Journal of Vestibular Research, Vol. 3, pp. 101-114, 1993 Printed In the USA. All rights reserved.

0957-4271/93 $6.00 + .00 Copyright © 1993 Pergamon Press Ltd.

VISUAL-VESTIBULAR INTERACTION IN HUMANS DURING ACTIVE AND PASSIVE, VERTICAL HEAD MOVEMENT

Joseph L. Derner, MD, PhD , * t John G. Oas, MD, t and Robert W. Balah, Mot

*Jules Stein Eye Institute and tDepartment of Neurology, University of California at Los Angeles Reprint address: Joseph L. Derner, M.D., Ph.D., Jules Stein Eye Institute, Comprehensive Division ,

100 Stein Plaza, UCLA, Los Angeles, CA 90024-7002

o Abstract - We studied visual-vestibular interac­tion (VVn in 9 normal human subjects using active and passive vertical head rotations. Gain and phase of the vertical vestibulo-ocular reflex (VOR) and visually enhanced vestibulo-ocular reflex (VVOR) were measured for single frequency sinusoidal mo­tion, as well as for sinusoidal motion of continu­ously increasing frequency, over the range of 0.4 to 4.0 Hz. In addition to measurement of VVOR during normal vision, telescopic spectacles having a magnification of 1.9x were used to challenge VVI to facilitate measurement of visual enhancement of VOR gain. In the mid.frequency range (1.6 to 2.4 Hz), the active VOR exhibited gain closer to compensatory than did the passive VOR; at other frequencies, active and passive VOR gains were similar. VVOR gain during normal vision was com­pensatory for both active and passive motion throughout the frequency range tested. VVOR gain with 1.9x telescopic spectacles was greater than VOR gain at all frequencies tested, including up to 3.2 Hz for passive head movements, and up to 4.0 Hz for active head movement. However, gain enhancement with telescopic spectacles was consis­tently greater during active than during passive head movement. Phase errors for the VOR and VVOR were smaJ[ under all testing conditions. AI­thougt. active VOR and: VVOR were directionall~'

symmetrical, gain of upward slow phases differed from that of downward slow phases for passive VOR and VVOR in a manner depending on rota­tional frequency. For both active and passive test­ing, gain and phase values obtained during swept frequency rotations were similar to those obtained

Presented at the Society for Neuroscience Annual Meet­ing, New Orleans, Louisiana, November 12, 1991.

during single frequency sinusoidal testing. These data indicate that VVI can enhance gain of the pas­sive vertical VOR even at frequencies above what is usually considered to be the upper limit of visual pursuit tracking. The additional enhancement ob­served during active head movements at these bigh frequences is attributable to use of efference copy of the skeletal motor command to neck musculature.

o Keywords - active head movement; passive head movement; vertical, visual-vestibular interaction.

Introduction

A large body of data indicates that the hori­zontal (1-4) and vertical (4-6) eye movements produced by the vestibulo-ocular reflex (VOR) imperfectly compensate for passively imposed head movements. Perfect compensation would imply that VOR gain (eye velocity divided by head velocity) be equal to 1.0, indicating that I the reflexive eye velocity is equal and oppo­site to head velocity. The presence of normal vision is extremely effective in improving the precision 0: compensatory eye movements dur­ing pass!v::, wioie-body ::'o;.ation, wj:~ gair. at low frequencies usually within a fe~.' percent of the ideal value, both for horizontal (7,8) and vertical eye movements (6). However, even during normal vision, gain departs from unity for either high rotational frequencies (4,8) or, somewhat paradoxically, small rota­tional amplitudes (9). Since high frequency ro­tations of the head are present over a large range of amplitudes during normal standing

RECEIVED 15 June 1992; REVISED MANUSCRIPT RECEIVED 24 August 1992; ACCEPTED 26 August 1992.

101

102

and ambulation (10), and since inappropriate gain leads to retinal image motion and con­sequent impairment in visual acuity (11,12), it is functionally advantageous for the ocular motor system to use other inputs to optimize gain.

Many everyday head rotations are self­generated, which allows the possibility that knowledge of the intended head movement could be used to optimize VOR gain. Indeed, nemons in the reticul:1r ~()r:naticn {!3, ld' 'l!ld

in the superior ,;oilicuius (15) that exhibit 4aze movement activity project to both oculomo­tor and head motor control centers. For ac­tively generated movements up to 5.0 Hz, the horizontal VOR has a gain of unity, unlike during passive movement (16). However, Skavenski and collaborators studied both small horizontal and vertical, active and pas­sive head movements over a broad frequency spectrum and found that normal vision in­creased gain from 0.65 in darkness to a maxi­mum of only 0.90, and then only at frequencies below 3.0 Hz (9). These differences are diffi­cult to interpret, partly because the VOR is normally close to compensatory and because of the small amount of gain enhancement re­quired by normal vision. There is a paucity of data on the effect of active as compared with passive head movement on visual-vestibular interaction (VVI) , particularly for vertical movement.

