The predictions of this simple model are consistent with the trends in the data, supporting the concept of a linear interaction between canal and otolith inputs. For rotations in the light, pursuit tracking probably adds another contribution to the response. The presence of an actual visible target also allows an accurate and compelling estimate of target distance. It is desirable to optimize control of subjects’ mental sets by providing an actual visible target but to be able to study purely vestibular responses independent of the confounding effects of visual tracking. We were able to do this by studying the initial VOR response in the presence of visible targets. In such tests only the initial compensatory eye movements are analyzed before the availability of visual feedback to generate pursuit eye movements.17,19 Because pursuit latency exceeds 100 msec, we limited analysis to the first 80 msec of the response. Unpredictable, passive, whole-body transient rotations were used to avoid predictive eye movements and prevent the use of motor efference copy to augment the VOR. Using sampling rates of 800 Hz, we have found it feasible to analyze VOR and VVOR responses to pseudorandom sequences consisting of about 20 unpredictable changes in rotator position. Automated analysis programs identified head acceleration events and then automatically fit regression lines to head and eye signals in either the position (yaw) or velocity (pitch) domains, depending on the characteristics of the head movement (Fig. 10). We have been able to reliably compute initial gains and to show dependence of gain on rotational eccentricity and target proximity, as illustrated in Fig. 11. We found no effect of axis or target distance on the initial 25 msec of the VOR response; however, by 30 to
DEMER and CRANE
85
Fig. 10. Typical responses of one subject to unpredictable transients of head rotation in the light (VVOR) about oculocentric and eccentric axes and with targets located at either 10 or 600 cm. Rightward rotations are positive. Filtering at 43 Hz to minimize noise precludes visualization of latency between the head and eye velocity profiles. Perturbations in eye velocity were commonly observed near 30 and 60 msec. Upper panel, VVOR during viewing a target at 10 cm with a posterior eccentric rotational axis (20 cm) and an oculocentric axis (0 cm). Responses were similar until 70 msec, after which a greater eye velocity was observed for the eccentric axis. Lower panel, VVOR during posterior eccentric rotation of 20 cm but with targets at 10 and 600 cm. The responses were similar up to 40 msec, after which the response was greater for the near target.
60 msec, clear effects of both axis of rotation and target distance were evident. Because analysis the ocular response from 25 to 80 msec prevents any direct contribution of visual tracking to observed gain, we attribute gain enhancement during rotations in the light to an effect of mental set. Despite an earlier report by others,20 we are not confident of the reliability of these data in estimating VOR latencies because torque limitations of the rotators make the onset of head rotations gradual, and the low early signal-to-noise ratio necessitates considerable low-pass filtering.
86
DEMER and CRANE
Fig. 11. Effect of rotational axis eccentricity and target distance (D, in centimeters) on mean VOR gains of 15 subjects undergoing unpredictable transient rotation. Points represent average of both VOR and VVOR data because visual feedback had no significant effect on gain. Solid lines, Theoretical gains required to stabilize image of the target on the retina. Error bars ± 1 SEM. Upper panel, Gain measured over first 25 msec of response was independent of both rotational axis and target distance ( p > 0.05). Lower panel, Gain measured from 25 to 80 msec after head rotation onset was progressively increased by target proximity for axis eccentricity posterior to eyes and decreased by target proximity for axis eccentricity anterior to eyes. Gain changes were smaller than during sinusoidal rotations under either lighting condition.
Implications
These data indicate that during ambulation a strong counterphase relationship exists between dorsoventral translation and pitch,21-23 as well as between mediolateral translation and yaw.12 These relationships were observed in every trial and every subject tested, so they are clearly robust. Perhaps the counterphase relationships are simply the result of passive head inertia. This idea is mechanically plausible for the pairings of pitch with dorsoventral translation, and yaw with mediolater-
Otolaryngology– Head and Neck Surgery July 1998
al translation, because the center of mass of the head is anterior to the support point of the head on the neck. Participation of active neural reflexes is also a possibility. Head stabilization during angular perturbations of the torso is believed to be accomplished by voluntary mechanisms at frequencies below 1.0 Hz, by the vestibulocolic and cervicocolic reflexes from 1.0 to 2.5 Hz and by mechanical considerations at higher frequencies.24,25 The 1 to 2 Hz fundamental frequencies of head motion during natural gait7,26 overlap the frequency range of active reflex head stabilization. Support for an active contribution is provided by the observation that patients with vestibular loss lack the counterphase relationship between dorsoventral translation and pitch21 and the observation that in normal walking subjects the ratio between pitch head rotation and dorsoventral translation is altered by reducing target distance.23 These observations suggest that the relationship between head translation and rotation may at least partially be governed by an active reflex system influenced by target distance, similar to the effect of context on head-neck dynamics hypothesized by Keshner et al.24 This reflex system may be deranged in patients with labyrinthine dysfunction who exhibit significantly more translational and rotational oscillations of head position than normal, especially in the vertical direction.10 Geometric analysis of the effects of head translation and rotation suggests that these two types of motion, when phase locked in the way observed here, act in a mutually compensatory manner to enhance gaze stability for typical target distances. There is an important implication of this finding for interpretation of tests of vestibular function. Conventional rotational clinical tests of the vestibular system involve passive rotation in darkness of the head and body about some vertical axis, at constant velocity or at sinusoidal frequencies from 0.0125 to 1.0 Hz.27 Often no care is taken to determine the precise axis of rotation, and no effort is made to control the subject’s estimate of target distance. We have shown here that both of these factors have important effects on optimal and on observed VOR gain. A further issue is possible adaptation to a habitual mode of eye-head coordination during natural activities. For example, the elderly subjects studied here exhibited lower angular VOR gains during most natural activities than did the young subjects. Does the decline in angular VOR gain represent deterioration of vestibular function with aging, and should this be expected to be reflected in poorer balance and dynamic visual acuity? Such questions cannot be answered with angular VOR data alone because we now appreciate that head translation, which is greater in the elderly, supplements the angular VOR in stabilizing gaze during ambulation for
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
