Dynamic Visual Acuity of Normal Subjects During VerticalOptotype and Head Motion

Joseph L. Demer^-f and Firooz AmjadiX

Purpose. To characterize the effect of passive vertical head motion on dynamic visual acuity ofyoung, normally sighted subjects wearing telescopic spectacles, and to relate this to the velocityof images on the retina.

Methods. Static visual acuity was measured without motion. Dynamic visual acuity was mea-sured during vertical, sinusoidal motion of either optotypes or of a servo-driven rotating chairin which subjects were seated. Dynamic visual acuity for head motion was measured unaided,as well as with 1.9X, 4X, and 6X telescopic spectacles. Vertical eye movements were recordedusing magnetic search coils.

Results. During optotype motion, acuity declined with increasing velocity to a minimum of~20 /200 at 100°/sec. Pursuit gain (eye velocity/optotype velocity) for moving optotypes waslow except for optotype velocities of 20°/sec of less. Dynamic visual acuity without telescopicspectacles was not sensitive to head motion. Static visual acuity improved with increasingtelescopic spectacle power, but dynamic visual acuity became progressively impaired by headmotion as telescopic spectacle power was increased. Compared with static visual acuity, headmotion with peak velocity of 40°/sec reduced acuity two-fold for 1.9X telescopic spectacles,fourfold for 4X telescopic spectacles, and eightfold for 6X telescopic spectacles. Visual vestibu-lo-ocular rellex gain with telescopic spectacles increased to values markedly above 1.0, but wasalways less than telescopic spectacle magnification. There was visual tolerance of slip velocitiesof 2°/sec or less, above which acuity declined in proportion to the 0.6 power of retinal slipvelocity. Above 2°/sec, retinal slip velocity accounted for 95% of the variance in dynamic visualacuity.

Conclusions. These results confirm that acuity is sensitive to retinal image motion in the verticaldirection, and extend this finding to indicate that sensitivity of acuity to vertical head motionduring wearing of telescopic spectacles is attributable to retinal image slip velocity. InvestOphthalmol Vis Sci 1993;34:1894-1906.

.Dynamic visual acuity (DVA) is that acuity obtainedduring relative motion of either optotypes or ob-server.1 It has long been known that motion reduces

From the * Jules Stein Eye Institute and ^Department oj Neurology, UCLA MedicalSchool, and %University of California at Los Angeles, Los Angeles, California.Supported by U.S. Public Health Service grants EY-08656 and NS-10940, andgrants from the Karl Kirchgessuer Foundation and Research to Prevent lilindness.JLD was a Research to Prevent lilindness William and Mary Creve InternationalResearch Scholar.Submitted for publication: December 31, 1991; accepted July 17, 1992.Proprietary Interest Category: N.Reprint Requests: Joseph L. Demer, Jules Stein Eye Institute, ComprehensiveDivision, 100 Stein Plaza, UCLA, Los Angeles. CA 90024-7002.

DVA as compared with static values obtained withoutrelative motion. Classical studies involving moving op-totypes have suggested that DVA decreases accordingto the cube of horizontal2-3 or vertical optotype ve-locity.4

A number of investigators have simultaneouslymeasured horizontal eye movements during DVA test-ing with moving optotypes, and related reduction inDVA to limitations on pursuit tracking.5-67 Murphymeasured horizontal eye movements and contrastthresholds for square wave gratings moving at veloci-

1894Investigative Ophthalmology & Visual Science, May 1993, Vol. 34, No. 6

Copyright © Association for Research in Vision and Ophthalmology

Vertical Dynamic Visual Acuity 1895

ties up to 7°/sec, concluding that visual sensitivity isnot reduced by pursuit itself, but is limited by velocityerrors in pursuit.8 Many previous studies of DVA havenot measured the tracking eye movements that serveto reduce the motion of images on the retina. In fact, itis common to employ DVA as an indirect indicator ofthe effectiveness of pursuit tracking,9 or of vestibular-evoked eye movements in stabilizing gaze during headrotation.10" Such studies highlight the importance ofDVA in relation to self-motion of the observer in theenvironment.12 Even when attempting to stand com-pletely at rest, humans exhibit significant involuntaryhead movements about the vertical axis (yaw), aboutthe interaural axis (pitch), and about the anteroposte-rior axis (roll).1314 Maximum head velocities duringrunning are up to 90°/sec, with predominant fre-quencies up to 2.7 Hz for yaw and 8.2 Hz for pitch.15

The mechanics of the head and neck are such that thefrequency of involuntary head movements in the pitchaxis tends to be substantially higher, and often athigher velocity, than head movements in the other tworotational axes.12131415 This implies that vertical self-rotation is likely to be significant for the visual system.

The major reflex producing compensatory move-ments during self-rotation is the vestibulo-ocular re-Ilex (VOR). During vision, other mechanisms augmentthe VOR, producing the visual vestibulo-ocular reflex(WOR). The gain of compensatory eye movements isdefined as the ratio of eye velocity to head velocity.Thus, the normal ideal gain of 1.0 implies that eyevelocity is equal and opposite head velocity, stabilizingthe retina in space. Very precise studies of active hori-zontal and vertical eye movements indicate that this isvery nearly the case, but that gain averages —1-2%less than the ideal value.16 Because of this residualerror, very rapid head rotations can overwhelm theWOR even in normal subjects.1017

Telescopic spectacles with magnifications of 2-8Xare commonly employed as visual aids for patients withlow vision, defined as significant but incomplete visualloss. Being affixed to the head, telescopic spectaclesmagnify the effects of head movement. With tele-scopic spectacles, the WOR gain required for retinalimage stabilization is no longer 1.0, but is equal to themagnification of the telescopes. If WOR gain is insuf-ficient, large retinal image slip velocities can occurwith only low or moderate velocities of head move-ment. Even physiologic head instability during stand-ing can reduce DVA with telescopic spectacles, as dem-onstrated for low vision patients wearing 4X telescopicspectacles.18