The literature would suggest that the visual contribution to gain enhancement should be limited to relatively low frequencies in the physiologic range. Effective horizontal (17) and vertical (6) pursuit tracking are commonly said to be limited to frequencies below about 2 Hz, while the spectrum of physiologic head movements during standing and walking ex­tends to appreciably greater frequencies (4,10). This limitation on vertical tracking would sug­gest that other mechanisms besides pursuit might make important contributions to VVI at high frequencies.

Telescopic spectacles afford an improved opportunity to study enhancement of VOR gain by vision. These devices are telescopes having magnification from around 2x to 8x

J. L. Demer et al

mounted in spectacle frames for use as visual aids for patients suffering from ocular dis­eases (18). Telescopic spectacles magnify the visual effects of head motion in proportion to their optical power, and thus require compen­satory eye movements with gain substantially greater than the ordinary ideal value of unity during normal vision (7). Magnification per­mits the gain-enhancing effect of vision to be more easily demonstrated, even at its upward limit of frequency. The present investigation ~hl:s ~!TIployed telescopic spectacles as a chal­lenge : 0 VVI and compared VVI during active vertical head movements with those during passive vertical head movements.

Materials and Methods

Subjects

Nine young adult, paid volunteers gave writ­ten informed consent to a protocol approved by the Human Subject Protection Committee. All underwent ophthalmological examination by the author, verifying visual acuities in each eye correctable to 20120 or better . Average age of subjects was 30.2 ± 6.8 years (mean ± standard deviation, SO, range 19 to 39); there were 6 women and 3 men. During experi­ments, the comfort of subjects was monitored using an intercom and closed circuit television.

Apparatus

The host computer for these experiments was a Macintosh II (Apple Computer, Cuper­tino, CA) equipped with analog-to-digital converter (ADC), digital-to-analog converter (DAC), and direct memory access devices (Na­tional Instruments, Austin, TX). Custom soft­ware was written for these experiments using the Lab View software package (National In­struments). Sampling was at 200 Hz for stim­ulus frequencies below 1.0 Hz, and 400 Hz for greater stimulus frequencies. Digitizer sensitivity was increased up to 8-fold accord-

Active and Passive Visual-Vestibular Interaction

ing to stimulus amplitude, increasing the ef­fective performance of the ADC to 15 bits as necessary.

Subjects were seated in a custom fabricated swing rotator having a horizontal interaural axis. Their heads were firmly strapped to a headrest on the rotator, and their torsos and extremities were comfortably secured to the rotator using harnesses and belts. The rotator was driven by a servomotor (l08 ft-lb, 4.5 kW, Inland Motors, Radford, VA) controlled by a computer via a DAC. This permitted sinu­soidal rotation of the subject's head and body at peak velocities of up to 40° to 75°/s, de­pending on subject weight and body habitus. Swing rotator position was registered by a pre­cision potentiometer. Pairs of magnetic field generator coils 1 m in diameter were mounted on the swing rotator to permit measurement of horizontal and vertical eye and head posi­tion using the magnetic search coil technique (19). Subjects wore a headband to which was attached a search coil for the measurement of head position relative to the swing rotator, since decoupling of head, body, and swing motion often occurred at high frequencies de­spite restraints. The signal from the head coil was used to compensate for decoupling, so that corrected eye-in-head and head-in-space measurements were analyzed. The mechani­cal properties of the swing rotator depended upon loading by individual subjects, particu­larly with respect to weight and weight dis­tribution. Many subjects' bodies exhibited mechanical resonances within the frequency range tested. Due to motor torque limitations in control of the swing rotator, as well as de­coupling of the head from the rotator. it was impossible to precisely comroi head velocity. Data are reported here at nominal velocities of the swing rotator, but gains were always computed using measured head velocities. Generally, head velocities were close to nom­inal at 0.8 Hz, exceeded nominal from 1.2 to 2.0 Hz, and declined to less than nominal at higher frequencies. Observed gain values were not strongly dependent upon variations in head velocity within the range of head veloc­ities employed.

103

For active head rotations, subjects were trained to make sinusoidal vertical head move­ments in synchrony with the pitch of an au­dible tone presented via a loudspeaker. The tone was provided by a loudspeaker driven by a voltage controlled oscillator under the control of a computer via a DAC. Head move­ments were recorded using a search coil at­tached to a headband, and subjects remained seated in the locked swing rotator with head restraints removed. During training, head po­sition and velocity were monitored in real time on a polygraph, and verbal feedback was given to subjects regarding the amplitude, fre­quency, and distortion of their head move­ments. With such feedback, most subjects learned to make sinusoidal head movements of the correct frequency and appropriate ve­locity amplitude after less than 10 min of training. The lowest frequencies were most difficult to produce without distortion, and maximum frequencies varied from subject to subject. Single frequency testing was generally performed at a constant velocity amplitude in the range of 40° to 1000/s, at the following approximate single frequencies: 0.5, 1.0, 1.6, 2.0,2.4,3.0, and 4.0 Hz. In each case, the ac­tual frequency achieved in the trial was veri­fied by Fourier analysis. Not every subject achieved every frequency, and some were un­able to achieve 4.0 Hz. Some subjects also made active head movements of constant fre­quency, but progressively increasing veloc­ity amplitude, from 30° to 150°/5. Active frequency sweeps were also made using con­stant velocity amplitude, from 0.4 to 4.0 Hz over 20 to 25 seconds.

