Suppression of sensitivity to surround displacement during vergence eye movements

EXPERIMENTAL

NEUROLOGY

105,300-305 (1989)

Suppression of Sensitivity to Surround Displacement during Vergence Eye Movements GEORGE K. HUNG,*

TIJUN WANG,*

KENNETH

J. CIUFFREDA,~

AND JOHN L. SEMMLOW*

*Rutgers University, Department of Biomedical Engineering, P.O. Box 909, Piscataway, New Jersey 08854; and tSchnurmacher Institute for Vision Research, SUNY/State College of Optomety, Department of Vision Sciences, 100 East 24th Street, New York City, New York 10010

a change in the retinal image is followed immediately by a stable image. Thus, the stable image dominates and masks the smeared image (17). Centrally, any residual signal from the smeared image appears to be inhibited, perhaps because it provides little useful information or computational overload related to eye movement control prevents its processing (37). Active central inhibition may also occur, coordinated by a “corollary discharge” signal (10) related to the neural command for eye or lid movement. As a result of both peripheral and central suppressions (7,31), the intermittent blurring and darkening that may otherwise occur are replaced instead by a stable and continuous view of the world. Experiments have shown that saccadic suppression of sensitivity to light (13, 14,20,21,23,28,36,37) and target displacement (2, 4, 6, 8, 22, 26) begins about 40 ms before the saccade, lasts until about 80 ms after saccadic onset, and has an amplitude of about 0.5 log units. A detailed review of the literature on saccadic suppression may be found in Matin (18). Suppression of sensitivity to a light pulse during a blink has also been demonstrated. Blink suppression was found to be similar in amplitude to that of saccadic suppression (0.5 log units), but lasting longer (about 400 ms) (27,29,30). Vergence suppression to light decrement has recently been reported (15,16,32). In these experiments, the subject was positioned near the center of a sphere 60 cm in diameter which acted as a Ganzfeld stimulus. Light decrements were delivered during vergence movements by fast rise-time lamps operated from a tachistoscopic power supply. Vergence stimulus angles were different for the three subjects used and varied from 6 to 9”. Eye movements were recorded using a low resolution electrooculgraphy (EOG) technique. The results showed that for brief light decrements, vergence suppression was similar to blink suppression, beginning about 250 ms before and ending 200 ms after the start of the eye movement. The maximum increase in threshold was about 0.5 log units. However, vergence suppression to image displacement has not been investigated. We describe below a series of experiments that quantify the suppression of sensitivity

Suppression was investigated psychophysically in three human observers by measuring their loss of sensitivity to brief (20 ms) simultaneous vertical displacement (up to 0.5”) of horizontal lines during 4” convergence eye movements. A two-alternative forced choice procedure was used in which the stimulus was presented either in the first or second portion of a trial. The amplitude of the displacement pulse, the time of the pulse relative to convergence onset, and the portion of a trial in which the stimulus was presented were randomized. The results showed that suppression began about 200 ms before, and continued until 350 ms after, convergence onset with maximum loss occurring at 25 to 125 ms after convergence onset. The maximum sensitivity loss was about 0.25 to 0.30 log units. Since peripheral factors were minimized by the use of a brief stimulus presentation and an eccentrically placed surround, the suppression found was primarily attributed to central neural mechanisms. Finally, the suppression of sensitivity to pulse displacement during the initial phase of the vergence movement is consistent with a recently developed dual-mode model of the vergence system, in which the initial transient portion of a step response is preprogrammed whereas the final sustained portion is maintained by continuous feedback control. 0 1989

Academic

Press,

Inc.