The magnification of involuntary head move-ments produced by telescopic spectacles appears tohave significant impact on functional use of these vi-sual aids spectacles by patients with low vision. A retro-spective study of telescopic spectacle use by low vision

patients identified a combination of two factors to be85% accurate in predicting successful use of telescopicspectacles by motivated low vision patients: sensitivityof visual acuity with telescopes to head motion, andinvoluntary head instability (while standing) in thepitch axis.12 This finding has been confirmed prospec-tively.19 Existing data on DVA with telescopic specta-cles are unsatisfactory in many respects, however. Allstudies employed passive head movements in the hori-zontal plane (yaw), despite the recognition that naturalhead movements are most challenging in the verticalplane (pitch). No previous study employing head move-ment with telescopic spectacles has related DVA toactual eye or retinal slip velocity during viewing, so therole of retinal slip has only been presumed to corre-spond to the literature obtained using moving opto-types.

Even the existing data relating retinal image mo-tion to DVA were obtained under conditions not rep-resentative of those encountered by low vision patientsusing telescopic spectacles. Reading employed Lan-doldt ring optotypes undergoing constant velocityramp motions of brief duration.35 The study ofWestheimer and McKee presented moving verniers orLandoldt ring optotypes so briefly that tracking eyemovements were impossible.20 The elegant studies ofBrown employed simultaneous eye movement record-ing during viewing of single Landoldt ring opto-types,2122 whereas Murphy employed variable contrastgratings repetitively alternating direction at constantspeed.8 In distinction to these laboratory conditions,visually impaired patients making daily use of tele-scopic spectacles must deal with retinal image motionproduced by substantially repetitive head movementsof varying velocities, produced by tremor, transmittedheartbeat, and environmental perturbations.14 Pa-tients may freely choose the duration of scrutiny. Al-phabetic letters, arranged in lines, are the commonoptotypes for clinical low vision testing and corre-spond to functional reading tasks, whereas Landoltring optotypes have practical drawbacks for clinicaluse.23 Existing data concerning DVA in normal sub-jects and low vision patients using telescopic spectaclesemployed rows of letter optotypes presented for ex-tended periods during repetitive, sinusoidal, horizon-tal head motion.1219'24 Since previous studies ofDVA with telescopic spectacles presented optotypesthrough the entire sinusoidal cycle of head motion, avariety of ocular motor strategies might have been em-ployed to avoid retinal image slip. Such strategiesmight include viewing during the low velocity portionsof the waveform,12 as is the case for foveation strate-gies of patients with congenital nystagmus.25 Practicaldesign considerations, particularly weight limitations,dictate that the field of view of telescopic spectaclesgenerally decreases as magnification is increased. This

1896 Investigative Ophthalmology & Visual Science, May 1993, Vol. 34, No. 6

implies that eye movement strategies might vary fromone telescopic spectacle power to another, whichmight also influence retinal image stability and DVAduring head motion. Thus, actual retinal image slipcannot be confidently estimated in the absence of eyeand head movement measurements under the viewingconditions employed.

The aim of the current study was thus to character-ize, in young, normally sighted subjects, the effect onDVA of imposed vertical head motion using telescopicspectacles of various optical powers. For comparisonwith previous studies of DVA with telescopic specta-cles during horizontal head motion, letter optotypesand sinusoidal waveforms were chosen.1219-24 A fur-ther goal was to relate DVA during head motion withtelescopic spectacles to actual retinal slip, the velocityerror of images on the retina, and to compare this todata from the literature. To make comparison withprevious studies easier, confirmatory measurementsof DVA were also made under comparable conditionsin the same subjects during vertical movement of op-totypes with the head stationary. These studies of DVAwith telescopic spectacles in young, normal subjectswere intended to provide normative data for contem-plated subsequent evaluations of the effects of aging,visual impairment, and neurologic disease.

MATERIALS AND METHODS

Thirteen young adult, paid volunteers gave written in-formed consent according to a protocol conformingto the tenets of the Declaration of Helsinki and ap-proved by the Human Subject Protection Committeeat the University of California at Los Angeles. All whoparticipated in the study underwent ophthalmologicexamination by one of the authors, and had visualacuities in each eye correctable to 20/20 or better us-ing standard clinical optotype presentations. Averageage of subjects was 30 ± 6 years (mean ± SD, range19-40); both men and women were represented.

The host computer for these experiments was aMacintosh II (Apple Computer, Cupertino, CA)equipped with 16-channel analog-to-digital converter,6-channel synchronous digital-to-analog converter(DAC), and direct memory access devices (National In-struments, Austin, TX). Custom software was writtenfor these experiments using the Lab View laboratorysoftware package (National Instruments).

Visual acuities were measured using single rows offive white Sloan letters optically projected at 98% con-trast and approximately 100 fL luminance against ablack background on a matte white screen measuring2.4 m2 and located 3 m from the subjects' eyes. Theroom was otherwise dark. The content of each row wasbalanced for equal difficulty,23 and the projector(Mast Development Company, Davenport, IA) was

controlled by the computer to rotate among three setsof letters for each optotype size to prevent memoriza-tion. Optotype size varied from line to line by 20%,representing a logarithmic change in the minimum an-gle resolvable (logMAR) of approximately 0.1 unit.Optotypes ranged in size from —0.7 (20/4 Snellen)through 1.7 logMAR (20/1000). For the optotypeslarger than 1.0 logMAR (20/200), fewer than five let-ters were presented per line, due to limitations in thesize of the transparencies in the projector. Subjects'responses were monitored using an intercom, andwere recorded on the audio channel of the data re-corder during trials when simultaneous eye movementrecording was performed. Threshold acuity was de-fined as the smallest optotype size for which the major-ity of letters were correctly identified in correct se-quence. If any errors occurred on a line, threshold wasspecified as 0.05 log unit greater than for all letterscorrect on that line.24