The apparatus was located in a roOlT. tha~ couie. be made to:aI!y dark; i; faced c; gra~ '

waH locatee! 3 m from Ine subject's eyes. ro, VVOR testing, the wall was illuminated with standard fluorescent lighting to a centrallu­minance of 50 Cd/m2 • On the wall directly in front of the subjects was a small cross flanked by 4 red bars extending its arms, serving as the central fixation target. Four other crosses flanked the central target at 10° horizontal and vertical eccentricity from it, serving as horizontal and vertical calibration targets. The

104

remainder of the central 10° of the wall fac­ing the subjects was filled by concentric dia­monds comprised of diagonal stripes spaced about 2 ° apart.

Measurement Conditions

Rotational stimuli consisted of trains of approximately 8 cycles of single frequency sinusoids, as well as trains of sinusoids I)f ~on­tinuously inc:-easing ~requenc:, over a ;ogarith­mic progression. For passive rotations, the command to the servomotor was set for a con­stant velocity amplitude of 300/s, and the fol­lowing single frequencies were tested: 0.8, 1.2, 1.6, 2.0, 2.4, and 3.2 Hz. The passive fre­quency sweeps also had a constant nominal velocity amplitude of 300/s and a frequency increasing in continuous, logarithmic fashion over 20 seconds, from 0.4 to 3.2 Hz. Actual velocity amplitudes varied somewhat with the size and weight of the subject, and generally were less than nominal above 2.4 Hz.

Unaided visual-vestibulo-ocular reflex (1 x VVOR condition) testing was performed in normal room illumination with instructions to the subject to ftxate the central target directly ahead. Magnified VVOR testing was per­formed during wearing of 1.9x binocular tele­scopic spectacles having a field diameter of -16.8°, with instruction to attend to visual details on the wall ahead. The visual field pe­ripheral to the telescopes was occluded. The VOR was tested in total darkness with instruc­tions to ftxate the remembered central ftxation target on the wall ahead.

Eye and Head Movement Measurement

Vertical eye position was measured using the magnetic search coil technique, with the scleral search coil contact lens (Skalar Medi­cal, Delft, The Netherlands) (20). Topical pro­paracaine 0.51170 was applied to the eye before insertion of the search coil contact lens, and the lens was removed within 30 minutes. Cal­ibration was obtained for vertical saccades to targets located at 10° eccentricity from the pri-

J. L. Derner et al

mary position. True head position in space was taken as the sum of swing rotator posi­tion plus head position relative to the swing rotator, thus compensating for any possible decoupling of the head from the rotator. To calibrate the sensitivity of the search coil on the head, the subject was instructed to fixate a target at the end of a 25-cm rigid rod held tightly in the teeth, as head position was changed in roughly 10° increments about pri­:nar~1 ~osition. Since the sensitivity of the searc!1 -:oil on the eye had 'Jeen previously cal­ibrated, the change head position was thus cal­ibrated against the search coil on the eye. Eye, head, target, and chair rotator position data were displayed on a polygraph, with electronic differentiation for real time monitoring of ve­locities. Data were recorded in pulse code modulated analog format on magnetic tape for possible later editing, and were digitally sampled and stored to hard disc memory.

Data Analysis

Data were automatically analyzed using custom software under the Lab View package. All sampled channels were digitally low pass filtered (6 pole Butterworth, 0 to 42 Hz) and differentiated. Large quick phases and sac­cades were removed using velocity and dura­tion criteria. Cross spectral analysis of eye compared with head velocity was then per­formed to determine the phase shift of the eye velocity response. After correction for phase, a linear regression of eye velocity against head velocity was obtained, and data points found to be statistical outliers from this ftt were also excluded as quick phases. Separate regressions were performed for the upward and down­ward directions, and the slopes so determined were taken to be the gains in each direction. The direction of gain is considered to be that of the slow phase response. After removal of quick phases, the remaining slow phase data points were fit cycle-by-cycle by least squares to a sinusoidal equation. Head veloc­ity, eye velocity, phase, and gain (eye velocity/ head velocity) were computed for each cycle, and outlying cycles were automatically ex-

Active and Passive Visual-Vestibular Interaction

eluded using a statistical criterion previously described to be effective for the removal of low gain artifacts (3). Points removed as quick phases and artifacts were then replaced by velocity values computed from the linear re­gressions, phase shifts, and instantaneously measured head velocities. Fourier spectral analysis was repeated to compute gain and phase at each of a range of frequencies. Phase lags were taken to be negative, while leads were taken to be positive. An ideal phase of 180° was taken as implicit for the VOR and VVOR, and reported phase values are relative to this. Values obtained by Fourier analysis were considered reliable only if the coherence at the single frequency in question exceeded 0.80; a coherence of 0.96 or greater was re­quired for frequency sweeps. Typical values of coherence for single frequency VOR and VVOR testing exceeded 0.98. Since relatively few data are obtained at anyone frequency for sweep testing, higher coherence values were required to assure reliability; typically this affected only scattered and extreme fre­quencies in the spectrum for VOR and VVOR.

