Data Loading...
Spatial aftereffects - Investigative Ophthalmology ... Flipbook PDF
Spatial aftereffects Ray Over Perception becomes distorted when the spatial properties of the visual stimulus are abrupt
121 Views
76 Downloads
FLIP PDF 867.4KB
Spatial aftereffects Ray Over
Perception becomes distorted when the spatial properties of the visual stimulus are abruptly changed. Psychophysical studies have shown some variables are critical for induction of spatial aftereffects and others have not. There have been few attempts to record single-cell response in the infrahuman visual system under stimulus conditions which yield aftereffects in human vision. The present paper outlines the assumptions of a model attributing aftereffects to neural inhibition. By consideration of psychophysical data within the context of this model, guidelines are provided for determination of the neural correlates of aftereffects.
A
the types of measures which could profitably be obtained within this approach.
stationary pattern is seen as moving from right to left when viewed following exposure to lines which moved left to right. A vertical line appears tilted in the clockwise direction when its orientation is judged following inspection of counterclockwise tilted lines. Similar aftereffects can be induced in the perception of size, length, proximity, visual direction, curvature, periodicity, and stereoscopic depth. Over the last 30 years a substantial body of psychophysical data has accumulated indicating that some variables are critical for induction of spatial aftereffects and others are not. Within the same period explanations1"4 have been offered of after-effects by reference to proactive neural inhibtory mechanisms. These accounts involve statements about functional interactions within populations of spatially tuned feature detectors. Such guesses about neural response characteristics can be educated only to the extent they are in accord with what can be demonstrated about neural functioning by direct methods. There unfortunately have been few attempts to record single-cell response in the infrahuman visual system under stimulus conditions which induce spatial aftereffects in human vision. The main aim of the present paper is to indicate
A general model of spatial aftereffects
Assumptions underlying a general model of spatial aftereffects can be developed from reference to the study in which Barlow and Hill5 measured the discharge of ganglion cells in the rabbit retina before, during, and after the eye was stimulated by moving contours. Motion in the cell's nonpreferred direction did not change discharge frequency from the cell's spontaneous activity level. However, movement in the opposite (preferred) direction resulted initially in a large increase in neural response before a steady rate was attained. Termination of motion immediately suppressed the cell's response, and recovery to spontaneous activity level was not complete (following one-minute exposure to a figure moving at 15 degrees per second) until 45 seconds later. Barlow and Hill5 proposed that the movement aftereffect occurs in human vision because the pattern of neural activity generated when a stationary figure is displayed after exposure to a moving figure is similar to that normally produced by slight, but maintained, motion in the direction which was null during inspection. This imbalance occurs as the result of suppression of cells which were active during inspection. To extend this account to the general
From the Department of Psychology, University of Queensland, St. Lucia 4087, Australia.
533
534
Over
class of aftereffects, four assumptions need to be made: 1. The human visual system contains input analyzers which are selectively tuned to specific spatial properties of the visual stimulus ("analyzer rule"). 2. The brain establishes the spatial properties of the visual stimulus by reference to differential excitation patterns between subsets of spatially tuned analyzers ("ratio rule"). 3. The degree to which an analyzer's activity is suppressed following a period of specific stimulation is a function of the analyzer's level of excitation during this stimulation period ("suppression rule"). 4. Suppression in the activity of an analyzer will have perceptual consequences only if the analyzer's spatial tuning includes the postinspection as well as the inspection stimulus value ("tuning rule"). Operating within these assumptions the present paper considers whether it is possible to establish links between aftereffects measured by psychophysical techniques and neural response functions established by microelectrode exploration of the infrahuman visual system. Attention should also be given to extensive indirect evidence on feature processing in the human visual system obtained from evoked potential measures0 and psychophysical studies employing masking paradigms.7 Such data are highly relevant to and complement accounts of aftereffects, but they must of necessity receive secondary attention here. Prior to the main discussion some general comments should be made about visual spatial aftereffects. First, they are topographically restricted in that they can be induced only when the inspection and test figures are displayed in the same or in closely adjacent regions of the visual field. Second, aftereffects can be induced with monoptic, dichoptic, or binocular display of contours, and the aftereffect is reduced under dichoptic condtions. Third, the aftereffects under consideration should not be confused with errors made in spatial judg-