The optotype projector beam was deflected froma front surface mirror mounted on a high perfor-mance, temperature-compensated, position feedbackgalvanometer (General Scanning, Watertown, MA)that pivoted about the horizontal axis. The servodriver of the galvanometer was optimized for mini-mum deflection time, and was under computer con-trol by a synchronous DAC. This permitted the com-puter to control the vertical position of optotypes asprojected on the screen. The projection beam alsopassed through a large aperture that ordinarily did notinterfere with the beam, but which completely oc-cluded the beam when a maximum deflection com-mand was provided to the mirror galvanometer. Usingthis arrangement, optotypes could be moved sinusoi-dally in the vertical plane at chosen frequencies andamplitudes, but could also be extinguished (blanked)by the abrupt deflection of the beam beyond the aper-ture. In the current experiments employing sinusoidaloptotype and head motion, the optotypes were extin-guished for 50% of each cycle (50% dwell), centeredabout the minimum velocity portions of each half cycle(Fig. 1). This permitted viewing when optotypes wereat or near the straight ahead position, which is alsowhen optotype or head velocity was greatest.

For acuity measurement, subjects were seated in acustom-made swing rotator pivoting about the inter-aural axis. Their heads were firmly strapped to a head-rest on the rotator, and their legs and bodies werecomfortably secured to the rotator using harnessesand belts. The rotator was driven by a servomotor (108ft-lb, 4.5 kW, Inland Motors, Radford, VA) with a cus-tom-designed position feedback servo under com-puter control via a synchronous DAC. This permittedsinusoidal rotation of the subject's head and body atpeak velocities of up to 40°/sec, synchronized withoptotype motion. Pairs of magnetic field generator

Vertical Dynamic Visual Acuity 1897

Dwell

Position

Velocity

FIGURE 1. Diagram of timing of optotype blanking relative tomotion of optotypes or the rotating chair. The horizontalaxis represents time. Position and velocity traces apply toeither type of relative motion, which was sinusoidal despitethe more circular curves shown in the diagram. Dark hori-zontal bars labeled "Dwell" indicate intervals of optotypeblanking, and the squares at the bottom illustrate the screenand apparent optotype motion as these would be viewed by asubject during a moving optotype presentation.

coils 1 m in diameter were mounted on the swing rota-tor to permit the measurement of horizontal and verti-cal eye and head position using the magnetic searchcoil technique.26 Subjects wore a headband to whichwas attached a search coil for the measurement ofhead position relative to the swing rotator. The headsearch coil sensor was used only to verify that subjects'heads were always moving with the rotator headrest,since decoupling of the motion of the head from theheadrest in known to occur at high frequencies ofhead rotation. However, at 1.0 Hz, the frequency ofhead rotation chosen for these experiments, such de-coupling was never observed to occur.

Acuity during optotype motion was measuredmonocularly for the right eye without optical devicesfor emmetropic subjects, or with the subjects' custom-ary spectacles or contact lenses for subjects with ame-tropia. The total duration of presentation for each op-totype line was held constant at 16 sec, but the fre-quency and peak velocity of sinusoidal optotypemotion were systematically varied as indicated in Table1. In each case, viewing was limited to the upper veloc-ity half of each sinusoidal cycle, for a total blankingdwell of 180°.

Acuity during head motion was measured monocu-larly for the right eye under several viewing condi-tions. Viewing with only simple refractive correction(if necessary) was considered the IX condition. Mea-surements of acuity were also made using the follow-ing telescopic spectacles: 1.9X (~16.8° field); 4X(-10.3° field); and 6X (~7.5° field). The visual field

peripheral to the telescope elements was masked, as ina previous study.24 Field diameters were sufficient toinclude the optotypes throughout stimulus motions.

Vertical eye position was measured in selected ex-periments using the magnetic search coil technique,with the scleral search coil contact lens (Skalar Medi-cal, The Netherlands) applied to the left eye.27 Topicalproparacaine 0.5% was applied to the eye before in-sertion of the search coil contact lens, and the lens wasremoved within 30 minutes. Calibration was obtainedfor vertical saccades to targets located at 10° eccentric-ity from the primary position. True head position inspace was taken as the sum of swing rotator positionplus head position relative to the swing rotator; how-ever, there was generally no decoupled head move-ment and this correction proved to be an unnecessaryprecaution. Eye, head, and chair position data weredisplayed on a digital polygraph, with electronic dif-ferentiation for display of velocities. All data were alsodigitally recorded on magnetic tape for later editingand digital sampling at 400 Hz by the computer. Foranalysis of eye movements during optotype motion,peak slow phase eye velocity was determined by in-spection of the eye velocity tracings on the polygraphobtained during the reading of letters. For analysis ofeye and head movements during head motion, the au-dio channel of the recorded data tape was used todetermine the times when letter reading occurred;5-15 consecutive sinusoidal cycles were then digitallysampled from this period for automated analysis.

For trials during which eye movements were re-corded during passive head movements, data were au-tomatically analyzed by digital low pass filtering (0-42Hz), differentiation and removal of large quick phasesand saccades using velocity and duration criteria, andcross spectral analysis of eye versus head velocity todetermine the phase shift of the eye velocity response.After correction for phase, a linear regression of eyevelocity against head velocity was obtained, and datapoints found to be statistical outliers from this fit werealso excluded as quick phases. This generally removedapproximately 20% of points. The remaining slowphase data points were fitted cycle-by-cycle by leastsquares to a sinusoidal equation. Head velocity, eyevelocity, and gain (eye velocity/head velocity) were

TABLE l. Characteristics of Optotype MotionPeak Velocity (deg/sec)

7.51530456075

100

Frequency (Hz)

0.0.

25.5

1.012.2.3.

