0042-6989/92 65.00 + 0.00 Copyright 0 1992 Pergamon Press Ltd
Vision Res. Vol. 32, No. 6. pp. 1169-1175, 1992 Printed in Great Britain. All rights reserved
Research Note Control of Interocular Suppression as a Function of Differential Image Blur TRACEY
J. SHORS,*$
KENNETH
WRIGHT,?
ERNEST
GREENE*
Received 18 June 1990; in revised form 18 July 1991
This experiment studied normal adults, normal children, and corrected strabismics with alternating fixation, evaluating their ability to detect information presented to one eye under viewing conditions which might affect their eye preference. An apparatus was used which allowed unconstrained, binocular viewing of a video monitor, but in which the image seen by one eye was blurred while that seen by the other eye was in sharp focus. Under these viewing conditions, subjects were asked to identify a sharply focused letter which was shown briefly to the eye which was seeing the blurred background image or to the one which was watching the well-focused image. The question was whether there would be general suppression of the eye that was shown a degraded background, causing a reduction in its ability to process other information being presented. The results indicate no performance penalty for children, a mild decrement in recognition accuracy for normal adults, and a severe deficit in the ability of corrected strabismics to attend to the eye which views a blurred image. These results suggest that the loss of binocular vision that accompanies childhood strabismus is not a result of an inherent sensitivity to blurring, insofar as such vulnerability is reflected by these acute testing conditions. Strabismus Interocular suppression Binocular rivalry Selective attention
Visual cortical suppression is the unconscious and involuntary inhibition of part or all of the visual field from one eye under binocular viewing conditions (Moses, 1981). The fundamental observation is that patients with strabismus and unilateral anisometropic ambliopia are unaware of visual information being presented to the nonfixating eye. The term “suppression” has been broadly applied to various manifestations of this effect, and when chronically present in childhood, is thought to progress into a total dependence on one eye-i.e. amblyopia-and/or loss of binocular vision. A blurred image to one eye and stabismus are the critical factors in the development of clinical suppression. In children, a differential refractive error from anisometropia causes monofixation by the eye with the sharper focus and suppression of the central visual field of the blurred eye. Thomas (1978) and Parks (1984) have suggested that these suppression mechanisms apply also for the patient with uncorrected stabismus even if each eye has 20120 vision. Where the two eyes are deviated, different images fall on corresponding retinal
*Department of Psychology, University of Southern California, Los Angeles, CA 90089, U.S.A. tDepartment of Ophthalmology, University of Southern California, Los Angeles, CA 90089, U.S.A. fPresent address: Department of Psychology, Princeton University, Princeton, NJ 08540, U.S.A.
Eye preference
locations and thus the objects which are presented to each eye will overlap-causing visual confusion. Thomas (1978) hypothesizes that the dominant eye selects the focal distance for both eyes, which makes it likely that the objects registered by the subordinate eye will be out of focus. By this analysis, it is the differential in focus between the two eyes-not rivalry mechanisms-which provides the substrate for clinical suppression. Thus an initial small difference in eye preference may progress into an irreversible amblyopia, or where the two eyes are well balanced, an inability to synthesize the two retinal images even though the deviation is surgically corrected. We evaluated the effects of background blur on visual control of normal adults, normal children, and patients that were “corrected strabismics with altemating fixation”, meaning that they previously had been surgically treated to correct strabismus, had adequate vision with either eye, and when tested could switch from one eye to the other depending on stimulus conditions. It is important to note that for these patients the surgical correction, which was done after the critical period for binocular development, remedied motor coordination of the eyes but came too late to prevent the loss of binocular vision. The present apparatus and methods were essential for testing such patients, since other devices-such as Polaroid glasses or haploscopic devices-generally induce dissociation of the visual system and break
1169
1170
RESEARCH
TABLE 1. Visual
Patient
Age at testing (Yr)
Age at surgery W
t
10
9
2 3 4 5 6 7
9 19 35 13 19 28
8 19 7 13 19 28
NOTE
~~~r~ct~~stjcs of clinical subjects Acuity (right eye)
Acuity (Ieft eye)
worth
20120 20/20 20130 20!20 xi20 2Oj2S 20125
2oj20 20/20 2OJ20 20120 20120 2OJ25 20125
suppressed suppmsed Suppressed Suppressed
4-dot
S%lppRSSed
SuQprMed SupQriWd
TitttlUS None None None None None None None