In order to correct for calibration errors in­herent in the derivation of passive gains from eye, head, and chair position, 1 x VVOR gain at 0.8 Hz as computed by Fourier analysis was assumed to be the ideal value of 1.0. "For each individual subject, all other measurements of passive VOR and VVOR gain were normal­ized against 1 x VVOR gain at 0.8 Hz. No normalization was performed for active head head rotations.

Results

Single Frequency Rotations­Gain and Phase

For both active and passive vertical head movements, normal and telescopically mag­nified vision were effective in enhancing VOR performance. Under all viewing conditions, eye movements during passive rotations were opposite the direction of head movement, having slight phase variations as described be­low. Trials conducted with active head veloc-

105

ities varying over the range of 30° to 1500/s demonstrated that VOR and VVOR gain were independent of peak head velocity.

Gain for the VOR was typically less than 1.0 for both passive and active head rotations, while gain during normal viewing (1 x) was almost perfectly compensatory. During 1.9 x viewing, gain exceeded 1.0 at lower frequen­cies and declined at higher frequencies, but gain enhancement was greater during active head rotations than during passive head rota­tions for all frequencies tested. Typical data illustrating this effect for single frequency ro­tation at 1.2 Hz are shown in Figure 1, which compares active and passive VOR responses with active and passive VVOR responses dur­ing l.9x viewing. Pooled gain data of all sub­jects obtained using Fourier analysis are plotted against frequency in Figures 2 through 4. Fig­ure 2 illustrates that VOR gain during passive rotations was less than during active rotations between 1.2 and 2.4 Hz; within this frequency range, the active VOR was almost perfectly compensatory. Figure 3 illustrates that 1 x VVOR gain was essentially independent of frequency for both passive and active head ro­tations. The slight differences between gains obtained under the two rotational conditions could easily be attributable to variations in calibration of measurements, but all values closely approximate the ideal value of 1.0. Figure 4 illustrates the marked increase in VVOR gain produced by the wearing of 1.9x telescopic spectacles for both passive and ac­tive rotations. There was a trend toward greater gains during active head rotations, which was statistically significant (P < 0.01 two-tailed t­:es~1 a: :.0 and :.4 !iz. Gain enhancemen~ during magnified vision occurred even at the highest frequencies tested for both types of rotations.

Phase values during passive and active head rotations are also illustrated in Figure 5. Un­der all viewing conditions, there was a trend toward phase lead (positive phase) below about 1.5 Hz, often with a slight phase lag for greater frequencies. However, phase errors were uniformly less than 5° and were indepen­dent of whether rotations were passively or ac­tively induced. Phase compensation of the

106 J . l. Derner et al

Passive VOR (.) Q) U) --a

>--'(j 0

Q) > Q)

>-W

:..> Q) U) -­a

>­.'!::! (.) o Q) > ~

100

50

a

-50

-100 I

0 .0

~ -100 1 ~ 0.0

I 1.0

1 1.0

I 2.0

I

3 .0

I 3 .0

I 4 .0

Time - sec

::(. :,'. ) ~,, {.~

I 5 .0

I 0 .0

(.) Q) U) -­a

G = 0.76

100 -r---,---r----,.---

80~~-~--~-_4----

60~~~~~~~--+._-~

40~~~~~~~-_4---

20~-~~~~~-~+_---->­;:: ~ :J ~---"+.~rF_7~r._;;:....,._:_+..:......--.., ! ~ -20 'I-:-----+.~~~~~~,__--

i ~:~~ _ ;~~~ -II------il-----+----+--..,;~-

-100

Head Velocity - a/sec

Passive 1.9X VVOR

IV >- -100 I W 0.0

I 1.0

I 2.0

I 3 .0

I 4.0

I :5 .0

I 6.0

(.) 100~---,----~--~----r---~---­IV en -­a

I 1.0

I 2 .0

I 3 .0

I I 4.0 :5.0

Time- sec

I 6 .0

(a)

(.) Q)

J!! a

>­.'!::! (.) o

Q) > IV >­

W

G = 1.35

100 -...---...--..._..,.....,.-,.-----r----

80~~~~~_4----~-----

60~---~1!I

40~--~~~~--~-----

20 -+-----=-~~ ~~-+_--o -+------+"""'""!idIilIiI

-20 -+------+--~ -40-+---~--~~~~---

-60 +--+-+~.fu:---80 -+----+---_4~~.~~

-100 -100 -50 o 50

Head Velocity - a/sec

Figure 1. Sampled eye movement data for one representative subject illustrating enhancement of VOR gain by viewing wHh 1.9x telescopic spectacles (1.9 x viewing) during head movements at 1.2 Hz. Note greater gain enhancement during active head movement (b) than during passive head movement (a). For each condition, the lett graphs plot slow phase eye velocity (upper) and sampled head velocity (lower) against time. (Figure continues on facing page.)