.5

.0

.533

1898 Investigative Ophthalmology & Visual Science, May 1993, Vol. 34, No. 6

computed for each cycle, and outliers were automati-cally excluded using a statistical criterion previouslydescribed to be effective for the removal of low-gainartifacts.28

During optotype motion, retinal image slip veloc-ity was considered to be the difference between opto-type velocity and eye velocity. During head motion,retinal image slip velocity e was computed from eyevelocity E and head velocity H using the followingequation:

e = E - MH, (1)

where M represents telescope magnification.

RESULTS

Moving Optotypes

DVA was measured in 13 subjects during controlledoptotype motion as specified in Table 1. These dataare plotted as a function of peak optotype velocity inFigure 2. Although there was some interindividualvariation, all subjects exhibited the same pattern ofdependence of acuity on optotype velocity. At zerooptotype velocity, representing static visual acuity(SVA), mean acuity was approximately 0 logMAR (20/20), and acuity declined monotonically with increasingpeak optotype velocity. At a peak optotype velocity of100°/sec, DVA declined by a full log unit (20/200)relative to SVA. At this optotype velocity, the range ofindividual acuities was 0.65-1.15 logMAR (~20/90-20/280). Individual subjects thus vary in their suscepti-bilities of DVA to optotype motion.

Moving HeadThe same 13 subjects also underwent DVA measure-ments during passive whole-body, sinusoidal rotationin the vertical plane at 1.0 Hz. Peak velocity was variedfrom 0°/sec to 40°/sec, and optotypes were blankedduring half of each sinusoidal cycle at intervals cen-tered around the velocity zero crossings. The effect ofthis head motion on DVA with telescopic spectacles isillustrated in Figure 3. During unmagnified vision,there was little dependence of acuity on peak sinusoi-dal head velocity. Although statistically significant (P< 0.05, two-tailed t test), the decline in acuity relativeto SVA associated with head motion at 10°/sec and40°/sec was only about one half line; the decline wasnot significant at 20°/sec.

In contrast to the case of unmagnified vision, pas-sive head motion with telescopic spectacles was asso-ciated with marked reductions in acuity in all subjects,although there was again some individual variation(Fig. 3). The 1.9X, 4X, and 6X telescopic spectacleswere associated with progressive increases in SVA, re-spectively, although the increases were less than pro-portionate to optical magnification. As indicated inFigure 3, as telescopic magnification increased, DVAbecame increasingly sensitive to head velocity. In allcases, mean DVA with telescopic spectacles duringhead motion at 20°/sec was less than with unmagni-fied vision at the same velocity. The following dataobtained at a peak head velocity of 40°/sec are illus-trative. Using 1.9X telescopic spectacles, mean acuitydeclined by more than two-fold relative to SVA, with arange of -0.05-0.65 logMAR (-20/18-20/90). Us-ing 4X telescopic spectacles, mean acuity declined by

1.2

O)

o

I75

1.0"

0.8

0.2

0.0

-0.2

Mean ± SEM

20/320

20/200

20/50

•20/32

20/20

20/1220 40 60 80

Optotype Velocity - Deg/Sec100

FIGURE 2. Dynamic visual acuity of 13 subjects during sinusoidal optotype motion. Abscissaindicates peak sinusoidal optotype velocity, under conditions specified in Table 1.

Vertical Dynamic Visual Acuity 1899

-0.6-5 5 10 15 20 25 30 35

Head Velocity at 1.0 Hz - deg/sec

20/80

20/50

20/32

20/20

• 20/12

20/8

20/5

FIGURE 3. Dynamic visual acuity of 13 subjects during sinusoidal, head rotation at 1 Hz,viewing with and without telescopic spectacles of indicated powers. 1 X, no telescopes.

five-fold relative to SVA, with a range of 0.15-0.55logMAR (~20/30-20/70). Using 6X telescopic spec-tacles, mean acuity declined by more than seven foldrelative to SVA, with a range of 0.35-0.85 logMAR(~ 20/45-20/140). These ranges indicate that individ-ual subjects also vary considerably in their susceptibili-ties of DVA to head motion while wearing telescopicspectacles.

Eye Movements

Compensatory eye movements during DVA were mea-sured in five of the subjects. Optotypes were extin-guished except around the straight-ahead position, asbefore. During measurements, subjects read opto-types approximately two lines larger than their individ-ually measured thresholds. The magnetic search coilcontact lens was applied to each subject's occluded lefteye, during viewing by the right eye. For moving opto-types, peak velocities of 7.5, 15, 45, and 75°/sec wereemployed. Typical eye movements during reading ofmoving optotypes are illustrated in Fig. 4 (upper). Sur-prisingly, despite optotype blanking except aroundthe straight-ahead position, slow-phase velocity was es-sentially sinusoidal, although only part of the sinusoi-dal optotype motion was visible. Catch-up saccades oc-casionally occurred (Fig. 4, upper), and eye positionclosely approximated that of the optotypes. The peak-to-peak position amplitude of the eye tracking re-sponses was typically equal to the total displacement ofthe optotypes during their presentation, 4.8°. Thissuggests that foveal position error was minimal. Theocular motor strategy of all subjects was simply to at-

tempt smooth sinusoidal tracking of optotype motionwhen optotypes were visible, and to continue sinusoi-dal tracking when they were not, with catch-up sac-cades as needed to correct for position errors. How-ever, peak smooth eye velocity approximated the stim-ulus only at the 7.5°/sec; at higher optotype velocities,eye velocity fell behind (Fig. 5). Peak eye velocity neverexceeded 12°/sec at any optotype velocity.

During passive head motion, slow phase eye veloc-ity was also sinusoidal, but much higher velocities wereobserved than during optotype motion (Fig. 4, middeand lower). The relationship between peak eye velocityand peak head velocity and the dependence upon tele-scopic spectacle power, are illustrated in Fig. 6. Athead velocities of 10°/sec and 20°/sec, increasing tele-scopic spectacle power was associated with increasingeye velocity, but at 40°/sec head velocity, eye velocitybegan to saturate at a maximum of 80°/sec. The rela-tionship between eye velocity and head velocity is re-plotted in terms of gain in Fig. 7. This illustrates thatgain approached but did not achieve optical magnifi-cation at a head velocity of 10°/sec, and declined athigher head velocities for 4X and 6X telescopic spec-tacles. For 1.9X telescopic spectacles, gain increasedas head velocity was increased from 10°/sec to 20°/sec, and did not change further as head velocity wasincreased to 40°/sec.