down binocular fusion. This confounds the issue of eye coor~nation. Age and visual char~te~s~~ of whether selective attention to a given eye is dete~in~ the patients are reflected in Table 1. Ocular deviations by the quality of its image or diptopia that has been for each patient were measured using standard strabisinduced by the apparatus. The present apparatus allows mus techniques, including: (a) cove&-uncover testing; undissociated binocular vision and also provides the (b) simultaneous prism cover testing; and (c) alternate potential to blur the background while simultaneously cover testing. These tests detect tropia, tropia with presenting a clear and focused target stimulus to that peripheral fusion and the full dissociated deviationeye. “pboria”. Under the demands of testing, I patient We ~va~~t~ the ability of these subjects to report showed slight esotropia with dissociated vertical deviletters that were presented in clear focus to a given eye ation. The other patients showed extremely good eye while blurring the ~ck~o~~d that was being shown coordination under ali test conditions. to that eye. This was accomplished by means of an apparatus containing three computer/video monitors, Apparatus and stimuli beam-splitters and a + 10 D lens which could be placed The general configuration of equipment for control in the viewing path of either the left or right eye. Thus of images to the two eyes is shown in Fig. f . A color the subject was provided with relatively fry-dewing television monitor, bordered by a white cire%ewith conditions in which they could watch an animated bisecting lines, was placed 0.5 m in front of the subcartoon with both eyes, but with the image provided to ject, whose head position was stabilized by the use one of the eyes being blurred while the other was sharp. of a chin rest. At this distance the image from the Further, by means of the lateral monitors and beam monitor subtended 6 deg of visual angle, A color cartoon splitters, sharply focused letters could be flashed to either was displayed on the monitor (from a video cassette eye, which the subjects were required to identify and recorder), which had a mean luminance of 75 cd/m2, and report, The key question was whether there would be with contrast set at 65% of the range of the instrument. a d~e~ntial in reporting accuracy as a function of the The cartoon image served as a ~ck~ound field for the focus of the background image-i.e. whether the eye detection task, and for the expe~menta~ control of eye being shown the bluned image would be refractory to preference as described below. A beam splitting mirror was placed in the path of these letters even though they were being presented in vision for each eye, which allowed images from. two sharp focus to the fovea. additional black-and-white monitors to be superimposed upon the subject’s view of the cartoon, and at tbe proper focus. The only mate~al shown on these mon~hrome mo~tors were white capital Mters that were presented for a 16 msec interval, one at a There were three groups of subjects, consisting of time, and whiih were always positi&ed at the center of 20 normal-vision adults (ages 23-4S), 6 normal-vision the screen. The remainder of the screen, as well as the children (age 6ll), and 7 patients (ages 9-35) who monitor itself and surrounding equipment were not had been surgicaliy corrected for strabismus. All subjects ill~inated, By this &on~guration, the letter stimuli were given a complete opht~rno~o~~~ examination, from each monochrome monitor cotdd be superwhich included the Snellen test for visual acuity and tests imposed upon the subject’s view of the color mo~tor, for stereoscopic fusion using the Worth 4 Dot and the and could be presented to either -the left or the right Titmus. In mkr to qualify for one of the normal-vision eye (depending on which mon~~r~rne monitor was groups, the sub&x$had to demonstrate acuity of 20120 used to dehver the stimulus). Thus the composite image (with corm&ion) for each eye, to fuse the elements of that was seen by each eye cot&ted of its view of the Worth 4 Dot at near and far distances, and to show the cartoon being displayed, and a white capital letter stereopsis of at least 40 set of arc on the Titmus. None which would appear periodically, For 16 msec, in the center of the screen. of the normal subjects showed ocular deviation. A major e~~~rn~n~ marination was made The s~b~s~~ patients were s&x%& on the basis of high acuity in each eye (with correction) and good possible by adding a + 10 D lens in the viewing path for
RESEARCH
as shown in Fig. I. The fens WAS placed between the beam splitting mirror and the color monitor, and served to degrade the retina1 image of the cartoon while leaving the image from the monochrome monitori.e. the capital letter-in sharp focus. The control condition was provided by biurring the background image for the opposite eye. As detailed below, the issue is not whether one can discriminate sharply focused letters against a blurred background, but whether the background is able to bias the use of that eye and thus work at odds with the demands of the discrimination task. In order to reduce the complexity of the display system, only symmetrical letters-i.e. A, H, M, T, V, W, X and Y-were used as the test stimuli. As seen by the subject, the vertical dimension of the letters subtended 47 min of visual angle. At1 sampling from the letter pool, selection of side for the display (left or right), as well as interstimulus interval was determined at random. The interval between successive presentations ranged from 2 to 4 sec.