typical target distances. Measurement of the VOR in all degrees of translational and rotational freedom, as illustrated in Fig. 6, would permit full characterization of ocular stabilization during natural motion. Measurements of VVOR gain were also made during ambulation while subjects wore ´2 telescopic spectacles. Although statistically significant VVOR gain enhancement with telescopic spectacles did occur during the three natural activities, the increase in gain was typically only about 15% to 20% and was thus far below the compensatory value equal to the magnification of the spectacles worn. During self-generated head rotations in both yaw and pitch, the same subjects attained very large augmentations of VVOR gain to values exceeding 1.5, although never quite equal to telescopic spectacle magnification of ´2. Experiments with telescopic spectacles of up to ´6 power have indicated that VVOR gain during self-generated head movements can exceed 3.0.15 Therefore the surprising finding was that during natural activities magnified vision was largely ineffective in significantly augmenting the VOR to accomplish gaze stabilization. This implies that patients with visual impairments requiring intermittent wearing of telescopic spectacles will be faced with substantial uncompensated retinal image instability during standing and ambulation.1,28 In the current studies performance during “autorotational” VOR and VVOR testing was not at all representative of vestibular performance during even the most common of natural activities. Perhaps this is because head movements during natural activities have a high frequency content and a major element of randomness29; because these head perturbations are truly unintended, motor efference copy is presumably unavailable to use in gaze stabilization. Enhancement of VVOR gain with telescopic spectacles was considerably greater during passive, predictable, whole-body rotations up to about 2.0 Hz than was observed during natural activities,3 although still less than during volitionally generated head rotations at similar frequencies.16 Because in both cases testing was with equally predictable sinusoidal motion, the greater gain enhancement during self-generated head movement might be attributed to motor efference copy or preprogramming, or perhaps to neck proprioception or the cervico-ocular reflex. The latter is generally considered to be insignificant in human beings30,31 but may become significant in subjects with compensated peripheral vestibular loss.32,33 In contrast to the small immediate influence of magnified vision on VOR gain during natural activities, data on VOR initiation indicate that intended target distance can modify the response to combined semicircular canal and otolith stimulation within a latency short-
DEMER and CRANE
87
er than that of visual tracking. Central regulation of the response to otolithic stimulation may be one means by which gaze stabilization is controlled during combined angular and linear head motion. Modeling is compatible with the idea that responses to otolithic stimulation are scales by the inverse of intended target distance. The findings summarized here indicate that angular VOR performance depends on intended visual target distance, which interacts with head translation. The result is consistent with recent theories of the VOR reflecting the idea that the vestibular input is only one input to a larger gaze control system.34-36 The VOR thus does not behave stereotypically for all head movements. We suggest that the vestibular system should be tested under more natural conditions7 and that both rotational and translational head movements be tested in the context of visual requirements. Results of such tests should be functionally validated by correlation with the most fundamental functional purpose of VORs, gaze stabilization. Nicolasa De Salles and Troy Howard provided technical assistance with the experiments. REFERENCES 1. Demer JL, Porter FI, Goldberg J, et al. Validation of physiologic predictors of successful telescopic spectacle use in low vision. Invest Ophthalmol Vis Sci 1991;32:2826-34. 2. Demer JL, Goldberg J, Porter FI. Effect of telescopic spectacles on head stability in normal and low vision. J Vestib Res 1991;1: 109-22. 3. Demer JL. Mechanisms of human vertical visual-vestibular interaction. J Neurophysiol 1992;68:2128-46. 4. Lisberger SG, Evinger C, Johanson GW, et al. Relationship between eye acceleration and retinal image velocity during foveal smooth pursuit in man and monkey. J Neurophysiol 1981; 46:229-49. 5. Pozzo T, Berthoz A, Lefort L. Head kinematic during various motor tasks in humans. Prog Brain Res 1989;80:377-83. 6. Grossman GE, Leigh RJ, Abel LA, et al. Frequency and velocity of rotational head perturbations during locomotion. Exp Brain Res 1988;70:470-6. 7. Grossman GE, Leigh RJ, Bruce EN, et al. Performance of the human vestibuloocular reflex during locomotion. J Neurophysiol 1989;62:264-72. 8. Demer JL, Amjadi F. Dynamic visual acuity of normal subjects during vertical optotype and head motion. Invest Ophthalmol Vis Sci 1993;34:1894-906. 9. Grossman GE, Leigh RJ. Instability of gaze during locomotion in patients with deficient vestibular function. Ann Neurol 1990;27: 528-32. 10. Takahashi M, Hoshikawa H, Tsujita N, et al. Effect of labyrinthine dysfunction upon head oscillation and gaze during stepping and running. Acta Otolaryngol (Stockh) 1988;106:348-53. 11. Gagnon J, Roth JM, Carroll M, et al. SuperANOVA. Berkeley: Abacus Concepts; 1989. p. 185-218. 12. Demer JL, Viirre ES. Visual-vestibular interaction during standing walking, and running. J Vestib Res 1996;6:295-313. 13. Demer JL. Effect of aging and visual impairment on dynamic visual acuity during vertical motion. In: Noninvasive assessment of the visual system. 1993 Technical digest series. Washington (DC): Optical Society of America; 1993. p. 192-5.
88
Otolaryngology– Head and Neck Surgery July 1998
DEMER and CRANE
14. Demer JL. Effect of aging on vertical visual tracking and visualvestibular interaction. J Vestib Res 1994;4:355-70. 15. Demer JL, Oas JG, Baloh RW. Visual-vestibular interaction during high-frequency, active head movements in pitch and yaw. In: Cohen B, Tomko DL, Guedry F, editors. Sensing and controlling motion. Vestibular and sensorimotor function. New York: New York Academy of Sciences; 1992. p. 832-5. 16. Demer JL, Oas JG, Baloh RW. Visual-vestibular interaction in humans during active and passive, vertical head movement. J Vestib Res 1993;3:101-14. 17. Viirre ES, Demer JL. The human vertical vestibulo-ocular reflex during combined linear and angular acceleration with near target fixation. Exp Brain Res 1996;112:313-24. 18. Crane BT, Viirre ES, Demer JL. The human horizontal vestibulo-ocular reflex during combined linear and angular acceleration. Exp Brain Res. In press 1996. 19. Crane BT, Viirre ES, Demer JL. Effect of target distance and eccentric rotation on human horizontal vestibulo-ocular reflex (VOR). Soc Neurosci Abstr 1995;21:519. 20. Johnston JL, Sharpe JA. The initial vestibulo-ocular reflex and its visual enhancement and cancellation in humans. Exp Brain Res 1994;99:302-8. 21. Pozzo T, Berthoz A, Lefort L, et al. Head stabilization during various locomotor tasks in humans. II. Patients with bilateral peripheral vestibular deficits. Exp Brain Res 1991;85:208-17. 22. Hirasaki E, Kubo T, Nozawa S, et al. Analysis of head and body movements of elderly people during locomotion. Acta Otolaryngol (Stockh) 1993;501(Suppl):25-30. 23. Bloomberg JL, Reschke MF, Huebner WW, et al. The effects of target distance on eye and head movement during locomotion. In: Cohen B, Tomko DL, Guedry F, editors. Sensing and controlling motion: vestibular and sensorimotor function. New York: New York Academy of Sciences; 1992. p. 699707. 24. Keshner EA, Cromwell RL, Peterson BW. Mechanisms control-
25. 26.