INTRODUCTION We live under the illusion that the world around us is seen continuously. The perception of continuity and the accompanying sense of stability persists despite disruptive changes in the retinal image that occur during highvelocity saccadic eye movements as well as during much slower movements such as blinks and vergence. However, under special laboratory lighting conditions, one may see smear of the retinal image during a saccade (5). Also, with greater attention under ordinary lighting conditions, one may notice the slight darkening of the visual scene during a blink and the shift in the surround during vergence (l&16). Two mechanisms appear to aid in the illusion of a continuous and stable percept. Peripherally, 0014-4886/89 $3.00 Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

300

SUPPRESSION

OF

SENSITIVITY

LE

FIG. 1. Overall layout of binocular eye movement stimulus system. The major components are T, binocular disparity target; I, displacement stimulus; L, alignment laser light source; Pl and P2, iris diaphragms; Bl-B3, beam splitters; Ll-L4, double convex, achromatic lenses; Ml and M2, front surface mirrors; M3 and M4, translating and rotating front surface mirrors; PC, photocells for infrared eye movement recording device; and LE and RE, the left and right eyes of the subject, respectively.

to surround displacement during vergence eye movements (12). The generality of the finding of suppression during vergence eye movements supports a recently developed dual-mode model of the vergence system (11). In this model, a fast preprogrammed component drives the transient part of the response to a step disparity stimulus, and then a slow closed-loop component maintains fusion of the retinal images. Suppression of sensitivity to both surround displacement and luminance change would be consistent with preprogrammed control of the transient part of the vergence movement, since such control only requires information about the initial disparity.

TO

SURROUND

301

DISPLACEMENT

an EG&G Electra-Optics Model 450 Photometer) generated on a low-persistence (P31) oscilloscope screen which appeared against a dark background (5.43 X 10m3 lx). The aperture, Pl, was reduced to
METHOD

The overall layout of the binocular eye movement stimulus system (Fig. 1) is a modification of that used by Semmlow and Venkiteswaran (25). The target (T) was placed 114 cm from aperture Pl, which was optically conjugate to the entrance pupils of the eyes (LE and RE). The optical components were also arranged so that lateral displacement of the target at (T) produced oppositely directed image displacements in the two eyes to give disparity stimulation. The disparity target was a vertical line (1” high and 0.1” wide; 6.2 X 10e2 lx illuminance at the entrance pupil of the eyes as measured with

,I’

___-.---__ -.. I’

f::’

LEFT

FIG. 2.

,’

,,,’ .-

__----.__. .-.~ ~\

RIGHT

Subject’s left and right eye views of the disparity target (short vertical line segments) and surround displacement stimulus (horizontal lines). The circular field diameter was 16.7”.

302

HUNG

eyes were dynamically measured using an infrared eye movement monitor. This method provides f10” linearity and a resolution of approximately 10 min of arc (35). Three subjects (G.H., J.S., and L.S.) participated in the experiments. G.H. (age 40) and J.S. (age 46) were experienced male subjects, whereas L.S. (age 25) was a female subject who was a naive observer. For each subject, a control experiment was initially performed to determine the threshold under the condition of steady viewing of the target. This also established the range of pulse displacement amplitudes that was appropriate for the convergence eye movement experiments. An HP9000 Model-550 Computer was used for stimulus generation and data recording (sampling frequency, 171 Hz). The data consisted of the binocular eye movements, the time of the pulse after convergence onset, and the subject’s response (first or second interval). In the analysis program, individual records containing right and left eye movements were first calibrated and then subtracted to give the vergence response (24). Records containing artifacts occurring near the transient portion of the convergence movements were rejected. Each response trace was then examined using an interactive graphical display to identify the beginning of the convergence movement. A movable cursor on the display screen was used to mark the position of the eye movement. As the cursor was moved from left to right, the beginning of the eye movement was indicated by the horizontal location just before an observable change in the cursor vertical position. The time delay between convergence onset and pulse stimulus, the amplitude of the pulse, and the subject’s response (first or second interval) were stored in the computer and grouped into time delay bins. For each bin, the percentage correct was plotted against stimulus amplitude. The threshold criterion for detection was arbitrally chosen as 75% correct responses in accordance with Manning (15). If the percentage correct was below 50%, it was given a score of 50% to represent chance level of occurrence. Combining the threshold results, the sensitivity (reciprocal of threshold) was plotted as a function of time delay. RESULTS

A typical vergence response is shown in Fig. 3. Displacement amplitude was converted to normalized logarithmic units (-LOG(AB/B)) with 0 = 4” of disparity target amplitude. The results were grouped into delay time bins and tabulated as percentage correct for different displacement amplitudes. An example is shown in Table 1 for subject GH. For each bin, a curve of percentage correct vs displacement amplitude was drawn. A typical curve is shown in Fig. 4 for the -50 to 0 ms bin. (subject G.H.). A horizontal line indicating 75% correct responses was drawn and its intersection with the curve

ET

AL.