NOTE
1171
an eye,
The subject was seated with his head in the chin rest, and the apparatus was aligned with the aid of stimulus markers until the monitors were properly centered for that individual. The beamsplitting mirrors were adjusted to align the image from the side monitors into the center of his visual field, and cover~uncover tests were used to verify the alignment. The subject was then instructed to concentrate (to focus) on the’cartoon movie being shown on the color monitor, but also was told to report the letters that would briefly appear. The subject was then given a trial run of 10 letters to each eye to familiarize him with the procedure, and to be certain that he could reliably report letters which were shown for 16 msec. A sequence of test trials was administered which included monocular and binocular viewing conditions. The monocular trials were used to screen subjects and reject any subjects who could not accurately detect letters when all of the stimulus material (letters as well as background) was presented to each eye alone.
FIGURE 1. A experimental configuration was designed to allow one eye to have a cleat, well-focused view of a central monitor while the image to the other eye was blurred by means of a IO D lens. Additionally, beam-splitters were placed in the path of view which allowed letters to be gashed briefly on a lateral monitor, thus providing a well-focused letter to the fovea of one eye. The task requires that the subject identify and name the letter which was shown. The upper panel illustrates the letter being presented to the eye which is viewing a blurred background image-this being designated the “ipsilateral eye” condition, For the “contralateral eye” condition the blur is presented to the opposite eye, and the letter must be identified by the eye which has a dear view of the background cartoon. The figure illustrates the presentation of letters to the left eye-with reiations and terminology reversing when they are shown to the right. (Note that positioning of elements is for the purposeof illustration, and distances are not to scale.)
1172
RESEARCH NOTE
Under these monocular test conditions, every subject that was included in the study was able to discriminate the letters with at least a 90% accuracy-whether or not the background was blurred. The experimental test conditions consisted of presenting a letter to one eye, while the background-the view of the color cartoon-was presented to both. Additionally, as described above, a lens was inserted into the viewing path of one eye which blurred its view of the cartoon. This background image was blurred for the eye which received and was required to detect the letter, or for the eye which did not have this discrimination demand; these are designated as the “ipsilateral” and “contralateral” viewing conditions respectively. An additional control measure was taken to establish a base-line of performance accuracy under the demands of the binocular viewing. I-Iere the subject was asked to detect the letter while both eyes were presented with a clear view of the background. This control proved to be useful as a correction for the binocular performance level of the children, who tended to score in the 80% range while the adults were at or about 90%. For the two adult groups, each eye was tested with 50 letters under each of the conditions (including the monocular screening tests), with the letters, task conditions and eye being varied at random for a total of 500 trials lasting for more than an hour. The children were provided with half this number of triah-i.e. 25 letters for each condition-so that the total testing time would be well under an hour. RESULTS
The data were averaged across stimulus trials to derive a measure of accuracy for each subject, expressed as percent correct. Figure 2 shows the means for the three experimental groups, with the two blurred and the nonblurred control measure from the left and right eyes plotted separately. This summary is useful for showing the similarity of the contralateral blur and the nonblur treatments, and for showing that the pattern of effects was comparable for the left and right eyes. Further, from this summary one can see that the overall performance of the children was about 10% iower than that of the adults. This decrement was probably due to the demand for sustained attention, and it is unlikely that it indicates a fu~amental diEerence in their perceptual skill These children were able to detect the letter with 90% accuracy under monocular viewing conditions. The lower baseline performance of the children is not particularly pertinent to the major hypothesis under examination-laterality of biurring-and could cloud interpretation of statistical analysis by adding an irrelevant factor to the mix. In order to keep a focus on the key issues, therefore, we have used the binocular control measure (in which the background was not blurred in either eye) to adjust for the overall differences in performance. Specifically, accuracy on the ipsilateral and contralateral test conditions was divided by the binqcular control measure, which served to equate the overall
RESEARCH
NOTE
1173