Active and Passive Visual-Vestibular Interaction 107

Active VOR

.~ u o

CD -50~~~----~~-M~~~¥-~~-­> Q)

.1i -100 I 0.0

I 1.0

I 2.0

I 3.0

I 4.0

I 5.0

I 0.0

U 100~--~----~---r----r---'----­Q)

~ >--'0 0~4-~~-+~~~~~~-+~+-~ o ~ -504-~-4~~W--+~~--~~~-L-­'C ~ -100 I

:::z:: 0.0 I

1.0 I

2.0 I

3.0 I

4.0

Time- sec

I 5.0

I 0.0

(J Q)

~ o

>--'0 o (j) > Q)

>­w

G = 1.04

100-r~--~----~---'r-----

80-P~~~----~----4------

60~~~~~~~---+~---

40~--~~~~~----~-----

20~~~~~~~--~~-----

o ~--'-~r+-':""-...:Jt,1!I -20~~~~~--~

-40 -+---~'---~--:~ I!;~---

-60~~--+---~~--~~---

-80~----+---~~---+~~

-100 I -100

I -50

I a

I 50

Head Velocity - o/sec

I 100

Active 1.9X VVOR U 100~~-'----~--~----r---~----CD :.

~ >--'0 o

CD -50~~~~+*+-~~~--~~~~~ > Q)

.1i -100 I

0.0 I

1.0 I

3.0 I

4.0

Time - sec

I 5.0

I 0.0

(b)

G = 1.56

100--~~~--~----~-----

80-l---~~

~ 60 -I-----~ ~ 40 -+-----~~i!ii ~_+----_+,....-----o

20 -I---__ ~JO ~~----+_--->-:= 0-1---4-:.....:;1 (J o -20-l-----~~--~~~+----

~ -40~----~----~~~-----­~ -60~----~----~~~~----­

~ -8C~t------------~~~~~---

-iO: J, -100 -50 50 100

Head Velocity - o/sec

Figure 1 continued. Eye velocity data are fitted to sinusoids (solid curves) for those cycles meeting the sta­tistical criterion for acceptance, while cycles rejected due to artifacts are indicated by horizontal bars. For each condition, the right graphs plot phase-corrected eye velocity against head velocity. The solid lines Indicate linear regressions of eye velocity against head velocity, and their slopes correspond to gain. The two flank­ing lines Indicate boundaries of the region In which data points were accepted for analysis. G = gain averaged for upward and downward slow phases.

108

c .-CO ~

a: o >

J. L. Derner et al

1.75~--------------------------------------------------~

1.50

1.25

1.00~

J

0.75

Normal Subjects

* Active ..... 0-.... Passive

Mean ± SEM

T --&

0.50+-----~----~----~-----r----~----~----~----~~~

0.0 1.0 2.0 3.0 4.0

Frequency - Hz Figure 2. VOR gain of 9 subjects tested at multiple single frequencies during passive and active head rota­tions. Gain was greater during active rotations than during passive rotations in the mid-frequency range.

c .-m C!' a:: o > > ><

1.75.-------------------------------------------------~

1.50

1.25

1.00

Normal Subjects

* Active ·····0-···· Passive

~ ... D ........ ~ ........ * * ............... ¢ • 0-······-0-···

~ 0.75

Mean ± SEM

0.50 +-----......-----,.-----......-----,.-----......-----,.-----......-----,.---' 0.0 1.0 2.0 3.0 4.0

Frequency - Hz Figure 3. VVOR gain during normal viewing (1 x) of 9 subjects tested at multiple single frequencies during passive and active head rotations. Gain was similar for both types of rotations.

Active and Passive Visual-Vestibular Interaction

1.75...----------------------------, Normal Subjects

c "m 1.50

e" a: o > > >< 0)

•

1.25

1.00

0.75

* Active ---.. 0--'" Passive

*

Mean ± SEM

* P < 0.01

0.50+---..-----.--~----r---.----.----...----,....-..... 0.0 1.0 2.0 3.0 4.0

Frequency - Hz

109

Figure 4. VVOR gain during viewing with 1.9x telescopic spectacles in 9 subjects tested at multiple single frequencies during passive and active head rotations. Gain enhancement was greatest at low frequencies and was greater during active than during passive head rotations.

5

4

3

C) 2

(1) "'0

0

Cl) --CfJ -m J: -2

a. -3

-4

-5 0.0

0,

oa __ .o_o ____ o:~:~:o~:~:~~ ___ Active VOR

-.- Active 1 X VVOR --.- Active 1.9X VVOR ----0--- Passive VOR ----0--- Passive 1X VVOR ----fr--- Passive 1.9X VVOR

1.0 2.0 3.0

Frequency - Hz

w

4.0

Figure 5. Phase of the VOR, VVOR during normal viewing (1 x), and VVOR with 1.9x telescopic spectacles dur­ing active and passive head rotations at multiple frequencies in 9 subjects. Phase errors were small, but there was a tendency for phase lead (positive) at low frequencies.

110

VOR and VVOR was thus always essentially ideal.