Investigative Ophthalmology June 1972
ment after an observer has been actively exposed to optically transformed visual input. There is compelling evidence8 that these errors reflect a behavioral compensation process rather than sensory inhibition. Fourth, with some few exceptions2 it is not possible to generate a visual aftereffect by spatial stimulation within another modality. Fifth, a number of judgmental and response factors can affect the magnitude of an aftereffect,9 and these should be controlled in parametric determination of stimulus effects. Last, techniques are available to measure spatial aftereffects with infrahuman observers,10 but there have been few such studies. Temporal determinants of spatial aftereffects
An aftereffect is generated by an abrupt change in the spatial value of the visual stimulus, and the error decays exponentially at a rate dependent on the period of time the observer was exposed to the inspection stimulus. If allowance is made for decay in the interval between cessation of inspection and completion of the test judgment, the initial magnitude of the aftereffect is independent of the period of inspection.11 Absence of a time-growth function for spatial aftereffects suggests, in terms of the Barlow and Hill5 model, that the initial extent of postexcitatory suppression should depend solely on a cell's prior activity level (determined by the spatial value of the inspection stimulus) while the rate of recovery should reflect the period of prior stimulation. Unfortunately physiologists have totally neglected parametric study of the dynamic response properties of spatial detectors. There are several complications, however, to the simple approach to decay of aftereffects just outlined. Masland12 found that the motion aftereffect induced by 15 minute inspection persists for more than a day; in addition long-lasting spatially linked color aftereffects can be generated by prolonged inspection.13 Most studies of aftereffect have employed inspection
Volume 11 Number 6
periods of 2 minutes or less. The question of interest (and it is basically an experimental issue) is whether prolonged exposure to a maintained stimulus results in long-term neural modification which differs in mechanism from the relatively brief postexcitatory suppression demonstrated by Barlow and Hill.5 The second complication is that the pattern characteristics of a stationary figure interpolated between cessation of inspection and presentation of the test stimulus can affect the rate at which the motion aftereffect decays.14 "Storage" in motion aftereffect can occur if a dark interval is interpolated. What happens to postexcitatory suppression under these conditions? Spatial determinants of magnitude of aftereffect The magnitude of a spatial aftereffect depends on the difference in spatial value between the inspection and test stimuli. This suggests the possibility of establishing linkages between spatial functions for aftereffects and neural tuning functions for spatial analyzers. Three measures of tuning are of interest: the activity level of a single unit at different stimulus values, variability in neural representation on repeated stimulus presentation, and the distribution of units in terms of the stimulus value at which each is maximally responsive. Physiologists characteristically report the first and third of these measures but not the second. Ideally, variability in spatial representation should be determined to allow the information-signaling capacities of neural analyzers to be expressed and conceptualized within a signal detection framework. This matter is important in that perceptual discrimination can be most readily compared with neural differentiation if both are expressed in the one metric. The following discussion is restricted to the spatial determinants of motion and tilt aftereffects. The general argument is that by recording the dynamic as well as maintained response properties of a large number of single units the physiologist can assemble a synthetic population. Output func-
Spatial aftereffects 535
tions can be obtained from this population by a process of response pooling in association with the "ratio rule" outlined earlier. These can be compared with aftereffect measures (which preferably have been obtained from the same species). The process of comparison is inevitably indirect, but clues about extraction of information in the visual system may follow from determining which pooling principle yields output functions that best fit psychophysical functions. The motion aftereffect is maximal if lines moving at velocities between 1.5 and 4 degrees per second had been inspected before the stationary test pattern was viewed.15 This inverted U-shaped function would be expected if motion detectors in human vision are broadly tuned to velocity, with most units maximally responsive in the medium range. The aftereffect would be minimal following inspection of rapidly moving contours because most motion detectors would be below the threshold for excitation during inspection and would not subseqeuntly be suppressed. Postexcitatory suppression would also be slight if few units were highly excited by slow motion. This analysis of the spatial function of the motion aftereffect could have received considerable indirect support if Pettigrew, Nikara, and Bishop10 had extended their study of velocity selectivity in the cat visual cortex to include measures of activity before and after cells with different velocity selectivity had been excited. By considering their sample as a synthetic population they could have determined pooled poststimulation response as a function of velocity during stimulation. The motion aftereffect is maximal when a stationary vertical grating is viewed following lateral movement of a vertical grating and is diminished with variation in the orientation and direction of motion of the inducing contours.17 The direction selectivity of the aftereffect extends over a 45 degree range. Single cells in the cat visual cortex show angular selectivity to moving gratings,ls and it would be simple to establish whether postexcitatory suppres-