Retinal Slip Velocity

Retinal slip velocity was computed from the data forboth optotype motion and head motion, and was plot-ted in logarithmic form against DVA measured in the

1900 Investigative Ophthalmology 8c Visual Science, May 1993, Vol. 34, No. 6

EyePosition

EyeVelocity

Moving Optotypes - 1.5 Hz, 45°/sec

EyePosition

EyeVelocity \J

Moving Chair, 1.0 Hz, 40°/sec, 1.9X

EyePosition

EyeVelocity

Moving Chair, 1.0 Hz, 40°/sec, 4X

FIGURE 4. Eye movements recorded using the magnetic search coil affixed to the nonviewingeye of one subject during dynamic visual acuity testing. Each pair of tracings consists ofvertical eye position and vertical eye velocity obtained by electronic differentiation. Upwardeye movements are signified by upward deflections of the tracings. Upper tracings, movingoptotypes. Note prominent catch-up saccades. Middle tracings, moving head, with ] .9X tele-scopic spectacles. Lower tracings, moving head, with 4X telescopic spectacles.

same subjects (Fig. 8). The use of multiple telescopicspectacle powers permitted examination of a broadrange of retinal slip velocities. The plots suggest thatfor 2°/sec or less the visual system tolerates retinalimage slip without appreciable decrement in DVA, butthat DVA declines markedly with slip above this. Be-cause the measure of acuity is logarithmic, the effectof telescopic spectacles is to shift the curves for headmotion downward according to telescope magnifica-tion. Because telescopic spectacles also magnify headmovements that are not fully compensated by eyemovements, increasing telescope power also shifts thecurves to the right. Above 2°/sec, the curves for eachviewing condition appear to be linear and approxi-mately parallel to one another. Linear regression for

each curve yielded coefficients of determination> 0.91, with slopes ranging from 0.53 for optotypemotion to 0.92 for 6X telescopic spectacles.

If acuity values are normalized for each telescopicspectacle power by the addition of the logarithmicvalue of magnification, DVA values for all telescopescan be pooled and plotted together against retinal slipvelocity (Fig. 9). For retinal slip velocities greater than2°/sec, all DVA values tightly clustered about a linearrelationship with the logarithm of slip velocity (R2

= 0.954). This was true regardless of viewing condi-tion and telescopic spectacle power, and strongly sug-gests that retinal slip velocity is responsible for thereduction in DVA. In contrast, for retinal slip veloci-ties at or below 2°/sec, linear regression suggested no

Vertical Dynamic Visual Acuity

20

1901

Optotype Velocity - deg/secFIGURE 5. Mean eye velocity of five subjects during dynamic visual acuity task near thresholdfor optotypes motion under conditions specified in Table 1. Not all subjects contributed dataat every point.

significant relationship between DVA and retinal slip(Fig. 9, R2 = 0.003). This suggests that acuity is notsensitive to retinal slip velocity of 2°/sec or less.

For retinal slip velocities greater than 2°/sec, theslope of the line relating DVA to the logarithm of reti-

nal slip velocity was 0.64 (Fig. 9). Thus, acuity A inlogMAR is related to retinal slip velocity s by the follow-ing equation:

A = K + 0.64 log(s), (2)

100

o0)

•

OOQ)

0)

LU

4010 20 30

Head Velocity at 1.0 Hz - deg/sec50

FIGURE 6. Mean vertical eye velocity of five subjects during dynamic visual acuity task nearthreshold for sinusoidal head motion at 1.0 Hz with and without telescopic spectacles ofindicated powers. IX, unmagnified vision.

1902 Investigative Ophthalmology & Visual Science, May 1993, Vol. 34, No. 6

4.0

cCOOt£O

>

3.5-

3.0-

2.5-

2.0-

1.5-

1.0"

0.5-

0.0

^

—*— ix* 1.9X

••"••• 4X

- • - 6X

Mean ± SEM

10 20 30 40

Head Velocity at 1.0 Hz - deg/sec50

FIGURE 7. Mean vertical VVOR gain of five subjects during dynamic visual acuity task nearthreshold for sinusoidal head motion at 1.0 Hz with and without telescopic spectacles ofindicated powers. 1 X, unmagnified vision.

where K is a constant. Taking exponentials, acuity a inMAR is related to retinal slip velocity s by

a = Ks0M. (3)

This implies that the minimum angle resolvable isapproximately proportional to the 0.64 power of reti-

nal slip velocity, regardless of whether the slip velocityis due to optotype or head motion, and regardless oftelescopic spectacle power used, if any.

DISCUSSIONAs shown in Figure 2, this investigation confirms thefinding of Miller that vertical motion of optotypes re-

1.0-

0.8-_o

i 0.6-I

a< 0.4-

~ 0.2 i

o

1c>»

Q -0.2 i

-0.4

- ° — Moving Optotypes• * • - 1X

-*••" 1.9X Telescopes*•-• 4X Telescopes•*"" 6X Telescopes

Mean ± SEM

1 10 100Retinal Slip Velocity - deg/sec

20/200

20/125

20/80

20/50

• 20/32

20/20

20/12

100020/8

FIGURE 8. Relationship between dynamic visual acuity and retinal image slip velocity in thesame five subjects, measured for sinusoidal vertical motion of optotypes, or of the head,during normal vision (IX), or of telescopic spectacles of the powers indicated in the key. Insome cases error limits were smaller than the size of symbols plotted.