performance of the children without introducing a bias with respect to laterality. As there was no indication of differential performance by right and left eyes, these data were averaged. The major experimental question was whether viewing of a blurred image by one eye would suppress perception by that eye, and in particular, reduce its effectiveness in detecting a letter which briefly was shown to it. Mean ipsilateral and contralatera! performance by the three subject groups (using the transformed and condensed data) is plotted in Fig. 3. As conf?rmed by the statistical analysis described below, the side of blurring had a substantial impact on performance of strabismic patients, but had much less effect on the performance of _-iii normal adults and children. All three groups were in the Normal UWdren Strabww Dotie !nts LOO%range when the blur was presented to the eye that was contralateral to the Ietter, and thus the eye which FIGURE 3. General differencesin the performance of the childrenlikely from poor vigilance--was compensated by expressing ipsilateral detected the letter was watching a sharp background and contralateral performance as a function of the control condition image. Further, with normal adults and children, per- where neither eye was blurred. Additionally, for purposes of statistical formance was in the 80-100% range even when the letter evaluation and for the plot shown here, left and right eye scores were combined. The summary values enforce the conclusion that children was presented to the eye which viewed the blurred do not show any interocular suppression, normal adults show mild background. However, under these conditions, detection accuracy by the strabismic patient dropped to a very suppression, and corrected strabismics are severely deficient in their ability to identify a letter that is presented to the ipsilateral side--to low level-less than a fifth of the accuracy they achieved the eye which is viewing a blurred background image. when the background image to that eye was not blurred. The groups and laterality conditions were evaluated using a 2-way repeated measured ANOVA. This analysis of the smaller sample size, or might be a valid difference indicated a significant difference among the subject in perceptual skill given that the ipsilatera!/contralateral difference is less than 10% for children and is more than groups (&,O = 67.2, P < O.OOl), a significant difference as a function of blurring laterality (F,.30= 247.8, twice that amount for adults. P -cO.OOl), and a significant interaction between these The source of treatment effects (F2.30= 93.3, P < 0.001). DISCUSSION the effects was evaluated using a Dunnett T3 procedure, The key issues in the present experiment were whether which divided alpha according to the number of post hoc tests being performed. Counting the tests described blur can control and determine the use of an eye, and subsequently, a total of eight post hoc tests were calcu- whether that influence differs for normal adults, normal lated, with decision criteria set a P < 0.05 across the children, and corrected strabismics with alternating entire set of comparisons. Given this stringent standard fixation. A refractive error was simulated by means of an one would be unlikely to make more than one error in apparatus which presents the subject with an uncontwenty experiments in which all these comparisons were stricted view of a binocular scene-an animated cartoon repeated. displayed on a color television monitor-in which the Comparison of the subject groups with respect to view of one eye can be blurred with a lens while leaving their contralateral performance did not show any the image to the other eye unaffected. For convenience significant differences. However, when the groups were we describe the cartoon display as a “background” compared on their ipsilatera! performance, the differ- image, by which we mean that it provides a spatial as ences suggested in Fig. 3 were confirmed by the post hoc well as temporal context for the briefly presented letters analysis. The strabismie patients were significantly differ- which probe the subject’s use of each eye. ent from normal adults and from chldren in their ability The subject’s access to information presented to the to detect the letter when the blur was provided to the eye-and more specifically, to the fovea-was evaluated ipsilateral eye. The children did not differ from normal by briefly presenting a sharply focused letter and asking adults in this regard. that the subject identity it. The expectation was that an Additional tests of the children and normal adultseye which had been watching a sharply focused backwere calculated to determine if there was any difference ground would be able to detect the letter with high between their ipsilateral and contralateral performaccuracy, and this was the case for a!! subject groups ance. The null hypothesis could not be rejected for the whether or not the image to the contralateral eye was children, but it was rejected for the adult subjects. It is blurred. However, blurring the background in the eye clear, therefore, that normal adults were impaired in which was required to identify the letter did affect the their ability to detect a briefly exposed letter when that performance of two out of the three groups. Under these eye was also being shown a blurred background. The ipsilateral blur conditions, adults with normal vision lack of significance for the children might be a function showed a mild but significant impairment in accuracy,
1174
RESEARCH NOTE
and strabismic patients were extremely impaired. By contrast, children showed no deficit in accuracy for identifying the letters whether the blur was presented to the ipsilateral or to the contralateral eye. It is often suggested that clinical suppression is derived from normal binocular suppre~ion (e.g. Fahle, 1983). The clinical suppression of strabismic patients is generally assumed to be adaptive, in that it eliminates the confusion of different image elements falling on corresponding retinal locations. Thomas (1978) argues that amblyopia may be caused by lack of image quality in the deviated eye, as the dominant eye will determine the plane of fixation. By default, the images which fall upon the fovea of the deviated eye will likely be out of focus. When the strabismus is not corrected early in childhood, it will generally eliminate the coordinated use of eyes for purposes of stereopsis, and often