536 Over
sion in the case of such cells depends on the direction of image motion during excitation relative to the orientation of the stationary grating subsequently displayed. The possible correlates of further determinants of the motion aftereffect can be examined in the same way. For example, the aftereffect is impaired if the stationary test figure is viewed under stroboscopic illumination19; what happens to postexcitatory suppression during stroboscopic illumination? In addition the motion aftereffect induced under scotopic conditions transfers to photopic vision20; is there common neural processing of motion within the two systems? Further, color-specific motion aftereffects21 can be induced in addition to motion-specific color aftereffects.15 What is the neural basis of perceptual interactions between wavelength and spatial features in human vision? Last, the motion aftereffect is markedly reduced unless the stationary test figure is seen against a patterned surround.22 This is of interest in view of a recent demonstration23 that the response to motion of cells in the cat visual cortex can be modified by a stationary surrounding pattern. The spatial determinants of the tilt aftereffects are well documented. In central vision a vertical line appears displaced about two degrees counterclockwise following exposure to lines tilted 20 degrees in the clockwise direction from vertical and inspection of lines tilted 75 degrees results in clockwise distortion of one degree or so. Why do lines repel each other in the former case and attract each other in the latter? It has been proposed4 that repulsion occurs only when the inspection and test lines fall within the excitatory range of single tilt analyzers. The aftereffect thereby occurs because some cells which signal the orientation of the test figure are excited during inspection and are consequently suppressed when the test figure is introduced, while the response of other cells which are tuned to the inspection but not to the test orientation is unaffected when the test figure is
Investigative Ophthalmology June 1972
presented following inspection. The aftereffect would be maximal following inspection of 20 degree tilted lines if single tilt analyzers have excitatory tuning extending 20 to 30 degrees from a preferred angle. To explain contour attraction found when a vertical test line is seen following inspection of 75 degree tilted lines, it is necessary to assume cells excited by vertical lines are suppressed below spontaneous activity level by display of a 75 degree grating and become overexcited when the vertical grating is subsequently shown. Do tilt analyzers exhibit an S-shaped excitatoryinhibitory tuning function, and how do they respond when orientation is abruptly changed? These questions could be answered by considering neural response within a synthetic population under stimulus conditions which induce the tilt aftereffect in human vision. Only the repulsion form of tilt aftereffect can be induced in peripheral vision24; contour attraction no longer occurs. This result is explicable if cells have broader orientation tuning in peripheral than central vision, for example, by an arrangement such that receptive fields in central vision have small concentrated centers and large strongly antagonistic surrounds and in peripheral vision large centers and less influential flanks. Individaul cells in the periphery would thereby be excited over a wider range of orientations than foveal cells and be suppressed rather than overexcited when the vertical line is introduced following exposure to an extreme tilt. Aftereffects and illusions Many sets of contours which yield aftereffects when displayed successively constitute geometric illusions when shown simultaneously. A satisfactory explanation of geometrical illusions has not yet been developed,25 and many difficulties confront attempts to explain all aspects of illusions in terms of a single error-inducing mechanism. In particular, there is strong spatial determination of the magnitude of an illu-
Volume 11 Number 6