Single Frequency Rotations­Directional Symmetry

Directional symmetry of the VOR and VVOR responses were evaluated under all testing conditions using phase-corrected Un­~3.r ;-eg:.-essicrrs of ~~re '/e!oc:!~, :lgainst head. 're­:oC:ty . These jara are ;Jiotted in Figure 5. :n general, pooled responses of multiple subjects during active rotations were directionally sym­metrical, and this was also typical of the re­sponses of each individual subject. In contrast, VOR and VVOR responses during passive ro­tations occasionally exhibited directional asym­metries, but these were not always consistent over all frequencies. Significant directional asymmetries were observed in pooled gain data only for the VOR and 1 x VVOR, and

J. L. Derner et al

only at some frequencies. Figure 6A illustrates that VOR gain for upward slow phases was significantly greater than for downward slow phases for passive rotations at 2.0 Hz (P < 0.01, two-tailed t-test), with a trend toward significance at both higher and lower frequen­cies. Figure 6B illustrates that 1 x VVOR gain was significantly greater for downward than for upward slow phases during passive rota­tions at 0.8 Hz, but was significantly greater for upward ,low phases from 1.6 to 2.4 Hz. Figure 5C :llustrates :hat 1.9x YVOR gain ex­hibited no significant directional asymmetries for either active or passive head rotation.

Frequency Sweeps - Gain and Phase

Frequency sweep waveforms permit the frequency spectrum of head rotation to be studied more rapidly than sinusoidal testing

1.75~--------------------------------------------------~

c: o-m (!)

a: 0 >

1.50

1.25

1.00

0.75

--.- Active Up

--- Active Down

.. ---/:l---. Passive Up ----0--- Passive Down

*

Normal Subjects

Mean ± SEM

* . P < 0.01

0.50+-----.....----r-------r-------r-------r---I 0.0 1.0 2.0 3.0 4.0

Frequency (Hz) (a)

Figure 6. Gain of 9 subjects analyzed for directional symmetry during passive and head rotations. Asterisk indicates statistical Significance of comparison of upward with downward gain for each type of rotation. (a) VOR was significantly asymmetrical only for passive rotation. (b) VVOR during normal viewing (1 x) was sym­metrical during active head rotation, but exhibited a complex asymmetry during passive rotation. (c) VVOR with telescopic spectacles (1.9 x ) was directionally symmetrical for active and passive rotation. (Figure continues on facing page.)

Active and Passive Visual-Vestibular Interaction 111

at discrete frequencies. Active and passive swept frequency testing was performed in all subjects for the VOR, 1 x VVOR, and 1.9x VVOR conditions. Gain and phase were ob-

tained using Fourier analysis, tabulated at the same frequencies used for single frequency testing. In general, gain and phase values ob­tained using frequency sweeps were quite sim-

s:: CO

(!)

c:: 0 > > >< ,..

1.75..-----------------------~

1.50

1.25

1.00

0.75

---a- Active Up Normal Subjects

---- Active Down

·····ll···· Passive Up ···-0··· Passive Down

~ ........ n ...... · =&= =-•.••• _.u. ~ .......... ~ ......... '¢..... ,1. •...•.•...•...•.•.... * * .... ~ ........ !

*

•

Mean ± SEM

* - p < 0.01

0.50+-------.,.--~---r---..------.--~---.----l

0.0 1.0 2.0 3.0 4.0

Frequency - Hz (b)

1.75 -,--------------------------,

Normal Subjects

s:: ._ 1.50

CO (!)

a: 0 > > >< en

• ,..

1.25

( .00 i

0.75

0.50 0.0

---t- Active Up ---- Active Down

·····ll···· Passive Up ····0-·· Passive Down

1.0

....... ~

...... ~ ... : .... ~ ... :: .. :.::::: .. :: ] '

1 "1" Mean ± SEM

2.0 3.0

Frequency - Hz (c)

Figure 6 continued.

4.0

112

Har to those obtained using single frequency testing, and the small differences were not statistically significant. The exception to this generality was for the active VOR at 0.4 and 1.0 Hz, where gains for sweep testing were about 25070 greater than during single frequency testing (P < 0.01).

Discussion

VOR Gain Enhancemem in Darkness

The present data indicate that the vertical VOR is more precise during active than dur­ing passive head rotation. This was most evi­dent at frequencies between 1.2 and 2.4 Hz, where VOR gain during active rotation was higher than during passive rotation, although gains were similar at lower and higher fre­quencies (Figure 2). Since the cervico-ocular reflex is insignificant in normal humans (21) and monkeys (22), the more precise VOR gain observed during active head movements could be explained through use of efference copy of the command to neck musculature to supple­ment the VOR. The anatomical substrate for such a sharing of efference copy is provided by neurons in reticular formation (13,14) and in the superior colliculus (15) encoding gaze movement signals that project to both oculo­motor and cerivcal motor centers.

Directional Symmetry

The present study is the first to examine the issue of symmetry of the passive vertical VOR and VVOR at frequencies exceeding 1.6 Hz. In this study, the vertical VOR and VVOR were found to occasionally exhibit compli­cated directional asymmetries, depending upon frequency. Asymmetries in human ver­tical VOR gain have not been consistently ob­served by other investigators (5,6), although asymmetries have been found in individual subjects. The asymmetries were small and fa­vored gain for downward slow phases below 1.0 Hz, but for upward slow phases from 1.6 to 2.4 Hz, so testing over a limited frequency

J . L. Derner et al

range might not disclose them or might yield inconsistent results. Asymmetries in VOR and VVOR gain observed during passive head ro­tation were corrected during active head ro­tation (Figure 6). In agreement with Ranalli and Sharpe, the present study found no direc­tional asymmetry of the active vertical VOR or VVOR (23). The correction of asymmetry during active head movement is probably also due to the availability of efference copy of the t:eac! :ncvement ,:ommaP..d. Consequently. since :nest natural pitch head :novements of 3ubstan­tial velecity are likely IO oe self-generated. asym­metries in the vertical VOR and VVOR are unlikely to be of behavioral significance in normal humans.