27. 28. 29. 30. 31. 32. 33. 34. 35. 36.
ling human head stabilization. II. Head-neck characteristics during random rotations in the vertical plane. J Neurophysiol 1995; 73:2302-12. Keshner FA, Peterson BW. Mechanisms controlling human head stabilization. I. Head-neck dynamics during random rotations in the horizontal plane. J Neurophysiol 1995;73:2302-12. Keshner EA, Cromwell R, Peterson BW. Head stabilization during vertical seated rotations and gait. In: Woollacott MH, Horak F, editors. Posture and gait: control mechanisms. Eugene (OR): University of Oregon Books; 1992. p. 105-8. Jenkins HA, Honrubia V, Baloh R. Evaluation of multiple-frequency rotatory testing in patients with peripheral labyrinthine weakness. Am J Otolaryngol 1982;3:182-8. Demer JL, Porter FI, Jenkins HA, et al. Predictors of functional success in telescopic spectacle use by low vision patients. Invest Ophthalmol Vis Sci 1989;30:1652-65. Maas EF, Huebner WP, Seidman SH, et al. Behavior of human horizontal vestibulo-ocular reflex in response to high-acceleration stimuli. Brain Res 1989;499:153-6. Berthoz A. Adaptive mechanisms in eye-head coordination. Rev Oculomot Res 1985;1:177-201. Jurgens R, Mergner T. Interaction between cervico-ocular and vestibulo-ocular reflexes in normal adults. Exp Brain Res 1989; 77:381-90. Chambers BR, Mai M, Barber HO. Bilateral vestibular loss, oscillopsia, and the cervico-ocular reflex. Otolaryngol Head Neck Surg 1985;93:403-7. Kasai T, Zee DS. Eye-head coordination in labyrinthine-defective human beings. Brain Res 1978;144:123-41. Collewijn H. The vestibulo-ocular reflex: an outdated concept? Prog Brain Res 1989;80:197-209. Collewijn H. The vestibulo-ocular reflex: is it an independent subsystem? Rev Neurol (Paris) 1989;145:502-12. Steinman RW, Kowler E, Collewijn H. New directions for oculomotor research. Vision Res 1990;30:1845-64.
Endoscopic and Image Guided Sinus Surgery
The Joint Center for Otolaryngology, Beth Israel Deaconess Medical Center and Brigham and Women’s Hospital of Harvard Medical School, is sponsoring a course, “Advanced Endoscopic and Image Guided Sinus Surgery—Beyond the Basics,” Nov. 6-7, 1998, at the Fairmont Copley Plaza Hotel in Boston, Mass. This year’s honored guest is Professor Heinz R. Stammberger of Graz, Austria. This course provides upto-date, state-of-the-art information and instruction to the practicing otolaryngologist regarding advanced techniques in endoscopic sinus surgery. For further information, contact Harvard MEDCME, PO Box 825, Boston, MA 02117-0826; phone (617)432-1525.
This article summarizes six recent degree-of-freedom studies of visual-vestibular interaction during natural activities and relates the findings to canalotolith interactions evaluated during eccentric axis rotations. Magnetic search coils were used to measure angular eye and head movements of young and elderly subjects. A flux gate magnetometer was used to measure three-dimensional head translation. Three activities were studied: standing quietly, walking in place, and running in place. Each activity was evaluated with three viewing conditions: a visible target viewed normally, a remembered target in darkness, and a visible target viewed with ´2 binocular telescopic spectacles. Canal-otolith interaction was assessed with passive, whole-body, transient, and steady-state rotations in pitch and yaw at multiple frequencies about axes that were either oculocentric or eccentric to the eyes. For each rotational axis, subjects regarded visible and remembered targets located at various distances. Horizontal and vertical angular vestibuloocular reflexes were demonstrable in all subjects during standing, walking, and running. When only angular gains were considered, gains in both darkness and during normal vision were less than 1.0 and were generally lower in elderly than in young subjects. Magnified vision with ´2 telescopic spectacles produced only small gain increases as compared with normal vision. During walking and running all subjects exhibited significant mediolateral and dorsoventral head translations that were antiphase locked to yaw and pitch head movements, respectively. These head translations and rotations have mutually compensating effects on gaze in a target plane for typical viewing distances From the Jules Stein Eye Institute, Departments of Ophthalmology and Neurology, University of California, Los Angeles. Supported by grants from the National Eye Institute (DC-01404, DC02952, EY-08656, and core grant EY-00331) and Medical Scientist Training Program (GM-0804). Joseph L. Demer was the recipient of a Research to Prevent Blindness Lew R. Wasserman Merit Award. Presented at the UCLA Conference on Vestibular Adaptation, Santa Monica, Calif., May 23-25, 1996. Reprint requests: Joseph L. Demer, MD, PhD, Jules Stein Eye Institute, 100 Stein Plaza, UCLA, Los Angeles, CA 90095-7002. Copyright © 1998 by the American Academy of Otolaryngology– Head and Neck Surgery Foundation, Inc. 0194-5998/98/$5.00 + 0 23/1/87913 78
and allow angular vestibulo-ocular reflex gains of less than 1.0 to be optimal for gaze stabilization during natural activities. During passive, whole-body eccentric pitch and yaw head rotations, vestibuloocular reflex gain was modulated as appropriate to stabilize gaze on targets at the distances used. This modulation was evident within the first 80 msec of onset of head movement, too early to be caused by immediate visual tracking. Modeling suggests a linear interaction between canal signals and otolith signals scaled by the inverse of target distance. Vestibulo-ocular reflex performance appears to be adapted to stabilize gaze during translational and rotational perturbations that occur during natural activities, as is appropriate for relevant target distances. Although immediate visual tracking contributes little to gaze stabilization during natural activities, visual requirements determine the performance of vestibulo-ocular reflexes arising from both canals and otoliths. (Otolaryngol Head Neck Surg 1998;119:78-88.)