sion, but attentional and experiental factors can vary the perceived error appreciably. In addition, with some displays the illusion and aftereffect are strikingly similar, but in other cases simultaneous and successive presentation of contours results in opposite perceptual errors.20 The orientation illusion in which a vertical line appears tilted when seen against a background of tilted lines can, however, be considered in terms of a neural inhibitory mechanism. Disinhibition as well as inhibition can be demonstrated with this illusion.27 Border contrast analogous to the Mach band effect can also be demonstrated by displaying a column of vertical lines separated from a row of horizontal lines by a series of lines varying in tilt from vertical to horizontal.28 The vertical and horizontal lines at the points of angular discontinuity within the display appear tilted in the direction opposite to the adjacent tilted line provided the step size within the gradient is not too great. In a recent microelectrode study, Burns and Pritchard29 attempted to establish the neural correlates of the tilt illusion by positioning a contour for optimal response and examining variation in discharge frequency when a further line was introduced into the cell's receptive field. They found that response did not shift in a manner consistent with the illusion, but all measures were taken with the two lines forming an angle of 30 degrees. This is an inappropriate value; in human vision perceptual distortion is maximal when lines differ in tilt by 10 or 20 degrees and minimal when they differ by 30 degrees. The mechanism of the tilt illusion can probably best be identified through study of tilt aftereffects in the electrophysiologic context. In perceptual terms the two effects are strikingly similar; for example, the tilt illusion differs between central and peripheral vision in the same way as the tilt aftereffect.30 It is probable the same neural enhancement process underlies the two effects but operates in proactive form in the case of the aftereffect.1
Spatial aftereffects 537
Nonvisual aftereffects
There have been many claims that test judgments in one modality can be modified by prior stimulation in another modality. Physiologists (and psychologists) should be careful about trying to interpret crossmodal aftereffects within the neural inhibition model by reference to polysensory functioning; it is quite likely crossmodal aftereffects reflect properties of the response system such as a shift in frame of reference.2 It is necessary to demonstrate that crossmodal aftereffects decay at the same rate as intramodal aftereffects before they can be considered in terms of inhibition. Intramodal aftereffects can be induced in all spatial senses. For example, a horizontal rod feels tilted following kinesthetic inspection of a tilted rod; an analysis of this aftereffect has been offered in terms of neural inhibition.31 Studies of similarities and differences in spatial aftereffects between senses can yield important clues about the generality of input-processing mechanisms. For example, the issue of whether directionally sensitive units exist in modalities other than vision can be examined indirectly by determining whether direction-of-motion aftereffects occur in audition or following mechanical stimulation of the skin. The classical concepts of punctate sensitivity and point-by-point processing still dominate electrophysiologic study of the response of the skin. Recent psychophysical studies32 have shown the skin has rich spatial selectivity under conditions of dynamic stimulation; the classical punctate concepts are applicable only to perception with static stimulation. The question inevitably arises as to whether edge detection, motion analysis, and periodicity processing occur in relation to input from the skin. Put in other terms, "What does the frog's skin tell the frog's brain?" The psychologist can provide guidelines by examining perceptual functioning (including aftereffects) on the skin.
Investigative Ophthalmology June 1972
538 Over
Concluding remarks
The psychologist measures input-output relationships, and from these attempts to infer the logic of the system mediating the obtained relationships. The perceptual events studied by the psychologist inevitably reflect the pooled response of neural units which differ in terms of breadth of tuning, preferred value, and temporal characteristics. There always will be alternative logics potentially underlying specific perceptual functions, and the approach to explanation will inevitably be indirect. The alternatives can be narrowed by compounding psychophysical observations. However many psychologists presently believe the most valuable form of circumstantial evidence can come from direct measurement of neural processing in the infrahuman visual system. It is somewhat paradoxical that the psychologist, who must work with perceptual measures reflecting pooled neural activity, can gain greater insight into potential neural mechanisms from single-cell data than from compound measures. It is of concern to many psychologists who study spatial aftereffects that physiologists have not in general been interested in measuring neural response under conditions which yield consistent and intriguing distortions in perceptual judgment. This concern is justified as there now is an extensive body of psychophysical data, as well as a number of models, which can provide physiologists with clear guidelines. REFERENCES 1. Canz, L.: Mechanism of figural aftereffects, Psychol. Rev. 73: 128, 1966. 2. Day, R. H., and Singer, G.: Issues in the explanation of sensory adaptation and aftereffect, in Jarvinen, J., editor: Contemporary research in psychology of perception, Helsinki, 1969, Soderstrom. 3. Coltheart, M.: Visual feature-analyzers and aftereffects of tilt and curvature, Psychol. Rev. 78: 114, 1971. 4. Over, R.: Comparison of normalization theory and neural enhancement explanation of negative aftereffects, Psychol. Bull. 75: 225, 1971. 5. Barlow, H. B., and Hill, R. M.: Evidence for a physiological explanation of the water-
6.
7.
8.
9. 10.
11.
12. 13. 14.
15.
16.
17.
18.
19.
20.
21. 22.