Visual Enhancement of Vertical VOR Gain

The present data confirm previous findings that the vertical VVOR in the presence of nor­mal vision is compensatory when tested both for active (2,5) and passive (6) head move­ments, despite considerable variations in VOR gain in darkness. This VVI is appropriate, but the amount of gain change required is small and not as challenging as that required dur­ing the wearing of telescopic spectacles. The present data also indicate that during passive, predictable vertical head movements consid­erable VVI occurs with telescopic spectacles up to frequencies of at least 2.0 Hz, and that phase is compensatory. For active head move­ments, VVI with telescopic spectacles had an even broader spectrum, being significant at all frequencies tested, even up to the highest fre­quencies subjects could voluntarily achieve. There was a consistent trend for VVOR gain with telescopic spectacles to be greater during active than passive rotations, reaching statis­tical significance at 2.0 and 2.4 Hz. Although predictive mechanisms have been demon­strated to contribute to gain enhancement dur­ing head motion with telescopic spectacles (24), the active and passive head movements studied here were equally predictable. It is likely that the superior VVI observed during active head movements is due to use of effer-

Active and Passive Visual-Vestibular Interaction

ence copy of the skeletal muscle command to the neck musculature. Efference copy during active head movements permits precise estima­tion of speed and timing, and probably ac­counts for the superior gain enhancement with telesopic spectacles observed during active as opposed to passive head movements (16). This performance may be further influenced by an­ticipatory intent (25).

High-Frequency VVI

Baloh and colleagues have found that ver­tical pursuit exhibits substantially declining gain as frequency increases toward 1.6 Hz (6). Such observations have led others to the con­clusion that normal vision (26) and other ex­tra vestibular influences have negligible effects on the VOR at frequencies above 2.0 Hz (16). However, other data indicate that if target ac­celeration is not excessive, pursuit at reduced gain does occur at high frequencies (17). This sort of visual tracking contribution may ex­plain the significant VVOR gain enhancement observed here at frequencies above 2.0 Hz. Even though compensation is imperfect, and insufficient to match the magnification of powerful telescopic spectacles, VVI contrib­utes in a functionally valuable manner beyond 2.0 Hz, and presumably throughout the phys­iologic range. This finding deserves consider­ation in the development of clinical tests of the active VOR.

Implications for Visual Function

Observing tha~ 1 x VVOR gair. du::-ing V~:-~" small amplitude head rotations is insufficient to prevent retinal image motion, Skavenski and colleagues have argued that VVOR gain is down-regulated to assure that a small amount of retinal image motion always occurs (9). This may be advantageous to prevent percep­tual fading of stationary images, but retinal image slip velocities exceeding 2°/s markedly degrade visual acuity (11). Although signifi­cant VVI was observed here with telescopic spectacles during even at high frequencies, it

113

should be stressed in the present case that the highest observed VVOR gains with telescopic spectacles were always less than the ideal val­ues equal to telescope magnification, and were comparable to those measured during a visual acuity task with telescopic spectacles where acuity was degraded by motion (11). This im­plies that retinal images are unstable during these head movements with telescopic specta­cles, and that this instability is functionally disadvantageous, probably reflecting limita­tions in the ocular motor system.

Swept Frequency Testing

As opposed to multiple cycles at each of several discrete frequencies, continuously in­creasing stimulus frequency over a short time interval represents one way of reducing the time required to test rotational VOR and VVOR performance over a spectrum of fre­quencies (16). While the swept frequency ap­proach suffers from the disadvantage. of having relatively little stimulus energy at any given frequency, the present data indicate that resulting gain and phase values are nonethe­less quite close to those obtained using discrete frequency testing in normal subjects. The only exception is for testing at "frequencies below 1.0 Hz, where even trained subjects experience difficulty making undistorted sinusoidal head movements. Use of swept frequency stimuli at frequencies above 1.0 Hz would thus permit essentially the same data to be obtained in about one-tenth the time required for discrete frequency testing, an advantage that might be of considerable value in clinical settings. now­~" e: " the vaiue of swept fiequency testing in subjects with vestibular disease remains to be demonstrated.

Acknowledgments- The authors thank Laura A. Hovis for assisting with recruitment and testing of subjects, and Douglas Wong and James Li for hardware and software support. This work was supported by U.S. Public Health Service grants EY-08656 and NS-10940. JLD is a Research to Prevent Blindness William and Mary Greve International Research Scholar.

114 J. L. Derner et al

REFERENCES

1. Barr CC, Schulteis LW, Robinson DA. Voluntary, non-visual control of the human vestibulo-ocular re­flex. Acta Otolaryngol (Stockh). 1976;81:365-75.