T
he human head undergoes incessant motion. Angular head movements occur constantly in daily life because of transmitted heartbeat, tremor, sway, ambulation, and vehicular travel.1 Even when a person is attempting to stand completely at rest, involuntary angular head movements occur about the vertical axis (yaw), the interaural axis (pitch), and the anteroposterior axis (roll).2 Such head movements during quiet standing have root-mean-square (RMS) velocities of around 1 degree/second and commonly occur at frequencies above 1.0 Hz, where smooth visual-tracking mechanisms have significant performance limitations.3,4 During quiet standing these movements typically have peak components at frequencies exceeding 5 Hz. Larger head velocities are encountered in all three rotational axes during gentle walking in place,2,5 with predominant frequencies sometimes exceeding 2.5 Hz for yaw and 8 Hz for pitch.6,7 Natural activities are also characterized by complex patterns of head rotation and translation during the viewing of objects at a variety of distances from the eyes. Vigorous activities such as running or hopping produce still greater head velocities. During these head movements the vestibulo-ocular reflex (VOR) serves vision by generating reflex eye movements that stabilize images on the retina. This
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
image stabilization is critical to vision because angular image motion exceeding about 2 to 3 degrees/second degrades visual acuity.8 Whereas the human angular VOR has been studied extensively for passive head rotations and self-generated head rotations about a single axis, the response of the VOR to head translations has been examined less extensively. For pure head rotations about the center of the eye, an ideal VOR would always generate an eye movement of equal magnitude and opposite direction to the head movement. This is the situation most commonly studied in the laboratory. Consideration of translation adds complexity to the issue because, unlike the case of pure rotation, target distance is a significant determinant of effective gaze stabilization. Further, there are two eyes, neither of which is generally colocated with the axis of rotation of the head, so that natural head movements normally consist of combinations of head rotation and translation. This article addresses the operation of the VOR in gaze stabilization during the combined rotational head movements that occur during standing and ambulation and relates this to the interactions of the semicircular canals and otoliths during off-axis rotations. ANGULAR VOR DURING NATURAL ACTIVITIES
Refinements of the magnetic search coil technique7,9 permit the recording of angular VORs during natural activities not associated with perceptible head movements. Reference magnetic field coils 2 M in diameter create a sufficiently homogeneous central field that the signal representing the angular position of the search coil is not appreciably dependent on translation within a substantial central region.9 Measurements during standing and walking were made in 9 normal young (average age 29 years) and 11 normal elderly (average age 78 years) subjects. Written informed consent was secured after approval of all protocols by the institutional review board for the protection of human subjects. Subjects had normal corrected vision and had no known vestibular pathology. Sense coils were placed on one of the patient’s eyes while he or she was under topical anesthesia (Skalar Medical, Delft, The Netherlands) and at the front of a snug headband for measurement of angular gaze and head positions, respectively. Each subject stood in the center of the magnetic field reference cube. Testing was performed for all combinations of five activities, two viewing conditions, and interleaving two conditions of illumination. Activities consisted of (1) standing motionless with heels and toes together; (2) walking in place so that each footfall was synchronized with a metronome beating at 2 Hz; (3) running in place so that each footfall was synchronized with a metronome beating at 3 Hz (young subjects only); (4)
DEMER and CRANE
79
Fig. 1. Head velocity in pitch and yaw during quiet standing and walking in place, for young and elderly subjects. Elderly subjects have higher angular head velocity than young subjects in both axes.
active, single-axis sinusoidal head movement at 0.8 Hz in pitch paced by frequency modulation of an audible tone; and (5) active, single-axis sinusoidal head movement at 0.8 Hz in yaw paced by frequency modulation of an audible tone. For walking and running in place we used the same gait frequencies chosen by Takahashi et al.,10 which corresponded well to the rates of uncoached gait reported by Grossman et al.7 For each activity subjects regarded a high-contrast target located at standing eye level on the wall 6 to 9 M distant. Viewing conditions consisted of (1) unmagnified viewing with refractive correction if necessary and (2) ´2 binocular telescopic spectacles with a magnified field of 16.8 degrees. The unmagnified visual field peripheral to the telescopes was masked. For each combination of activity and viewing condition, measurements were interleaved first under normal indoor illumination (visually enhanced VOR [VVOR]) and then again in complete darkness with instructions to look at the remembered target location (VOR). For each of the test conditions, values of RMS angular head velocity, frequency of the peak Fourier component of angular head velocity, and VOR or VVOR gain were each obtained in duplicate trials. Duplicate measurements were averaged for subsequent analysis, which was performed with analysis of variance and covariance.11 Gaze and head positions were low-pass filtered over a bandwidth of 0 to 100 Hz before digital computer sampling with 16-bit precision at 400 Hz. Eye position in the head was obtained by subtraction of head position from gaze position. After automated removal of saccades,3 VOR and VVOR gains were taken to be the least squares slope of the plot of slow-phase eye veloc-
80
DEMER and CRANE
Otolaryngology– Head and Neck Surgery July 1998
Fig. 2. Eye and head velocities illustrating angular vertical VOR of normal young subject during standing, walking, and running in place. Upward velocities are positive. VOR gain and RMS head velocity (0 to 20 Hz) are indicated for each activity. Note 10-fold greater velocity sensitivity during standing than during walking and running. Top row, Gains are defined by linear regression of eye velocity on head velocity, with analysis of points lying within noise-defined limits of the fit (parallel lines). Saccades were truncated at around 2, 25, and 35 degrees/second for the three activity conditions, respectively, and these data points are visible outside the acceptance limits of the fits. Spontaneous 1 degree/second downward bias present for all activities reflects a drift common in normal subjects in darkness.
ity against head velocity. The RMS value of head velocity was computed for pitch and yaw. Head velocity was also subjected to Fourier transformation to obtain the frequency of the peak component in pitch and yaw. Understanding of the VOR during natural activities understanding of the characteristics of head motion that serve as the stimulus to the reflex. Neither the presence of a visible target nor the presence of spectacle magnification significantly influenced rotational or translational head motion when data were averaged across young or elderly subjects, so it was appropriate to pool data across viewing conditions. Predominant frequency components on Fourier analysis of head velocity were related to harmonics of the rate of footfalls. The observed RMS (0 to 20 Hz) velocities for angular head motion in roll, pitch, and yaw are summarized in Fig. 1. Angular head velocity in young standing subjects for both pitch and yaw exceeded by sixfold to eightfold the instrumentation noise in our search coil system (approximately 0.1 degree/second, 0 to 20 Hz) and was greater in the elderly. As expected, angular head velocity was greater for walking than standing (Fig. 1).
Interestingly, head velocity was 15% to 20% greater in the elderly than the young subjects for both pitch and yaw during both standing and walking in place (p < 0.01). Young subjects had even greater angular head velocities during running than during walking in place, but because few of the elderly subjects were able to run in place, quantitative comparisons of the two groups were not performed for this condition. Physiologic angular head instability thus provides a sufficient stimulus for reliable measurement of VOR gain even in subjects attempting to stand motionless. Using these techniques, we were able, under both lit and darkened conditions, to observe eye movements that were directed in an opposite direction from angular head movements during all activities studied, as illustrated for pitch in Fig. 2. Saccades and quick phases were occasionally present and were removed with velocity and acceleration criteria. For the remaining points eye velocity values were plotted against head velocity values, and the slope of the linear regression was taken to represent angular VOR or VVOR gain. Phase data are of course unavailable with this tech-
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
DEMER and CRANE
81
Fig. 3. VVOR gains with normal vision and with ´2 telescopic spectacles in young and elderly subjects during quiet standing, walking in place, and active sinusoidal head rotation at 0.8 Hz. Horizontal broken line indicates a gain of 1.0.