0

I 0.5

I 1.0

5

TIME (set) FIG. 3. Vergence response to 4” disparity stimulus obtained from subtraction of movements of the two eyes. The surround displacement pulse could occur anywhere from 0 to 600 ms after disparity target onset. Subject G.H. The (+) symbol represents a movable cursor used during data analysis to indicate the start of the vergence response.

gave the detection threshold. This was done for all delay time bins for each subject. The threshold values were then plotted to give a sensitivity curve for different delay times relative to convergence onset. The sensitivity curves of subjects G.H. (A), J.S. (Cl), and L.S. (0) are shown in Fig. 5. Sensitivity at steady fixation was found to be 1.58, 1.59, and 1.78 log units for subjects G.H., J.S., and L.S., respectively. From the suppression curves, it is evident that the subjects exhibited loss of sensitivity even before convergence onset. Suppression encompassed a range of -200 to 350 ms relative to convergence onset with maximum loss occurring between 25 and 75 ms for G.H., 75 to 125 ms for J.S., and 50 to 100 ms for L.S. The maximum sensitivity loss was 0.26,0.31, and 0.26 log units for G.H., J.S., and L.S., respectively. The shapes of the suppression curves were similar except that for L.S., whose curve was displaced upward relative to the others. DISCUSSION

The reduction of sensitivity found in this study appears to be due primarily to central rather than peripheral factors. This could be attributed in part to the stimulus configuration, which was designed to minimize peripheral masking effects. Lateral backward masking, or metacontrast (1, 19), was essentially eliminated by the use of the brief 20-ms pulse displacement stimulus. The extent of lateral masking had been found to be about 23” from the target (34). The displacement stimulus used in this study consisted of horizontal lines that were at least 2.5” from the 1” disparity target. If the suppression found were due mainly to central neural inhibition, then a question arises as to how the visual system knows when to suppress the incoming signal. The central suppression signal must be time-locked to the vergence eye movement. It has been hypothesized that in the saccadic

SUPPRESSION

OF

SENSITIVITY

TO

TABLE Probability

SURROUND

303

DISPLACEMENT

1

of Detection of Surround Displacement

during Convergence (Subject G.H.)

Stimulus LOG

(ABIB):

0.09

A0 (degree): Delay

-1.65

-1.47

-1.35

-1.25

-1.17

-1.10

0.135

0.179

0.224

0.270

0.318

time

(ms)

n

-2oo--101 -lOO--51 -50-o l-50 51-100 101-200 201-300 301-350

77 20 58 23 31 39 15 9

Note.

n, number

of trials;

%C

n

%C

n

%C

n

%C

n

%C

n

%C

54.5 35.0 50.0 47.8 45.1 59.0 60.0 55.5

53 13 31 24 39 55 17 15

64.2 61.5 61.3 62.5 46.2 72.7 76.5 80.0

44 9 38 20 20 47 15 10

86.4 88.9 81.6 70.0 65.0 93.6 100 100

20 8 35 13 27 37 12 9

100 100 97.3 64.6 100 97.3 100 100

16 6 13 9 5 8 12 11

93.8 100 100 100 80.0 100 100 100

18 2 6 12 8 11 15 10

100 100 100 100 100 100 100 100

%C, percentage

correct

detection.

system, a corollary discharge, or outflow signal concomitant with the saccadic neural command, provides the frame of reference information and also inhibits processing of sensory input signals. This has been used to account for the finding that with extraocular muscle proprioception eliminated, monkeys compensated for perturbation of eye position produced by stimulation of the superior colliculus during intended saccadic movements in the dark (9). Corollary discharge has also been used to account for the finding that strabismic patients pointed at targets with their unseen arms in a direction opposite to that of the corrective surgical rotation of the eye (3). In this case, the patient pointed in the direction which would have been appropriate prior to surgery.