the dominance of one eye will lead to permanent clinical suppressioni.e. amblyopia. We included children in the present study to see whether they were especially sensitive to the effects of blur. We thought it possible that their ability to detect the well-focused letter would be affected by whether the eye was viewing a defocused background image, causing them to show a greater than normal performance penalty when using that eye. To the contrary, the children showed no decrement-they detected the letter as well when it was presented to the eye which had been watching the blur as to the one which was watching a sharp background image. This may reflect enhanced sensitivity to onset transients or to some other-as yet unspecified-kind of attentional flexibility. It does not support the view that children are abnormally affected by image quality with respect to the control of attention to a given eye. The corrected strabismic patients, on the other hand, were almost totally unable to attend to info~ation which was presented to the eye that was viewing a degraded image. Based on the work of Sireteanu (1982) we have reason to believe that the quality of monocular vision is normal in such patients. This investigator evaluated acuity thresholds as a function of retinal and found symmetric-i.e. normaleccentricity, psychophysical curves for strabismic patients who were classified as having aftemating fixation. Hamasaki and Flynn (1981) say that there is a lower speed of response to a stimulus presented to an amblyopic eye of a strabismic patient, but no difference between the eyes of controls or of strabismic patients who do not suffer from amblyopia. There is firm evidence, however, that such patients will suppress under conditions in which normal. observers do not. Holopigian, Blake and Greenwald (1988) evaluated suppression by presenting different spatial frequency gratings to the two eyes. Due to the effects of binocular rivalry, normals saw one of the gratings and not the other when the orientation of the gratings was orthogonal-this result also being reported by strabismics with alternating fixation. However, the patients also supp~ss~ one eye even when the gratings to the two eyes were of identical orientation and frequency (the
stimuli being distinguishable on the basis of sma]] identifying icons). These investigators also timed the speed of eye alternation under conditions similar to ours, in that the subject was asked to identify a stimulus which was briefly shown. They found that most of these individuals averaged 1000 msec or longer of viewing time for a given eye, and thus had a low accuracy in reporting any stimulus which was flashed to the nonattended eye. The subjects with the fastest rate of alternation averaged 500 msec for each viewing state. These results indicate that switching time between the two monocular pools is relatively slow, and is unaffected by the stimulus onset transients which can trigger attentional shift in normals. For the patients in which strabismus was corrected in late childhood. it is generally said that these individuals do not have binocular vision-meaning that stereopsis has been lost. It is implicit, furthermore, that form vision is accomplished by monocular systems, those which serve one eye when the use of the other is absent or weak, or by switching back and forth between the monocular systems of each eye for subjects such as our own who have alternating fixation. Are these the pr~on~ious monocular systems that are used by normal subjects? We think not, as physiological studies in animals suggest that conditions which work to the detriment of stereopsis result in changes of connection by cells which were initially binocular-rendering them monocular in their responsiveness. In other words, we think it likely that the cell population whose activity provides for conscious percept in the normal viewer provides that experience for the corrected strabismic as we&--the difference being that the option for influence by either eye has been lost. Binocular responsive cells in the normal viewer have become monocular responsive cells in the late-corrected strabismic, with some of the options for stimulus cross-influen~ being greatly diminished or lost. In particular, from the results of the present study, we would contend that the normal binocuiar system allows for abrupt onset cues to disengage suppression which may be in place. We suggest that the binocular population has the capacity to respond selectively to monocular input which contains de~nitive images, and to suppress monocular inputs which are blurred. However, the onset transient which is produced by the presentation of a sharp image can destabilize that suppression, allowing the image to be registered with the binocular population. Thus we view the suppression as being conditional-it precludes influence by the defocused background image, but may not completely preclude other stimulus attributes such as the rapid onset of a sharpIy defined letter. Relative to children, normal adults showed greater suppression. One might speculate that this difference is due to maturation of rivalry mechanisms. In adults the presentation of a sharp background image to one eye and a blurred image to the other might induce rivalry, resulting in control of vision by the eye which is shown the clear image, Under these conditions~ the central area which must detect the target woukl be inhibited, and the
RESEARCH
quick flash of the letter would have to overcome this inhibition in order to be identified. Children, however, may be developmentally immature with respect to rivalry (Cobb, Morton & Ettlinger, 1967; Lawwill 8c Biersdorf, 1968; Wright, Ary, Shors & Erikson, 1986), and thus may synthesize the disparate images in a way which does not
reduce
their
attention
to the fovea1
zone.