nique, but they can be obtained over a range of frequencies with Fourier analysis. In this typical example VOR gain as measured in darkness during standing was less than the conventionally assumed ideal value of 1.0 during both standing and walking. Averaged across both activities, VOR gains for the young were 0.91 ± 0.01 (mean ± SEM) and 0.84 ± 0.02 in pitch and yaw, respectively. For the elderly, corresponding gains were 0.87 ± 0.02 and 0.72 ± 0.03. Whereas VVOR gain in light was modestly higher than VOR gain, VVOR gains during normal vision were similar in both subject groups and remained in the same sense suboptimal for both pitch and yaw. Mean VVOR gain data are summarized for the yaw axis in Fig. 3; for yaw, VVOR gain during standing and walking did not exceed 0.9 in either the young or the elderly subjects. Magnified vision with ´2 telescopic spectacles would conventionally be expected to require an angular VVOR gain of 2.0. In both subject groups observed pitch and yaw gains with telescopic spectacles were considerably less than 2.0 during standing and walking, reflecting very little of the potential visualvestibular interaction (Fig. 3). In fact, VVOR gain with ´2 telescopic spectacles exceeded 1.0 in only the young subjects, and even in them it did not exceed 1.2. During running, VVOR gain in the young subjects averaged less than 1.0.12 The above data suggest that during natural activities vision has at most a modest effect in enhancing VOR gain. The influence of vision was strikingly greater during volitionally generated head rotations, as illustrated
Fig. 4. Records of VOR,VVOR with normal vision, and VVOR with ´2 telescopic spectacles in yaw axis for a young normal subject during actively generated head rotation. Left column displays plots of slow-phase eye and head velocity against time, with rightward velocities depicted as positive. Each cycle of eye velocity was automatically fit with a sinusoid by least squares, except that individual cycles whose gains were statistical outliers for low gain were instead marked by horizontal lines. Right column displays plots of slow-phase eye velocity against head velocity; distribution is equal to gain. Note that magnified vision markedly increases gain to a value considerably greater than that observed for VOR in either yaw or pitch during ambulation (Fig. 2).
for representative active yaw head rotations in Fig. 4. Mean VVOR gain values during active yaw head rotations are shown in Fig. 3. Similar results were observed for pitch (not shown). As reported previously gain enhancement by magnified vision was greater in young than elderly subjects for both pitch and yaw.13,14 Gains measured during active head rotation at 0.8 Hz were higher than during natural activities in both age groups,
82
DEMER and CRANE
Fig. 5. Peak ocular translation in three degrees of freedom during quiet standing and walking in place for young and elderly subjects. Note significantly greater translation in elderly than in young subjects.
reaching values of as great as 1.5 for both pitch and yaw in the young, and 1.3 in the elderly. Other experiments have indicated that this potent visual-vestibular interaction persists at even the highest frequencies attainable by subjects during self-generated head rotations.15,16 Thus, during self-generated head rotation at a moderate frequency, visual-vestibular interaction apparently has a much more efficacy than it has during natural activity. We speculate that the greater VVOR gain enhancement during self-generated head rotation is not a direct visual effect at all but might instead reflect contributions from motor preprogramming, efference copy of skeletal muscle commands, or prediction.3 Head Translation During Natural Activities
Of course, rotational measurements do not fully characterize head motion. Head translation was also measured with a flux gate magnetometer array (Flock of Birds; Ascension Technologies, Burlington, Vt.), with the receiver affixed to the vertex of the headband above the interaural line. The flux gate magnetometer system determined linear and angular head position in three dimensions at 13.9 Hz, operating simultaneously but not synchronously with more rapidly sampled angular position recording with search coils. Flux gate magnetometer data were used primarily to determine the limits of translation during trials, to estimate the velocity of translation in each degree of freedom, and to define the approximate phase relationship between translation and rotation. It should be noted that the head is a rigid body with nonnegligible size compared with the eye, so that head rotations about an axis distant from the eye could cause the eye to translate. To account for
Otolaryngology– Head and Neck Surgery July 1998
this effect, we made a geometric correction on the sampled data to obtain eye translation from head translation and head rotation, taking into account the eccentricity of the eye from the flux gate sensor on the head. Experiments with the head clamped to a support showed that physiologic linear head velocity during quiet standing place exceeded the velocity noise of the flux gate sensor by a factor exceeding 2 in young subjects and by a factor of 4 in the normal elderly. Measurable eye translations were present during standing, walking, and running in both young and elderly subjects, with data on peak displacement summarized in Fig. 5. Peak mediolateral (ML) displacement was similar in both groups during standing (possibly due to a floor effect from system noise), but was about 50% greater in the elderly during walking (p < 0.05). Peak dorsoventral eye displacement was about twice as great in the elderly as the young during both standing and walking (p < 0.05). Both of these differences were mainly attributable to greater head translation in the elderly. The most striking finding arising from study of translation during walking and running was the universal occurrence in both subject groups of a tight phase relationship between eye translation and head rotation. Upward translation was tightly phaselocked to downward pitch rotation, and rightward translation was tightly phase-locked to leftward yaw (Fig. 6). Ocular translation arises from two sources during head motion: first, from the direct translation of the head, and second, from the eccentric location of the eyes relative to the axis of rotation of the head (Fig. 7).12 It may be seen that for moderately remote target distances upward head translation has the effect of offsetting the gaze disturbance produced by downward head rotation; a similar effect occurs for the relationship between leftward head translation and rightward yaw. It is possible to use geometric relationships to compute the parameters of the line of sight (gaze line) from measurements of head rotation and translation. The general kinematic effect of translation on required ocular rotation was explored computationally with the assumption that the mean observed ratio of dorsoventral translation to pitch rotation remains constant regardless of target distance.12 Results for yaw would be similar. The computations have been plotted in Fig. 8. For remote targets the required compensatory ocular rotation would of course be opposite to head rotation, but a surprising finding is that for sufficiently near targets (i.e., 10 cm), the compensatory ocular rotation must be suppressed or even be in the same direction as the head. Telescopic spectacles would magnify both the rotational and translational effects of head movements,
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
DEMER and CRANE
83
Fig. 6. Horizontal and vertical head, eye, and gaze rotation, and eye translation in normal subject during running in place in darkness with remembered distant target, as measured with search coils and flux gate magnetometer sensor. Note that translation was consistently out of phase with rotation, which reduces angular VVOR gain required to stabilize the retinal image of a target located 6 M distant. Thus angular gaze in the head shows small perturbations with gait.
but translation would still reduce the VVOR gain required for image stabilization on the retina. The preceding findings indicate that, in addition to angular head motion during natural activities, the VOR faces a visual demand from head translation. Comparison of angular VOR gains between groups such as the elderly and young may not fairly evaluate their ability to stabilize gaze because differences in phase-locked head translation may alter the angular VOR gain required for a particular target distance. The gain of the angular VOR may be chronically adapted to the value most commonly required in daily life, perhaps as dictated by habitual relationships between head translation and rotation. Finally, visual-vestibular interaction mediated by magnified vision has a smaller effect during head motion incidental to natural activities than is the case during repetitive, volitional head movements. Passive Eccentric Rotations
As noted above, rotation of the head about an axis offset from that of the eyes produces simultaneous angular and linear motion that must be compensated by the oculomotor system to stabilize gaze. With geometry it is possible to compute the angular eye movement required to maintain foveation of proximate or distant targets.17 For example, during rotation about an axis 20
cm posterior to the eyes, a gain of around 3.0 is required to stabilize the retinal image of a target 20 cm distant. The effects of passive eccentric rotation in pitch and yaw were studied in young, normal subjects ranging in age from 19 to 44 years. Magnetic search coils were used to directly measure eye and head movements, avoiding potential confounding by mechanical decoupling of subjects’ heads from the rotators. Pitch rotations were either about an oculocentric axis or about an axis 15 cm posterior to the eyes. Yaw rotations were about one of five axes: 10 cm anterior to the eyes (oculocentric), 4 cm posterior to the eyes, 7 cm posterior to the eyes (otolith centric), or 20 cm posterior to the eyes. Real and remembered targets were located 600 or 300 cm from the eyes, 25 cm from the eyes, and 10 cm from the eyes. Trials with visible targets in the light (VVOR) were interleaved with VOR trials with instructions to the subjects to visualize in darkness the remembered targets at the distances previously viewed. Two motion profiles were used: sinusoids at a constant peak velocity of 30 degrees/second and frequencies ranging between 0.8 and 2.0 Hz, and unpredictable, pseudorandom step changes in head position having typical head accelerations in pitch of 900 to 1100 degrees/second2 (146 N-M = 108 ft-lb rotator), and in yaw of 200 to 300 degrees/second2 (27 N-M rotator).