100 1

A simple demonstration of a corollary discharge signal in the vergence system can be found from a subjective experiment. If the outer canthi of the two eyes are pushed simultaneously, the world appears to move, whereas during normal vergence movements the world appears stable unless special attention is paid to the surroundings. As in the saccadic system, a vergence corollary discharge signal may provide the information on the

-5OtoOms

90

1 50 - 1.7

-200 - 1.6

- 1.5

STIMULUS

- 1.4

( LOG

- 1.3

- 1.2

- 1.1

(A9 I6 ) )

FIG. 4. Example of probability of detection curve for delay times of -50 to 0 ms relative to convergence onset (subject G.H.). Threshold for detection was arbitrarily chosen as 75% correct, giving a relative displacement threshold of -1.39 log units (solid line) for this subject. Probability curves for other time bins were similarly drawn to obtain the thresholds. This was repeated for the other subjects.

RELATIVE

-100

0

I 100

TIME OF STIMULUS TO CONVERGENCE

I 200

ONSET

300

(ms)

FIG. 5. Suppression of sensitivity to surround displacement as a function of stimulus timing relative to convergence onset. Note that suppression begins about 200 ms before and lasts until 350 ms after convergence onset. The maximum loss for G.H. (A) was 0.26 log units occurring at 25-75 ms, for J.S. (0) was 0.31 log units occurring at 75125 ms, and for L.S. (0) was 0.26 log units occurring at 50-100 ms.

304

HUNG

change, in depth, of the frame of reference, and also inhibit the signals related to changes in the surround. In this way, the visual scene appears stable despite the disruption from the vergence movement. The central suppression, however, could be accounted for by an alternative theory. The overload theory, proposed by Stark et al. (26), hypothesized that computation and execution of the intended movement so overwhelms the computational capacity of the visual system that further sensory information cannot be processed and is therefore discarded. In support of this theory are the findings that the duration of suppression is longer than the movement itself for each of the three systems investigated (saccade, blink, and vergence). A purely inhibitory mechanism would not require further inhibition after the completion of the movement. As discussed earlier, the reason central factors predominated in the results above was that stimuli which could have driven peripheral mechanisms were eliminated or minimized by the stimulus configuration. In fact, under ordinary room conditions, peripheral factors indeed do predominate, as complex visual scenes are particularly susceptible to masking effects. That is why “omission” (5) due to eye movements is so powerful under normal viewing conditions. This peripheral effect (1.22 log units for saccadic omission, Ref. (4)) is much larger than the 0.25 log units threshold increase to image displacement found in the present study and the 0.5 log units threshold increase to light flash found previously (1532). The time course for suppression of sensitivity to surround displacement is similar to that for light flash. This has been shown in the saccadic system (2), where the durations for both stimuli were about 150-200 ms with maximum suppression near the saccade onset. The results presented above now reveal that, for the vergence system as well, the time courses for suppression of sensitivity to displacement and light are similar. The total duration of suppression was found to be about 450-500 ms with maximum suppression occurring at about 50 ms after convergence onset. These findings add support to the concept that suppression is independent of the type of change in the surround, whether it is caused by a light flash or surround displacement. The suppression curve for L.S. (the naive, and also youngest, subject) was elevated relative to the curves for other subjects. Clearly, if experience in visual experimentation was a factor, the opposite would be expected since experienced subjects should show higher sensitivity. Most likely, this simply demonstrates variability of overall sensitivity for different individuals. Manning’s (15) sensitivity curves for light decrement stimuli showed a similar variability among individuals. The finding of suppression of sensitivity to surround displacement and light flash during vergence, along with the notion of “omission” for complex scenes, is consis-