Vertical Dynamic Visual Acuity 1903

1.4

DC<

cQ

1.2-

O 1.0

~ 0.8 i3U

— 0.6-(Q3(Aj> 0.4-O

I 0.2

0.0

-0.2

Normalized Acuity

Slope = 0.01

R2 r 0.003i

Slope = 0.64

R2= 0.954

Mean ± SEM

10 100

Retinal Slip Velocity - deg/sec

20/500

20/320

20/200

•20/125

20/80

20/50

^20/32

20/20

20/121000

FIGURE 9. The data of Fig. 8 have been normalized by addition of the logarithmic value ofmagnification of each telescopic spectacle used, equivalent to multiplying the thresholdacuity values by magnification. This demonstrates the effect of retinal slip on acuity, indepen-dent of the effect of telescope magnification on image size. For retinal slip velocities up to2°/sec (dotted line), DVA was independent of retinal slip. For higher slip velocities, there wasa close linear correlation (solid line) relating decreasing DVA (increasing logMAR) with thelogarithm of retinal slip velocity.

duces acuity progressively with increasing optotype ve-locity, although because Miller employed Landolt ringoptotypes, the somewhat greater reduction in acuityfound here is probably attributable to differences inoptotype presentation.4 The behavior reflected in Fig-ure 2 is also typical of many studies of DVA for hori-zontally moving optotypes.6>7>21>29f30'31 Thus the resultsof the sinusoidally moving optotype experiment sug-gest that the Sloan optotypes employed here produceDVA data comparable to that obtained by previousinvestigators who used different optotypes and typesof motion.

A number of studies have investigated eye move-ments during DVA tasks for horizontal optotype mo-tion. The studies of Reading,5 Barmack,6 and Brown7

employed infrared reflectance or electro-oculographyto monitor horizontal eye movement during singleramp movements of optotypes. Murphy used an opti-cal lever technique to measure tracking of gratingsmoving horizontally with a repetitive waveform.8 Allfour authors observed tracking eye movements thatfailed to match optotype motion except at low opto-type velocities. Barmack also demonstrated that thereis a predictive component to the tracking.6 Althoughthe present experiments employed a repetitive sinusoi-dal motion, the observation of poor pursuit tracking atthe higher velocities (Figs. 4 and 5) is consistent with

these previous investigations. The contribution of apredictive mechanism is also suggested by the persis-tently sinusoidal waveform of eye movements ob-served here (Fig. 4), despite blanking of optotypesduring half of each cycle. There is. good evidence inhumans of the continuation of horizontal predictivepursuit for disappearing targets,32>33>34>35>36 and pre-diction has also been reported to contribute to verticalvisual-vestibular interaction.37 Aside from the use ofprediction to continue smooth sinusoidal motion dur-ing optotype blanking, no other ocular motor strategywas evident during viewing of moving optotypes.

The current study also examined the effect of ver-tical rotation of the subject on DVA. Sinusoidal mo-tion was chosen, both for mechanical feasibility andfor comparison with previous data.24 The frequency of1.0 Hz was chosen both for comparison with previousdata and because it is in the range of physiologic headmovements.131415 Under conditions when the VVORwas effective in fully compensating for head rotation,acuity was insensitive to head motion (Fig. 3). Withtelescopic spectacles, W O R gain did not match spec-tacle magnification and acuity became increasingly im-paired by head motion as telescopic spectacle powerincreased from 1.9X to 6X (Fig. 3). This effect is quitesimilar to that of horizontal head motion under similarconditions.24 Even at 10°/sec head velocity with 1.9X

1904 Investigative Ophthalmology & Visual Science, May 1993, Vol. 34, No. 6

telescopic spectacles, WOR gain was not sufficient toprevent retinal slip, although gain with the same tele-scopes was greater at higher head velocities (Fig. 7),and much higher eye velocities were achieved underother testing conditions (Fig. 6). Thus, although theocular motor system was capable of achieving WORgain sufficient to prevent retinal slip at low head veloc-ity, it did not do so, supporting the supposition ofSkavenski et al that some small amount of retinalimage motion is preferred for optimal visual process-ing.38 When telescopic spectacles are worn, however,the retinal image motion at higher velocities creates anadverse effect on DVA.

It is interesting to note that at a peak head velocityof 20°/sec or greater, DVA with telescopic spectaclesof any power was less than without any telescopic spec-tacles. Such head velocities, which are well within thephysiologic range during ambulation, completely de-feat the visual advantage of telescopic spectacles.1415

This explains the general inability of visually impairedpatients to make use of telescopic spectacle low visionaids while walking, and the common strategy of period-ically stopping to "spot" objects of interest throughthe telescopes.1219 Presumably, those with poorer vi-sual-vestibular interaction, and consequently lowerWOR gain, would experience greater reduction inDVA. Since significant involuntary head movementsall three rotational axes occur even during attempts tostand motionless,14 inadequate WOR gain explainsthe observed reduction in acuity with 4X telescopicspectacles during free standing, as compared with sta-bilization of the head against a fixed object.18 This alsoagrees with reports that some low vision patients withparticularly severe head tremor can make use of tele-scopic spectacles only when bracing their headsagainst fixed objects such as lamp posts or walls.1219

For such patients with severe head tremor, an alterna-tive might be the use of handheld telescopic visualaids, which do not magnify head movements becausethey are not mechanically coupled to the head.