84
DEMER and CRANE
Otolaryngology– Head and Neck Surgery July 1998
Fig. 7. Geometric illustration of the interrelationship between translation and rotation in stabilization of gaze.
Fig. 8. Calculated ideal VOR gain during walking in place as a function of target distance with an assumed average observed relationship between pitch head rotation and dorsoventral translation.12 For remote targets, ideal VOR gain approaches the conventionally assumed value of 1.0. For near targets the compensatory value of VOR gain may be zero or even reversed.
These studies indicate orderly effects of both target distance and axis of head rotation on VOR and VVOR gains for both pitch17 and yaw.18 Mean values of gain obtained during sinusoidal rotation at frequencies at 1.2 Hz are illustrated here for yaw in Fig. 9. Target proximity reduced both VVOR and VOR gain for very anterior axes of rotation, and increased gain for axes posterior to the eyes (p < 0.001). Gains never achieved the geometrically ideal values depicted by the solid and broken lines, and gains were more modified by rotational axis and target distance for the VVOR than for the VOR. In darkness, rotation about an axis between the otoliths was associated with VOR gain independent of target distance. Note that for this axis the otoliths experienced no net linear acceleration. However, in light, the axis where
Fig. 9. Effect of axis of rotation and target distance on mean VVOR and VOR gains of 15 subjects undergoing sinusoidal rotation at 1.2 Hz about multiple axis eccentricities while regarding targets at various distances (D, in centimeters). Differences in gain caused by variations in target distance were statistically significant (p < 0.01) at all eccentricities and frequencies except those marked with asterisks. Solid lines represent theoretic gains required to stabilize the image of the target on the retina. Note that the theoretic lines intersect at the eyes (0 degrees eccentricity). Error bars are drawn to represent the ±1 SEM, although the errors are often smaller than the plot symbols. Upper panel, VVOR gain measured in the light with a visible target. VVOR gain was progressively increased by target proximity for axis eccentricity posterior to eyes and progressively decremented by target proximity for eccentricity anterior to eyes. Effect of target distance on VVOR gain was least at an eccentricity 4 cm posterior to eyes. For VVOR data at frequencies greater than 1.2 Hz, variations in target distance were not statistically significant (p > 0.01) at eccentricity of 7 cm. Lower panel, VOR gain measured in darkness with remembered targets. VOR gain was progressively increased by target proximity for axis eccentricity posterior to eyes and progressively decremented by target proximity for eccentricity anterior to eyes. Gain changes were smaller than during rotations with a visible target (VVOR). Effect of target distance on VOR gain was least at an eccentricity 7 cm posterior to the eyes, corresponding to location of otoliths.
VVOR gain became independent of target distance was between the eyes and the otoliths. We interpret these findings to indicate that otolithic input interacts with canal input to provide gain enhancement during eccen-
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
tric rotations and that this effect may be further modified by vision during steady-state rotation in light. For both pitch and yaw rotations very posterior to the eyes (i.e., 15 to 20 cm), VVOR gain with 10 cm targets did substantially exceed unity at relatively low frequencies (i.e., 0.8 Hz), but gain declined with increasing frequency up to 2.0 Hz, the highest frequency tested. This VVOR gain enhancement is similar to the behavior of pursuit tracking, which has a declining gain with increasing stimulus frequency. During sinusoidal eccentric rotation in both pitch and yaw, the VOR exhibited a phase change systematically related to frequency of head rotation. With the axis of rotation well behind the otoliths, near target viewing was associated with a gain increase and phase lag, both increasing with frequency. With the axis of rotation anterior to the otoliths, near target viewing was associated with a gain decrement and with phase lead increasing with frequency. We modeled this behavior as a linear interaction between canal input, and otolith input scaled according to the inverse of target distance.18,19 The predictions of this simple model are consistent with the trends in the data, supporting the concept of a linear interaction between canal and otolith inputs. For rotations in the light, pursuit tracking probably adds another contribution to the response. The presence of an actual visible target also allows an accurate and compelling estimate of target distance. It is desirable to optimize control of subjects’ mental sets by providing an actual visible target but to be able to study purely vestibular responses independent of the confounding effects of visual tracking. We were able to do this by studying the initial VOR response in the presence of visible targets. In such tests only the initial compensatory eye movements are analyzed before the availability of visual feedback to generate pursuit eye movements.17,19 Because pursuit latency exceeds 100 msec, we limited analysis to the first 80 msec of the response. Unpredictable, passive, whole-body transient rotations were used to avoid predictive eye movements and prevent the use of motor efference copy to augment the VOR. Using sampling rates of 800 Hz, we have found it feasible to analyze VOR and VVOR responses to pseudorandom sequences consisting of about 20 unpredictable changes in rotator position. Automated analysis programs identified head acceleration events and then automatically fit regression lines to head and eye signals in either the position (yaw) or velocity (pitch) domains, depending on the characteristics of the head movement (Fig. 10). We have been able to reliably compute initial gains and to show dependence of gain on rotational eccentricity and target proximity, as illustrated in Fig. 11. We found no effect of axis or target distance on the initial 25 msec of the VOR response; however, by 30 to
DEMER and CRANE
85
Fig. 10. Typical responses of one subject to unpredictable transients of head rotation in the light (VVOR) about oculocentric and eccentric axes and with targets located at either 10 or 600 cm. Rightward rotations are positive. Filtering at 43 Hz to minimize noise precludes visualization of latency between the head and eye velocity profiles. Perturbations in eye velocity were commonly observed near 30 and 60 msec. Upper panel, VVOR during viewing a target at 10 cm with a posterior eccentric rotational axis (20 cm) and an oculocentric axis (0 cm). Responses were similar until 70 msec, after which a greater eye velocity was observed for the eccentric axis. Lower panel, VVOR during posterior eccentric rotation of 20 cm but with targets at 10 and 600 cm. The responses were similar up to 40 msec, after which the response was greater for the near target.