ET AL.

tent with a recent theory of vergence eye movement control (11). The dual-mode model proposes that a fast preprogrammed, open-loop component responds to rapid changes in disparity, whereas a slow feedback controlled component follows slowly moving, small amplitude disparities. Due to the relatively long time delay (about 200 ms) and the time constant (about 200 ms) of the vergence response, feedback control could exhibit instability oscillations (11). A preprogrammed movement equal to the initial disparity stimulus, on the other hand, would give a stable response. Additionally, since the retinal images are smeared during a vergence response, continuous processing of the disparity signal could be rather inaccurate. Suppression serves to diminish this feedback information. Thus, suppression may not only be useful for the perception of image stability, but could also assist in oculomotor stability. Future studies may reveal additional properties of vergence suppression to light and movement. The vergence step disparity stimulus used in the movement increment study was fixed at 4”. An investigation of suppression as a function of vergence step size may provide additional parallels between the fast component of vergence and saccades, where the amount of suppression increases with increasing saccadic step size. Suppression during divergence, which had been shown for light flash (15,16) could also be examined for surround movement stimulus. Finally, increasing contours in the peripheral field, which has been found to increase suppression in the saccadic system, could be explored in the vergence system. REFERENCES 1. ALPERN, M. 1953. Metacontrast. J. Opt. Sot. Am. 29: 631-646. 2. BEELER, G. 1967. Visual threshold changes resulting from spontaneous saccadic eye movements. Vision Res. 7: 769-775. 3. BOCK, O., AND G. KOMMERELL. 1986. Visual localization after strabismus surgery is compatible with the “outflow” theory. Vision Res. 26: 1825-1829. 4. BROOKS, B. A., AND A. FUCHS. 1973. Influence of stimulus parameters on visual sensitivity during saccadic eye movements. Vision Res. 15: 1389-1398. 5. CAMPBELL, F. W., AND R. H. WURTZ. 1978. Saccadic omission: Why we do not see a grey-out during a saccadic eye movement. Vision Res. 18: 1297-1303. 6. DITCHBURN, R. 1955. Eye movements in relation to retinal action. Opt. Actu 1: 171-176. 7. DODGE, R. 1900. Visual perception during eye movement. Psychol. Rev. 7: 454-465.

8. FESTINGER, L., AND J. D. HOLTZMAN. 1978. Retinal image smear as a source of information about magnitude of eye movement. J. Exp. Psychol.

4: 573-585.

9. GUTHRIE, B. L., J. D. PORTER, AND D. L. SPARKS. 1983. Corollary discharge provides accurate eye position information to the oculomotor system. Science 221: 1193-1195. 10. VON HELMOLTZ, H. 1924. In Treatise on Physiological Optics (Translated by J. P. C. Southall), 3rd ed., Vol. 2, Optical Society of America, Rochester, NY.

SUPPRESSION

OF

SENSITIVITY

11. HUNG, G., J. L. SEMMLOW, AND K. J. CIUFFREDA. 1986. A dualmode dynamic model of the vergence eye movement system. IEEE Trans. Biomed. Eng. 33: 1021-1028. 12. HUNG, G., T. J. WANG, K. J. CIUFFREDA, AND J. L. SEMMLOW. 1989. Suppression of sensitivity to surround motion during vergence eye movements. Pages 203-204 in Proceedings, Fifteenth Annual

Northeast

Bioengineering

Conference,

Boston,

MA.

TO

25.

26. 27.