fall phenomenon and figural aftereffects, Nature 200: 1434, 1963. Campbell, F. W., and Maffei, L.: Electrophysiological evidence for the existence of orientation and size detectors in the human visual system, J. Physiol. 207: 635, 1970. Weisstein, N.: What the frog's eye tells the human brain: Single cell analyzers in the human visual system, Psychol. Bull. 72: 157, 1969. Day, R. H., and Singer, C : Sensory adaptation and behavioral compensation with spatially transformed vision and hearing, Psychol. Bull. 67: 307, 1967. Over, R.: Individual differences in figural aftereffects, Psychol. Bull. 74: 405, 1970. Scott, T. R., and Powell, D. A.: Measurement of a visual motion aftereffect in the rhesus monkey, Science 140: 57, 1963. Oyama, T.: Experimental studies of figural aftereffects: I. Temporal factors, Jap. J. Psychol. 23: 239, 1953. Masland, R. H.: Visual motion perception: Experimental modification, Science 165: 819, 1969. Hepler, N.: Color: A motion contingent aftereffect, Science 162: 376, 1968. Strelow, E., and Day, R. H.: Aftereffect of visual motion: Storage in the absence of a patterned surround, Percept. Psychophysics. 9: 485, 1971. Scott, T. R., and Noland, J. H.: Some stimulus dimensions of rotating spirals, Psychol. Rev. 72: 344, 1965. Pettigrew, I. D., Nikara, T., and Bishop, P. O.: Responses to moving slits by single units in cat striate cortex, Exp. Brain Res. 6: 373, 1968. Over, R., Broerse, J., Lovegrove, W., and Crassini, B.: Effect of the periodicity, velocity, and direction of motion of the inducing grating on movement aftereffect. In preparation. Campbell, F. W., Cleland, B. C , Cooper, G. F., and Enroth-Cugell, C : The angular selectivity of visual cortical cells to moving gratings, J. Physiol. 198: 237, 1968. Anstis, S. M., Gregory, R. L., Rudolf, N., and MacKay, D. M.: Influence of stroboscopic illumination on the after-effect of seen movement, Nature 199: 99, 1963. Anstis, S. M.: After-effect of seen motion: Transfer from rods to cones and vice versa, Nature 201: 952, 1964. Lovegrove, W., Over, R., and Broerse, J.: Colour selectivity in motion aftereffect. In press. Day, R. H., and Strelow, E.: Reduction or disappearance of visual aftereffect of move-
Volume 11 Number 6
23.
24.
25. 26. 27.
28. 29.
30.
31.
32.
ment in the absence of patterned surround, Nature 230: 55, 1971. Burns, B. D., and Webb, A. C : The effects of stationary retinal patterns upon the behaviour of neurons in the cat's visual cortex, Proc. Roy. Soc. B. 178: 63, 1971. Muir, D., and Over, R.: Tilt aftereffects in central and peripheral vision, J. Exp. Psychol. 85: 165, 1970. Over, R.: Explanations of geometrical illusions, Psychol. Bull. 70: 527, 1968. Over, R.: Lateral inhibition explanation of geometrical illusions, Nature 222: 99, 1969. Blakemore, C , Carpenter, R. H. S., and Georgeson, M. A.: Lateral inhibition between orientation detectors in the human visual system, Nature 228: 37, 1970. Over, R., and Alexander, R.: Border contrast in the perception of contour orientation. In preparation. Burns, B. D., and Pritchard, R.: Geometrical illusions and the response of neurons in the cat's visual cortex to angle patterns, J. Physiol. 213: 599, 1971. Over, R., Broerse, J., and Crassini, B.: Orientation masking and illusion in central and peripheral vision, J. Exp. Psychol. In press. Over, R.: Effect of the angle of tilt of the inspection figure on the magnitude of a kinesthetic tilt aftereffect, J. Exp. Psychol. 74: 249, 1967. White, B. W., Saunders, F. A., Scadden, L., Bach-y-Rita, P., and Collins, C : Seeing with the skin, Percept. Psychophysiol. 7: 23, 1970.