60 msec, clear effects of both axis of rotation and target distance were evident. Because analysis the ocular response from 25 to 80 msec prevents any direct contribution of visual tracking to observed gain, we attribute gain enhancement during rotations in the light to an effect of mental set. Despite an earlier report by others,20 we are not confident of the reliability of these data in estimating VOR latencies because torque limitations of the rotators make the onset of head rotations gradual, and the low early signal-to-noise ratio necessitates considerable low-pass filtering.
86
DEMER and CRANE
Fig. 11. Effect of rotational axis eccentricity and target distance (D, in centimeters) on mean VOR gains of 15 subjects undergoing unpredictable transient rotation. Points represent average of both VOR and VVOR data because visual feedback had no significant effect on gain. Solid lines, Theoretical gains required to stabilize image of the target on the retina. Error bars ± 1 SEM. Upper panel, Gain measured over first 25 msec of response was independent of both rotational axis and target distance ( p > 0.05). Lower panel, Gain measured from 25 to 80 msec after head rotation onset was progressively increased by target proximity for axis eccentricity posterior to eyes and decreased by target proximity for axis eccentricity anterior to eyes. Gain changes were smaller than during sinusoidal rotations under either lighting condition.
Implications
These data indicate that during ambulation a strong counterphase relationship exists between dorsoventral translation and pitch,21-23 as well as between mediolateral translation and yaw.12 These relationships were observed in every trial and every subject tested, so they are clearly robust. Perhaps the counterphase relationships are simply the result of passive head inertia. This idea is mechanically plausible for the pairings of pitch with dorsoventral translation, and yaw with mediolater-
Otolaryngology– Head and Neck Surgery July 1998
al translation, because the center of mass of the head is anterior to the support point of the head on the neck. Participation of active neural reflexes is also a possibility. Head stabilization during angular perturbations of the torso is believed to be accomplished by voluntary mechanisms at frequencies below 1.0 Hz, by the vestibulocolic and cervicocolic reflexes from 1.0 to 2.5 Hz and by mechanical considerations at higher frequencies.24,25 The 1 to 2 Hz fundamental frequencies of head motion during natural gait7,26 overlap the frequency range of active reflex head stabilization. Support for an active contribution is provided by the observation that patients with vestibular loss lack the counterphase relationship between dorsoventral translation and pitch21 and the observation that in normal walking subjects the ratio between pitch head rotation and dorsoventral translation is altered by reducing target distance.23 These observations suggest that the relationship between head translation and rotation may at least partially be governed by an active reflex system influenced by target distance, similar to the effect of context on head-neck dynamics hypothesized by Keshner et al.24 This reflex system may be deranged in patients with labyrinthine dysfunction who exhibit significantly more translational and rotational oscillations of head position than normal, especially in the vertical direction.10 Geometric analysis of the effects of head translation and rotation suggests that these two types of motion, when phase locked in the way observed here, act in a mutually compensatory manner to enhance gaze stability for typical target distances. There is an important implication of this finding for interpretation of tests of vestibular function. Conventional rotational clinical tests of the vestibular system involve passive rotation in darkness of the head and body about some vertical axis, at constant velocity or at sinusoidal frequencies from 0.0125 to 1.0 Hz.27 Often no care is taken to determine the precise axis of rotation, and no effort is made to control the subject’s estimate of target distance. We have shown here that both of these factors have important effects on optimal and on observed VOR gain. A further issue is possible adaptation to a habitual mode of eye-head coordination during natural activities. For example, the elderly subjects studied here exhibited lower angular VOR gains during most natural activities than did the young subjects. Does the decline in angular VOR gain represent deterioration of vestibular function with aging, and should this be expected to be reflected in poorer balance and dynamic visual acuity? Such questions cannot be answered with angular VOR data alone because we now appreciate that head translation, which is greater in the elderly, supplements the angular VOR in stabilizing gaze during ambulation for
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
typical target distances. Measurement of the VOR in all degrees of translational and rotational freedom, as illustrated in Fig. 6, would permit full characterization of ocular stabilization during natural motion. Measurements of VVOR gain were also made during ambulation while subjects wore ´2 telescopic spectacles. Although statistically significant VVOR gain enhancement with telescopic spectacles did occur during the three natural activities, the increase in gain was typically only about 15% to 20% and was thus far below the compensatory value equal to the magnification of the spectacles worn. During self-generated head rotations in both yaw and pitch, the same subjects attained very large augmentations of VVOR gain to values exceeding 1.5, although never quite equal to telescopic spectacle magnification of ´2. Experiments with telescopic spectacles of up to ´6 power have indicated that VVOR gain during self-generated head movements can exceed 3.0.15 Therefore the surprising finding was that during natural activities magnified vision was largely ineffective in significantly augmenting the VOR to accomplish gaze stabilization. This implies that patients with visual impairments requiring intermittent wearing of telescopic spectacles will be faced with substantial uncompensated retinal image instability during standing and ambulation.1,28 In the current studies performance during “autorotational” VOR and VVOR testing was not at all representative of vestibular performance during even the most common of natural activities. Perhaps this is because head movements during natural activities have a high frequency content and a major element of randomness29; because these head perturbations are truly unintended, motor efference copy is presumably unavailable to use in gaze stabilization. Enhancement of VVOR gain with telescopic spectacles was considerably greater during passive, predictable, whole-body rotations up to about 2.0 Hz than was observed during natural activities,3 although still less than during volitionally generated head rotations at similar frequencies.16 Because in both cases testing was with equally predictable sinusoidal motion, the greater gain enhancement during self-generated head movement might be attributed to motor efference copy or preprogramming, or perhaps to neck proprioception or the cervico-ocular reflex. The latter is generally considered to be insignificant in human beings30,31 but may become significant in subjects with compensated peripheral vestibular loss.32,33 In contrast to the small immediate influence of magnified vision on VOR gain during natural activities, data on VOR initiation indicate that intended target distance can modify the response to combined semicircular canal and otolith stimulation within a latency short-
DEMER and CRANE
87
er than that of visual tracking. Central regulation of the response to otolithic stimulation may be one means by which gaze stabilization is controlled during combined angular and linear head motion. Modeling is compatible with the idea that responses to otolithic stimulation are scales by the inverse of intended target distance. The findings summarized here indicate that angular VOR performance depends on intended visual target distance, which interacts with head translation. The result is consistent with recent theories of the VOR reflecting the idea that the vestibular input is only one input to a larger gaze control system.34-36 The VOR thus does not behave stereotypically for all head movements. We suggest that the vestibular system should be tested under more natural conditions7 and that both rotational and translational head movements be tested in the context of visual requirements. Results of such tests should be functionally validated by correlation with the most fundamental functional purpose of VORs, gaze stabilization. Nicolasa De Salles and Troy Howard provided technical assistance with the experiments. REFERENCES 1. Demer JL, Porter FI, Goldberg J, et al. Validation of physiologic predictors of successful telescopic spectacle use in low vision. Invest Ophthalmol Vis Sci 1991;32:2826-34. 2. Demer JL, Goldberg J, Porter FI. Effect of telescopic spectacles on head stability in normal and low vision. J Vestib Res 1991;1: 109-22. 3. Demer JL. Mechanisms of human vertical visual-vestibular interaction. J Neurophysiol 1992;68:2128-46. 4. Lisberger SG, Evinger C, Johanson GW, et al. Relationship between eye acceleration and retinal image velocity during foveal smooth pursuit in man and monkey. J Neurophysiol 1981; 46:229-49. 5. Pozzo T, Berthoz A, Lefort L. Head kinematic during various motor tasks in humans. Prog Brain Res 1989;80:377-83. 6. Grossman GE, Leigh RJ, Abel LA, et al. Frequency and velocity of rotational head perturbations during locomotion. Exp Brain Res 1988;70:470-6. 7. Grossman GE, Leigh RJ, Bruce EN, et al. Performance of the human vestibuloocular reflex during locomotion. J Neurophysiol 1989;62:264-72. 8. Demer JL, Amjadi F. Dynamic visual acuity of normal subjects during vertical optotype and head motion. Invest Ophthalmol Vis Sci 1993;34:1894-906. 9. Grossman GE, Leigh RJ. Instability of gaze during locomotion in patients with deficient vestibular function. Ann Neurol 1990;27: 528-32. 10. Takahashi M, Hoshikawa H, Tsujita N, et al. Effect of labyrinthine dysfunction upon head oscillation and gaze during stepping and running. Acta Otolaryngol (Stockh) 1988;106:348-53. 11. Gagnon J, Roth JM, Carroll M, et al. SuperANOVA. Berkeley: Abacus Concepts; 1989. p. 185-218. 12. Demer JL, Viirre ES. Visual-vestibular interaction during standing walking, and running. J Vestib Res 1996;6:295-313. 13. Demer JL. Effect of aging and visual impairment on dynamic visual acuity during vertical motion. In: Noninvasive assessment of the visual system. 1993 Technical digest series. Washington (DC): Optical Society of America; 1993. p. 192-5.