13. LATOUR, P. 1962. Visual threshold during eye movements. Vision Res. 2: 261-262.

14. LETTVIN, J. Y. 1960. Human vision. Q. Prog. Rep. Res. Lab. Electronics MIT 58: 254-258. 15. MANNING, K. A. 1986. Eye-movement-dependent loss in vision and its time course during vergence. J. Neurosci. 6: 1976-1982. 16. MANNING, K. A., AND L. A. RIGGS. 1984. Vergence eye movements and visual suppression. Vision Res. 24: 521-526. 17. MATEEFF, S., N. YAKIMOFF, AND L. MITRANI. 1976. Some characteristics of the visual masking by moving contours. Vision Res. 16:489-492. 18. MATIN, E. 1974. Saccadic suppression: A review and an analysis. Psychol. Bull. 81: 899-917. 19. MATIN, E., A. CLYMER, AND L. MATIN. 1972. Metacontrast and saccadic suppression. Science 178: 179-182. 20. MITRANI, L., N. YAKIMOFF, AND ST. MATEEFF. 1970a. Smearing of the retinal image during voluntary saccadic eye movements. VisionRes. 10:405-409. 21. MITRANI, L., N. YAKIMOFF, AND ST. MATEEFF. 1970b. Temporal and spatial characteristics of visual suppression during voluntary saccadic eye movements. Vision Res. 10: 417-422. 22. ORBAN, G., J. DUYSENS, AND M. CALLENS. 1973. Movement perception during voluntary saccadic eye movements. Vision Res. 13:1343-1353. 23. PEARCE, D. G., AND E. PORTER. 1970. Changes in visual sensitivity associated with voluntary saccades. Psychon. Sci. 19: 225227. 24. SEMMLOW, J. L., AND G. HUNG. 1986. Quantitative assessment of disparity vergence components. Znuest. Ophthalmol. Vis. Sci. 27:558-564.

28. 29. 30. 31.

32. 33. 34. 35.

SURROUND

37.

305

SEMMLOW, J. L., AND N. VENKITESWARAN. 1976. Dynamic accommodative response in binocular vision. Vision Res. 16: 403411. STARK, L., R. KONG, S. SCHWARTZ, AND D. HENDRY. 1976. Saccadic suppression of image displacement. Vision Res. 16: 11851187. STEVENSON, S. B., F. C. VOLKMANN, J. P. KELLY, AND L. A. RIGGS. 1986. Dependence of visual suppression on the amplitudes of saccades and blinks. Vision Res. 26: 1815-1824. VOLKMANN, F. C. 1962. Vision during voluntary saccadic eye movements. J. Opt. Sot. Amer. 52: 571-578. VOLKMANN, F. C., L. A. RIGGS, AND R. K. MOORE. 1980. Eyeblinks and visual suppression. Science 207: 900-902. VOLKMANN, F. C., L. A. RIGGS, A. G. ELLICOTT, AND R. K. MOORE. 1982. Measurements of visual suppression during opening, closing and blinking of the eyes. Vision Res. 22: 991-996. VOLKMANN, F. C., L. A. RIGGS, R. K. MOORE, AND K. D. WHITE. 1978. Central and peripheral determinants of saccadic suppression In J. W. SENDERS, D. A. FISHER, AND R. A. MONTY, Eds., Eye Movements and the Higher Psychological Functions. Erlbaum, Hillsdale, NJ. WANG, T. J. 1989. Suppression of Motion and Light Detection during Vergence Eye Mouements. M.S. thesis. Rutgers University, Piscataway, NJ. WARD, P. A., AND W. N. CHARMAN. 1985. Effect of pupil size on steady state accommodation. Vision Res. 25: 1317-1376. WEISSTEIN, N. 1972. Metacontrast. In D. JAMESON AND L. HURVICH, Eds., Handbook of Sensory Physiology, Vol. 4. SpringerVerlag, Berlin. YOUNG, L., AND D. SHEENA. 1975. Methods and design: Survey of eye movement recording methods. Behau. Res. Methods Instrum.

36.

DISPLACEMENT

7: 397-429.

ZUBER, B., A. CRIDER, AND L. STARK. 1964. Saccadic suppression associated with microsaccades. Q. Prog. Rep. Res. Lab. Electronics MIT 74: 244-249. ZUBER, B., AND L. STARK. 1966. Saccadic suppression: Elevation of visual threshold associated with saccadic eye movements. Exp. Neurol. 16: 65-79.