REFERENCE Cobb, W. A., Morton, H. B. & Ettlinger, G. (1967). Cerebra) potentials evoked by pattern reversal and their suppression in visual rivalry. Nature, 216. 1123-I 125.
Fahle, M. (1983). Non-fusable stimuli and role of binocular inhibition in normal and pathological vision, especially strabismus. Documenfa
NOTE
1175
Holopigian, K., Blake, R. & Greenwald, M. 1. (1988). Clinical suppression and amblyopia. investigative ~phthaImofo~y and Visual Science, 29, 444-451.
Lawwill, T. & Biersdorf, W. R. (1968). Bionocular rivalry and visual
evoked responses.investigative Ophthalmology, 7, 378-385. Moses,R. A. (1981). Adler’s physiology of the eye. New York: Mosby. Parks. M. M. (1984). Amblyopic ocular motility and strabismus. In Duane, T. (Ed.), Clinical ophthalmology (Vol. 1, Chap. 14). Hagerstown, Md: Harper & Row. Sireteanu, R. (1982). Binocular vision in strabismic humans with alternating fixation. Vision Research, 22, 889-896. Thomas, J. (1978). Normal and amblyopic contrast sensitivity functions in central and peripheral retinas. Inuestigatioe ~phtha~moIogy and Visual Science,
17, 746-753.
Wright, IL W., Ary, J. P.. Shors, T. J. & Eriksen, K. J. (1986). Suppression and the pattern visual evoked potential. Journal of Pediatric Ophthalmology & Strabismus 23, 252-251.
Ophthalmologica, 55, 323-340.
Hamasaki, D. 1. & Flynn, J. T. (1981). Amblyopic eyes have longer reaction times. lnuestigative Ophthalmology and Visaal Science, 21, 846-853.
Ack~owledge~enf-This psychology Foundation.
study was supported in part by the Neuro-
Vision Res. Vol. 32, No. 6. pp. 1169-1175, 1992 Printed in Great Britain. All rights reserved
Research Note Control of Interocular Suppression as a Function of Differential Image Blur TRACEY
J. SHORS,*$
KENNETH
WRIGHT,?
ERNEST
GREENE*
Received 18 June 1990; in revised form 18 July 1991
This experiment studied normal adults, normal children, and corrected strabismics with alternating fixation, evaluating their ability to detect information presented to one eye under viewing conditions which might affect their eye preference. An apparatus was used which allowed unconstrained, binocular viewing of a video monitor, but in which the image seen by one eye was blurred while that seen by the other eye was in sharp focus. Under these viewing conditions, subjects were asked to identify a sharply focused letter which was shown briefly to the eye which was seeing the blurred background image or to the one which was watching the well-focused image. The question was whether there would be general suppression of the eye that was shown a degraded background, causing a reduction in its ability to process other information being presented. The results indicate no performance penalty for children, a mild decrement in recognition accuracy for normal adults, and a severe deficit in the ability of corrected strabismics to attend to the eye which views a blurred image. These results suggest that the loss of binocular vision that accompanies childhood strabismus is not a result of an inherent sensitivity to blurring, insofar as such vulnerability is reflected by these acute testing conditions. Strabismus Interocular suppression Binocular rivalry Selective attention
Visual cortical suppression is the unconscious and involuntary inhibition of part or all of the visual field from one eye under binocular viewing conditions (Moses, 1981). The fundamental observation is that patients with strabismus and unilateral anisometropic ambliopia are unaware of visual information being presented to the nonfixating eye. The term “suppression” has been broadly applied to various manifestations of this effect, and when chronically present in childhood, is thought to progress into a total dependence on one eye-i.e. amblyopia-and/or loss of binocular vision. A blurred image to one eye and stabismus are the critical factors in the development of clinical suppression. In children, a differential refractive error from anisometropia causes monofixation by the eye with the sharper focus and suppression of the central visual field of the blurred eye. Thomas (1978) and Parks (1984) have suggested that these suppression mechanisms apply also for the patient with uncorrected stabismus even if each eye has 20120 vision. Where the two eyes are deviated, different images fall on corresponding retinal
*Department of Psychology, University of Southern California, Los Angeles, CA 90089, U.S.A. tDepartment of Ophthalmology, University of Southern California, Los Angeles, CA 90089, U.S.A. fPresent address: Department of Psychology, Princeton University, Princeton, NJ 08540, U.S.A.