2. Ferman L, Collewijn H, Jansen TC, Van den Berg AV. Human gaze stability in the horizontal, vertical, and torsional direction during voluntary head move­ments, evaluated with a three-dimensional scleral in­duction coil technique. Vision Res. 1987;27:811-28.

3. Demer JL, Poner FI, Goldberg J, Jenkins HA, Schmidt K. Adaptation w telescopic spectacles: ves­tibula-ocular reflex plasticity. Invest Ophthalmol Vis Sci. ;989; .30: 1:9-70.

... ':Jrossman GE. ~;!Igh :U, -\bel LA. Lansb :)1, T!1urston SE. ?,equency :md 'Ielocity of rotational head perturbations during locomotion. Exp Brain Res. 1988;70:470-6.

5. Baloh RW, Demer JL. Gravity and the vertical ves­tibula-ocular reflex. Exp Brain Res. 1991;83:427-33.

6. Baloh RW, Honrubia V, Yee RD, Jacobson K. Ver­tical visual-vestibular interaction in normal human subjects. Exp Brain Res. 1986;64:400-6.

7. Derner JL, Goldberg J, Porter FI, Jenkins HA, Schmidt K. Visual-vestibular interaction with tele­scopic spectacles. J Vestibular Res . 1991;1:263-77.

8. Gauthier GM, Piron J-P, Roll J-P, Marchetti E, Martin B. High-frequency vestibulo-ocular reflex ac­tivation through forced head rotation in man. Aviat Space Environ Med. 1984;55:1-7.

9. Skavenski AA, Hansen RM, Steinman RM, Winter­son BJ. Quality of retinal image stabilization during small natural and artificial body rotations in man. Vision Res. 1979;19:675-83.

10. Derner JL, Goldberg J, Porter Fl. Effect of telescopic spectacles on head stability in normal and low vision. J Vestibular Res. 1991;1:109-22.

11. Derner JL, Amjadi F. Dynamic visual acuity of nor­mal subjects during vertical optotype and head mo­tion. Invest Ophthalmol Vis Sci. [In press).

12. Derner JL, Porter FI, Goldberg J, Jenkins HA, Schmidt K. Dynamic visual acuity with telescopic spectacles: improvement with adaptation. Invest Oph­thalmol Vis Sci. 1988;29:1184-9.

13. Grantyn A, Berthoz A. Reticulo-spinal neurons par­ticipating in the control of synergic eye and head movements during orienting in the alert cat; 1: Be­havioral properties. Exp Brain Res. 1987;66:339-54.

14. Whittington DA, Lestienne F, Bizzi E. Behavior of preoculomotor burst neurons during eye-head coor­dination. Exp Brain Res . 1984;55 :215-22.

15 . Munoz DP, Pelisson D, Guitton D. Movement of neural activity on the superior colliculus motor map during gaze shifts. Science. 1991;251:1358-60.

16. Tomlinson RD, Saunders GE, Schwartz DWF. Anal­ysis of human vestibulo-ocular reflex during active head movements. Acta Otolaryngol. 1980;90: 184-90.

17 . Lisberger SG, Evinger C, Johanson GW. Fuchs AF. Relationship between ~ye acceleration and retinal im­age '1eioc:ty .lUring :ov~ai ,mooth pursuit in man and monkey . j -IeurophyslOl. ! 981;~6:229-"9 .

18 . Derner jL. Poner ? !. Goldberg " l enkins HA, Schmidt K. Ulrich 1. Predictors of :unctional success in telescopic spectacle use by low vision patients. In­vest Ophthalmol Vis Sci 1989;30: 1652-65.

19. Fuchs AF, Robinson DA. A method for measuring horizontal and vertical eye movements chronically in the monkey. J Appl Physio!. 1966;21:1068-70.

20. Collewijn H, Van der Mark F, Jansen TC. Precise recording of human eye movements. Vision Res. 1975;15:447-50.

21. Bronstein AM, Hood JD. The cervico-ocular reflex in normal subjects and patients with absent vestibu­lar function. Brain Res. 1986;373:399-408.

22. Bohmer A, Henn V. Horizontal and vertical vestibulo­ocular and cervico-ocular reflexes in the monkey dur­ing high frequency rotation. Brain Res. 1983;227: 241-8.

23 . Ranalli PJ, Sharpe JA. Vertical vestibulo-ocular re­flex, smooth pursuit and eye-head tracking dysfunc­tion in internuclear ophthalmoplegia. Brain. 1988; Ill: 1299-317.

24. Derner JL. High frequency, vertical visual-vestibular interaction in humans: evidence for predictive mech­anisms. Invest Ophthalmol Vis Sci. 1992;33:1358.

25 . Zangemeister WH, Schluter D, Kunze K. Compen­satory eye movement gain and head-eye latencies change with verbal feedback: voluntary adjustment of gaze types. Adv Oto-Rhino-Laryng. 1988;41 :82-8.

26. O'Leary DP, Davis LL. High-frequency autorota­tional testing of the vestibulo-ocular reflex. Neurol Clin. 1990;8:297-312.