Simultaneous measurements were made of eyemovements so the current data permit estimation ofthe contribution of retinal image slip velocity, velocityerror, to the reduction in DVA during vertical motion.However, the problem is complicated by the covaria-tion in these studies of three potential factors in thereduction of acuity. Retinal position and velocityerrors are necessarily correlated, to a degree depend-ing on the duration of optotype presentation, whichitself is a significant determinant of acuity.7 Consistentwith realistic conditions of telescopic spectacle use, inthe current study relative motion was repetitive andoptotype presentation was of such long duration thatviewing time may be safely assumed not to limit acuity.Conditions were always chosen so that relative motion

crossed the straight-ahead position, the assumed fo-veal direction, and saccades were permitted to re-center the presentation. Thus, it may be assumed thatretinal position error was zero for at least portions ofall stimulus presentations. During moving optotypepresentations, the maximum position displacement ofthe optotypes was 2.4°, further limiting the potentialposition error even in the absence of tracking eyemovements. The effects of retinal position error havebeen estimated. Using optotypes moving in circularpaths of different radii, Ludvigh estimated that theposition error of images relative to the fovea contrib-utes relatively little to the reduction in acuity due toimage motion.39 Ludvigh's data suggested that acuityof better than 20/30 is present with a position error of2°. The relative contribution of retinal position errorhas also been studied by Brown using brief presenta-tions of horizontally moving optotypes at various ec-centricities from fixation.22 Brown found that dynamicvisual acuity at 2.5° eccentricity varies with image ve-locity in a manner quite similar to that at fixation, al-though static acuity decreases by roughly half.

The current finding of an effect on DVA of retinalslip only more than 2°/sec (Fig. 9) confirms, in a largergroup of subjects, the data of Westheimer and McK.ee,who employed presentations of horizontally and verti-cally moving optotypes so brief that tracking was im-possible, and found no reduction in DVA for slip veloc-ities of less than 2.5-3.0°/sec.20 It also agrees with thefindings of Murphy, who employed horizontally mov-ing gratings and found a marked increase in thresholdcontrast for retinal image speeds exceeding 2°/sec.8

Above this range of retinal slip velocity, acuity in thepresent study declined with increasing slip in similarfashion regardless of viewing condition. The presentstudy thus demonstrates the applicability of prior dataon retinal image velocity and acuity to the special butclinically important situation of vertical self-rotationduring the wearing of highly magnifying telescopicspectacle visual aids. The source of retinal image mo-tion—optotype movement, eye movement, or headmovement—is thus unimportant to the effect ofimage motion on DVA. Although the present findingsregarding DVA during optotype motion merely con-firm and extend the cited studies to a higher range ofretinal slip, our quantitative findings about the effectof telescopic spectacles during head motion are new.Telescopic spectacles merely shifted the parallelcurves relating DVA to retinal slip as dictated by opti-cal magnification of the images and of the imposedhead movements (Fig. 8). This finding was not entirelypredictable, because a number of other optical char-acteristics of telescopic spectacles change concur-rently with magnification, including field size andimage brightness. The present data suggest that, in

Vertical Dynamic Visual Acuity 1905

young normal subjects, the major effects of head mo-tion on DVA with telescopic spectacles can be ac-counted for by consideration of W O R eye move-ments and magnification only. Other optical factorstherefore play minor roles at best.

The present data, although demonstrating a con-sistent relationship between vertical motion and DVAin a group of normally sighted, young adult subjects,also demonstrate individual variability in both DVAand concurrent eye movements. These individual dif-ferences have potential significance, because DVA hasbeen suggested to be an important determinant offunction for many important tasks. Among the nor-mally sighted elderly, DVA has been found to be thefactor most significantly associated with frequency ofautomobile driving,40 and is significantly better amongstudent athletes than among their nonathletic peers.41

It has been well established that DVA and spontaneousvertical head motion are predictive of rehabilitativesuccess in patients who attempt use of telescopic spec-tacles as low vision aids.19 Although the effect of ageon vertical W O R gain is unknown, horizontal W O Rgain with telescopic spectacles is known to decline sig-nificantly with age.42 The present study, linking visualtracking and W O R eye movements to DVA via themechanism of retinal image slip, suggests that verticaleye movement reflex performance may have consider-able functional implications for everyday life. Thissupposition could be tested by further studies of DVAand retinal slip in populations, such as the elderly andvisually impaired, in whom greater variations in perfor-mance might be expected.

Key Words

dynamic visual acuity, low vision, telescopic spectacles, ves-tibulo-ocular reflex, visual-vestibular interaction.

Acknowledgments

The authors thank Laura A. Hovis for assisting with recruit-ment and testing of subjects, Douglas Wong for softwaresupport, James Li for production of optotype transparen-cies, and Hongxing Dai for construction of optical appa-ratus.

References

1. Miller JW, Ludvigh EJ. The effect of relative motionon visual acuity. Surv Ophthalmol. 1 962; 7:83-116.

2. Ludvigh E, Miller JW. Study of visual acuity duringthe ocular pursuit of moving test objects. I. Introduc-tion. J Opt Sac Am. 1958;48:799-802.

3. Reading VM. Visual resolution as measured by dy-namic and static tests. Pjlugers Arch. 1972;333:17-26.

4. Miller JW. Study of visual acuity during the ocularpursuit of moving test objects. II. Effects of directionof movement, relative movement, and illumination. JOptSocAm. 1958;48:803-808.

5. Reading VM. Analysis of eye movement responses anddynamic visual acuity. Pjlugers Arch. 1972; 333:27-34.

6. Barmack NH. Dynamic visual acuity as an index of eyemovement control. Vision Res. 1970; 10:1377-1391.

7. Brown B. Dynamic visual acuity, eye movements andperipheral acuity for moving targets. Vision Res.1972; 12:305-321.

8. Murphy BJ. Pattern thresholds for moving and sta-tionary gratings during smooth eye movement. VisionRes. 1978; 18:531-536.

9. Borland RC, Nicholson AN. Visual motor co-ordina-tion and dynamic visual acuity. Br J Clin Pharmacol.1984;18(suppl 1):69S-72S.

10. Benson AJ, Barnes GR. Vision during angular oscilla-tion: The dynamic interaction of visual and vestibularmechanisms. Aviation Space Environ Med. 1978;49:340-345.

11. Longridge NS, Mallinson Al. A discussion of the dy-namic illegible E test: A new method of screening foraminoglyccoside vestibulotoxicity. Otolaryngol HeadNeckSurg. J984;92:671-677.