Discussion DAY: Singer and I were unable to induce intermodal aftereffects between vision and kinesthesis. Stimulation of the kinesthetic system did not produce a visual aftereffect, and vice versa. On a different matter, it has recently been demonstrated there is no movement aftereffect on the skin (Hazlewood, V.: A note on failure to find a tactile motion aftereffect, Aust. J. Psychol. 23: 59, 1971). OVER: It was the Day and Singer data2 that led me to urge caution in interpreting reports of crossmodal aftereffects within the neural inhibition model by reference to polysensory functioning. Concerning the tactile motion aftereffect, it would occur only if direction of motion is analyzed in the tactile system through opposed detectors and perception is measured when these detectors have been selectively adapted. I would like to see use of dynamic stimulation of the skin in the manner employed by White et al.32
Spatial aftereffects 539
RODDECK: It is interesting that you mentioned the Barlow-Hill experiment. I always have taken their evidence to indicate exactly the opposite, that this illusion is not found at this level in the rabbit. The argument is that neurons can signal better by an increase rather than a decrease in firing particularly when their spontaneous rate is low. If one would expect that this illusion occurs here one would also expect the firing rates in the null direction would dramatically increase at the termination of the movement and this doesn't happen. OVER: The Barlow and Hill model simply requires that an imbalance occur. The low velocity of motion represented within the aftereffect is consistent with a mechanism involving suppression alone. LEVICK: It is interesting that the apparent velocity of the aftereffect of seen motion is very much less than the optimal stimulus velocity for inducing it. This fact would fit in well with the Barlow-Hill hypothesis that the aftereffect is attributable to an imbalance in the poststimulation discharges of pairs of direction-selective retinal ganglion cells sensitive to oppositely directed motions. Experimentally, the imbalance appeared as a depression of the discharge of the unit excited by the inducing stimulus, the other partner of the pair being unaffected during or after stimulation. Thus, the maximum imbalance attainable would equal the maintained discharge of the latter partner. Since the maintained discharge of directionselective units is relatively low, so also would be the maximum imbalance and therefore the subjective experience of the aftereffect. The hypothesis further suggests the idea of estimating the maintained discharge of specific classes of visual neurons from the maximum magnitude of the aftereffect induced by stimuli for which the particular class of neurons is selective. OVER: I fully agree with this account. DAW: It wasn't clear to me either from your talk or Professor Day's talk how much the tilt aftereffect is due to cyclorotational movements of the eye. OVER: The tilt aftereffect is totally independent of this factor. (See Howard, I. P., and Templeton, W. B.: Visually-induced eye torsion and tilt adaptation, Vision Res. 4: 433, 1964.) DAW: YOU seem to be suggesting a generalization. The longer the inspection the longer the aftereffect but the strength of the aftereffect is independent of inspection period. Do you think this is true of color afterimages? Subjectively it doesn't seem to be true? OVER: Many- experimenters have reported timegrowth functions for spatial aftereffects, but test settings were not measured immediately (within 20 msec.) on cessation of inspection. They obtained
Investigative Ophthalmology June 1972
540 Over
development functions because aftereffects decay at a rate dependent on the period of inspection. It is necessary to do careful psychophysics11 to unconfound these factors. There might be value in examining color and brightness adaptation by the same approach. SCHILLER: YOU showed an illusion occurring as vertical oriented lines rotated around a corner to become horizontally oriented lines, and you attributed it to a Mach band-like effect between orientated units. I found that if I fixed my eye
on one point I couldn't get the effect and wonder whether it is due instead to eye movements? OVER: This could be answered by measuring the illusion with tachistoscopic presentation of the figure or during retinal stabilization. We have not studied the effect in this way, but Pritchard found that geometrical illusions are still produced with stabilization (Pritchard, R.: Visual illusions viewed as stabilized retinal images, Q. J. Exp. Psychol. 10: 77, 1958).
Cyclopean perception and neurophysiology Bela Julesz
Xhysiologic psychologists interested in visual perception have frequently exploited binocular techniques in order to distinguish retinal from cerebral processes. These classical binocular techniques either exploit interocular transfer or are based on dichoptic stimulation. In the first case only one eye is stimulated while in the latter the two eyes receive very different stimulation and therefore binocular rivalry is usually unavoidable. Because of this rivalry, negative results do not imply that peripheral processes are at work (as so many researchers erroneously assumed), and even in case of a positive outcome the implications are far from clear. Interocular transfer, dichoptic techniques, and cyclopean methodology In order to illustrate the difficulties with classical binocular techniques, Fig. 1 shows a positive dichoptic result taken from a study by Schiller and Wiener.1 Here the Ebbinghaus illusion is portrayed such that the test discs are presented to one eye's view, while the inducing discs are given to the other. Undoubtedly, despite strong Frem the Bell Telephone Laboratories, Incorporated, Murray Hill, N. J.
binocular rivalry, the combined binocular view yields an optical illusion. They tried to reduce binocular rivalry by using brief presentations and showing the left and right fields in brief succession. Even so, for most of the illusions tried, the dichoptic conditions yielded less illusory effects than the classical conditions (when both test and inducing figures were combined and were presented to the same or both eyes). The positive outcome implies that the processes responsible for the illusion are partly central, but the reduced illusory effect can have two interpretations: It might be that the illusory effect is the result of both central and peripheral processes at work, and with dichoptic stimulation only the central component is stimulated. It might also be that the illusory effect is entirely due to central processes but, because of binocular rivalry, the left and right fields cannot be fully combined into a single view. The question arises whether it would be possible to find some technique which would portray the desired message only at a central site and would accomplish this without binocular rivalry. In the case of classical stereoscopic fusion there is, of course, no rivalry; however, the left and