88
Otolaryngology– Head and Neck Surgery July 1998
DEMER and CRANE
14. Demer JL. Effect of aging on vertical visual tracking and visualvestibular interaction. J Vestib Res 1994;4:355-70. 15. Demer JL, Oas JG, Baloh RW. Visual-vestibular interaction during high-frequency, active head movements in pitch and yaw. In: Cohen B, Tomko DL, Guedry F, editors. Sensing and controlling motion. Vestibular and sensorimotor function. New York: New York Academy of Sciences; 1992. p. 832-5. 16. Demer JL, Oas JG, Baloh RW. Visual-vestibular interaction in humans during active and passive, vertical head movement. J Vestib Res 1993;3:101-14. 17. Viirre ES, Demer JL. The human vertical vestibulo-ocular reflex during combined linear and angular acceleration with near target fixation. Exp Brain Res 1996;112:313-24. 18. Crane BT, Viirre ES, Demer JL. The human horizontal vestibulo-ocular reflex during combined linear and angular acceleration. Exp Brain Res. In press 1996. 19. Crane BT, Viirre ES, Demer JL. Effect of target distance and eccentric rotation on human horizontal vestibulo-ocular reflex (VOR). Soc Neurosci Abstr 1995;21:519. 20. Johnston JL, Sharpe JA. The initial vestibulo-ocular reflex and its visual enhancement and cancellation in humans. Exp Brain Res 1994;99:302-8. 21. Pozzo T, Berthoz A, Lefort L, et al. Head stabilization during various locomotor tasks in humans. II. Patients with bilateral peripheral vestibular deficits. Exp Brain Res 1991;85:208-17. 22. Hirasaki E, Kubo T, Nozawa S, et al. Analysis of head and body movements of elderly people during locomotion. Acta Otolaryngol (Stockh) 1993;501(Suppl):25-30. 23. Bloomberg JL, Reschke MF, Huebner WW, et al. The effects of target distance on eye and head movement during locomotion. In: Cohen B, Tomko DL, Guedry F, editors. Sensing and controlling motion: vestibular and sensorimotor function. New York: New York Academy of Sciences; 1992. p. 699707. 24. Keshner EA, Cromwell RL, Peterson BW. Mechanisms control-
25. 26.
27. 28. 29. 30. 31. 32. 33. 34. 35. 36.
ling human head stabilization. II. Head-neck characteristics during random rotations in the vertical plane. J Neurophysiol 1995; 73:2302-12. Keshner FA, Peterson BW. Mechanisms controlling human head stabilization. I. Head-neck dynamics during random rotations in the horizontal plane. J Neurophysiol 1995;73:2302-12. Keshner EA, Cromwell R, Peterson BW. Head stabilization during vertical seated rotations and gait. In: Woollacott MH, Horak F, editors. Posture and gait: control mechanisms. Eugene (OR): University of Oregon Books; 1992. p. 105-8. Jenkins HA, Honrubia V, Baloh R. Evaluation of multiple-frequency rotatory testing in patients with peripheral labyrinthine weakness. Am J Otolaryngol 1982;3:182-8. Demer JL, Porter FI, Jenkins HA, et al. Predictors of functional success in telescopic spectacle use by low vision patients. Invest Ophthalmol Vis Sci 1989;30:1652-65. Maas EF, Huebner WP, Seidman SH, et al. Behavior of human horizontal vestibulo-ocular reflex in response to high-acceleration stimuli. Brain Res 1989;499:153-6. Berthoz A. Adaptive mechanisms in eye-head coordination. Rev Oculomot Res 1985;1:177-201. Jurgens R, Mergner T. Interaction between cervico-ocular and vestibulo-ocular reflexes in normal adults. Exp Brain Res 1989; 77:381-90. Chambers BR, Mai M, Barber HO. Bilateral vestibular loss, oscillopsia, and the cervico-ocular reflex. Otolaryngol Head Neck Surg 1985;93:403-7. Kasai T, Zee DS. Eye-head coordination in labyrinthine-defective human beings. Brain Res 1978;144:123-41. Collewijn H. The vestibulo-ocular reflex: an outdated concept? Prog Brain Res 1989;80:197-209. Collewijn H. The vestibulo-ocular reflex: is it an independent subsystem? Rev Neurol (Paris) 1989;145:502-12. Steinman RW, Kowler E, Collewijn H. New directions for oculomotor research. Vision Res 1990;30:1845-64.
Endoscopic and Image Guided Sinus Surgery
The Joint Center for Otolaryngology, Beth Israel Deaconess Medical Center and Brigham and Women’s Hospital of Harvard Medical School, is sponsoring a course, “Advanced Endoscopic and Image Guided Sinus Surgery—Beyond the Basics,” Nov. 6-7, 1998, at the Fairmont Copley Plaza Hotel in Boston, Mass. This year’s honored guest is Professor Heinz R. Stammberger of Graz, Austria. This course provides upto-date, state-of-the-art information and instruction to the practicing otolaryngologist regarding advanced techniques in endoscopic sinus surgery. For further information, contact Harvard MEDCME, PO Box 825, Boston, MA 02117-0826; phone (617)432-1525.