12. Demer JL, Porter FI, Goldberg J, Jenkins HA,Schmidt K, and Ulrich I. Predictors of functional suc-cess in telescopic spectacle use by low vision patients.Invest Ophthalmol Vis Sd. 1989; 30:1652-1665.

13. Demer JL, Porter FI, Goldberg J, Schmidt K, Had-dock RD. Effect of telescopic spectacles on head stabil-ity in three axes. In: Topical. Meeting on NoninvasiveAssessment of the Visual, System, 1989 Technical Digest.Series, vol. 7. Washington, DC: Optical Society ofAmerica; 1989:144-147.

14. Demer JL, Goldberg J, Porter FI. Effect of telescopicspectacles on bead stability in normal and low vision. /Vestibular Res: Equilibrium and. Orientation.. 1991;1:109-122.

15. Grossman GE, Leigh RJ, Abel LA, Lanska DJ, Thur-ston SE. Frequency and velocity of rotational beadperturbations during locomotion. Exp Brain Res.1988;70:470-476.

16. Ferman L, Collewijn H, Jansen TC, Van den Berg AV.Human gaze stability in the horizontal, vertical, andtorsional direction during voluntary head movements,evaluated with a three-dimensional scleral inductioncoil technique. Vision Res. 1987;27:81 1-828.

17. Guedry RE, LentzJM, Jell RM. Visual-vestibular in-teractions: 1. influence of peripheral vision on sup-pression of the vestibulo-ocular reflex and visualacuity. Avial Space Environ Med. 1979;50:205-21 1.

18. Porter FI, Goldberg J, Denier JL. Sensitivity of visualacuity to spontaneous head motion while wearing tele-scopic spectacle low vision aids. ARVO Abstracts. In-vest Ophthalmol. Vis Sd. 1989;30(suppl):399.

19. Demer JL, Porter FI, Goldberg J, Schmidt K. Valida-tion of physiologic predictors of successful telescopicspectacle use in low vision. Invest Ophthalmol Vis Sci.1991;32:2826-2834.

20. Westheimer G, McKee SP. Visual acuity in the pres-ence of retinal image motion. J Opt Soc Am.1975;65:847-850.

1906 Investigative Ophthalmology & Visual Science, May 1993, Vol. 34, No. 6

21. Brown B. The effect of target contrast variation ondynamic visual acuity and eye movements. Vision Res.1972;12:1213-1224.

22. Brown B. Resolution thresholds for moving targets atthe fovea and in the peripheral retina. Vision Res.1972; 12:293-304.

23. Strong G, Woo GC. A distance visual acuity chart in-corporating some new design features. Arch Ophthal-mol. 1985;103:44-46.

24. Demer JL, Porter FI, Goldberg J, Jenkins HA,Schmidt K. Dynamic visual acuity with telescopic spec-tacles: improvement with adaptation. Invest Ophthal-mol VisSci. 1988; 29:1184-1189.

25. Dell'Osso LF, Daroff RB. Congenital nystagmus wave-forms and foveation strategy. Doc Ophthal. 1975; 39:155-182.

26. Fuchs AF, Robinson DA. A method for measuring hor-izontal and vertical eye movements chronically in themonkey.y Appl Physiol. 1966;21:1068-1070.

27. Collewijn H, Van der Mark F, Jansen TC. Precise re-cording of human eye movements. Vision Res.1975; 15:447-450.

28. Demer JL, Porter FI, Goldberg J, Jenkins HA,Schmidt K. Adaptation to telescopic spectacles: Vestib-ulo-ocular reflex plasticity. Invest Ophthalmol Vis Sci.1989;30:159-l70.

29. Fergenson P, Suzansky JW. An investigation of dy-namic and static visual acuity. Perception. 1973; 2:343-356.

30. Long GM, Garvey PM. The effects of target borderson dynamic visual acuity: Practical and theoretical im-plications. Perception. 1988; 17:745-751.

31. Long GM, Crambert RF. The nature and basis of age-related changes in dynamic visual acuity. Psychol Ag-ing. 1990; 5:138-143.

32. Becker W, Fuchs AF. Prediction in the oculomotorsystem: Smooth pursuit during transient disappear-ance of a visual target. Exp Brain Res. 1985; 57:562-575.

33. Mitrani L, Dimitrov G. Pursuit eye movements of adisappearing moving target. Vision Res. 1978; 18:537-539.

34. Muratore R, Zee DS. Pursuit after-nystagmus. VisionRes. 1979; 19:1057-1059.

35. Whittaker SG, Eaholtz G. Learning patterns of eyemotion for foveal pursuit. Invest Ophthalmol Vis Sci.1982;23:393-397.

36. Kowler E, Steinman RM. The effect of expectationson slow oculomotor control—I. periodic target steps.Vision Res. 1979; 19:619-632.

37. Demer JL. Mechanisms of human vertical visual-vesti-bular interaction. J Neurophysiol. 1992;68:2128-2146.

38. Skavenski AA, Hansen RM, Steinman RM, WintersonBJ. Quality of retinal image stabilization during smallnatural and artificial body rotations in man. VisionRes. 1979; 19:675-683.

39. Ludvigh E. Visual acuity while one is viewing a movingobject. Arch Ophthalmol. 1949;42:14-22.

40. Retchin SM, Cox J, Fox M, Irwin L. Performance-based measurements among elderly drivers and non-drivers.y Am Geriatr Soc. 1988;36:813-819.

41. Rouse MW, DeLand P, Christian R, Hawley J. A com-parison study of dynamic visual acuity between ath-letes and nonathletes.y Am Optom Assn. 1988; 59:946-950.

42. Demer JL, Goldberg J, Porter FI, Jenkins HA,Schmidt K. Visual-vestibular interaction with tele-scopic spectacles. J Vestibular Res. 1991; 1:263-277.