Eye preference
locations and thus the objects which are presented to each eye will overlap-causing visual confusion. Thomas (1978) hypothesizes that the dominant eye selects the focal distance for both eyes, which makes it likely that the objects registered by the subordinate eye will be out of focus. By this analysis, it is the differential in focus between the two eyes-not rivalry mechanisms-which provides the substrate for clinical suppression. Thus an initial small difference in eye preference may progress into an irreversible amblyopia, or where the two eyes are well balanced, an inability to synthesize the two retinal images even though the deviation is surgically corrected. We evaluated the effects of background blur on visual control of normal adults, normal children, and patients that were “corrected strabismics with altemating fixation”, meaning that they previously had been surgically treated to correct strabismus, had adequate vision with either eye, and when tested could switch from one eye to the other depending on stimulus conditions. It is important to note that for these patients the surgical correction, which was done after the critical period for binocular development, remedied motor coordination of the eyes but came too late to prevent the loss of binocular vision. The present apparatus and methods were essential for testing such patients, since other devices-such as Polaroid glasses or haploscopic devices-generally induce dissociation of the visual system and break
1169
1170
RESEARCH
TABLE 1. Visual
Patient
Age at testing (Yr)
Age at surgery W
t
10
9
2 3 4 5 6 7
9 19 35 13 19 28
8 19 7 13 19 28
NOTE
~~~r~ct~~stjcs of clinical subjects Acuity (right eye)
Acuity (Ieft eye)
worth
20120 20/20 20130 20!20 xi20 2Oj2S 20125
2oj20 20/20 2OJ20 20120 20120 2OJ25 20125
suppressed suppmsed Suppressed Suppressed
4-dot
S%lppRSSed
SuQprMed SupQriWd
TitttlUS None None None None None None None
down binocular fusion. This confounds the issue of eye coor~nation. Age and visual char~te~s~~ of whether selective attention to a given eye is dete~in~ the patients are reflected in Table 1. Ocular deviations by the quality of its image or diptopia that has been for each patient were measured using standard strabisinduced by the apparatus. The present apparatus allows mus techniques, including: (a) cove&-uncover testing; undissociated binocular vision and also provides the (b) simultaneous prism cover testing; and (c) alternate potential to blur the background while simultaneously cover testing. These tests detect tropia, tropia with presenting a clear and focused target stimulus to that peripheral fusion and the full dissociated deviationeye. “pboria”. Under the demands of testing, I patient We ~va~~t~ the ability of these subjects to report showed slight esotropia with dissociated vertical deviletters that were presented in clear focus to a given eye ation. The other patients showed extremely good eye while blurring the ~ck~o~~d that was being shown coordination under ali test conditions. to that eye. This was accomplished by means of an apparatus containing three computer/video monitors, Apparatus and stimuli beam-splitters and a + 10 D lens which could be placed The general configuration of equipment for control in the viewing path of either the left or right eye. Thus of images to the two eyes is shown in Fig. f . A color the subject was provided with relatively fry-dewing television monitor, bordered by a white cire%ewith conditions in which they could watch an animated bisecting lines, was placed 0.5 m in front of the subcartoon with both eyes, but with the image provided to ject, whose head position was stabilized by the use one of the eyes being blurred while the other was sharp. of a chin rest. At this distance the image from the Further, by means of the lateral monitors and beam monitor subtended 6 deg of visual angle, A color cartoon splitters, sharply focused letters could be flashed to either was displayed on the monitor (from a video cassette eye, which the subjects were required to identify and recorder), which had a mean luminance of 75 cd/m2, and report, The key question was whether there would be with contrast set at 65% of the range of the instrument. a d~e~ntial in reporting accuracy as a function of the The cartoon image served as a ~ck~ound field for the focus of the background image-i.e. whether the eye detection task, and for the expe~menta~ control of eye being shown the bluned image would be refractory to preference as described below. A beam splitting mirror was placed in the path of these letters even though they were being presented in vision for each eye, which allowed images from. two sharp focus to the fovea. additional black-and-white monitors to be superimposed upon the subject’s view of the cartoon, and at tbe proper focus. The only mate~al shown on these mon~hrome mo~tors were white capital Mters that were presented for a 16 msec interval, one at a There were three groups of subjects, consisting of time, and whiih were always positi&ed at the center of 20 normal-vision adults (ages 23-4S), 6 normal-vision the screen. The remainder of the screen, as well as the children (age 6ll), and 7 patients (ages 9-35) who monitor itself and surrounding equipment were not had been surgicaliy corrected for strabismus. All subjects ill~inated, By this &on~guration, the letter stimuli were given a complete opht~rno~o~~~ examination, from each monochrome monitor cotdd be superwhich included the Snellen test for visual acuity and tests imposed upon the subject’s view of the color mo~tor, for stereoscopic fusion using the Worth 4 Dot and the and could be presented to either -the left or the right Titmus. In mkr to qualify for one of the normal-vision eye (depending on which mon~~r~rne monitor was groups, the sub&x$had to demonstrate acuity of 20120 used to dehver the stimulus). Thus the composite image (with corm&ion) for each eye, to fuse the elements of that was seen by each eye cot&ted of its view of the Worth 4 Dot at near and far distances, and to show the cartoon being displayed, and a white capital letter stereopsis of at least 40 set of arc on the Titmus. None which would appear periodically, For 16 msec, in the center of the screen. of the normal subjects showed ocular deviation. A major e~~~rn~n~ marination was made The s~b~s~~ patients were s&x%& on the basis of high acuity in each eye (with correction) and good possible by adding a + 10 D lens in the viewing path for
RESEARCH
as shown in Fig. I. The fens WAS placed between the beam splitting mirror and the color monitor, and served to degrade the retina1 image of the cartoon while leaving the image from the monochrome monitori.e. the capital letter-in sharp focus. The control condition was provided by biurring the background image for the opposite eye. As detailed below, the issue is not whether one can discriminate sharply focused letters against a blurred background, but whether the background is able to bias the use of that eye and thus work at odds with the demands of the discrimination task. In order to reduce the complexity of the display system, only symmetrical letters-i.e. A, H, M, T, V, W, X and Y-were used as the test stimuli. As seen by the subject, the vertical dimension of the letters subtended 47 min of visual angle. At1 sampling from the letter pool, selection of side for the display (left or right), as well as interstimulus interval was determined at random. The interval between successive presentations ranged from 2 to 4 sec.
NOTE
1171
an eye,
The subject was seated with his head in the chin rest, and the apparatus was aligned with the aid of stimulus markers until the monitors were properly centered for that individual. The beamsplitting mirrors were adjusted to align the image from the side monitors into the center of his visual field, and cover~uncover tests were used to verify the alignment. The subject was then instructed to concentrate (to focus) on the’cartoon movie being shown on the color monitor, but also was told to report the letters that would briefly appear. The subject was then given a trial run of 10 letters to each eye to familiarize him with the procedure, and to be certain that he could reliably report letters which were shown for 16 msec. A sequence of test trials was administered which included monocular and binocular viewing conditions. The monocular trials were used to screen subjects and reject any subjects who could not accurately detect letters when all of the stimulus material (letters as well as background) was presented to each eye alone.
FIGURE 1. A experimental configuration was designed to allow one eye to have a cleat, well-focused view of a central monitor while the image to the other eye was blurred by means of a IO D lens. Additionally, beam-splitters were placed in the path of view which allowed letters to be gashed briefly on a lateral monitor, thus providing a well-focused letter to the fovea of one eye. The task requires that the subject identify and name the letter which was shown. The upper panel illustrates the letter being presented to the eye which is viewing a blurred background image-this being designated the “ipsilateral eye” condition, For the “contralateral eye” condition the blur is presented to the opposite eye, and the letter must be identified by the eye which has a dear view of the background cartoon. The figure illustrates the presentation of letters to the left eye-with reiations and terminology reversing when they are shown to the right. (Note that positioning of elements is for the purposeof illustration, and distances are not to scale.)
1172
RESEARCH NOTE
Under these monocular test conditions, every subject that was included in the study was able to discriminate the letters with at least a 90% accuracy-whether or not the background was blurred. The experimental test conditions consisted of presenting a letter to one eye, while the background-the view of the color cartoon-was presented to both. Additionally, as described above, a lens was inserted into the viewing path of one eye which blurred its view of the cartoon. This background image was blurred for the eye which received and was required to detect the letter, or for the eye which did not have this discrimination demand; these are designated as the “ipsilateral” and “contralateral” viewing conditions respectively. An additional control measure was taken to establish a base-line of performance accuracy under the demands of the binocular viewing. I-Iere the subject was asked to detect the letter while both eyes were presented with a clear view of the background. This control proved to be useful as a correction for the binocular performance level of the children, who tended to score in the 80% range while the adults were at or about 90%. For the two adult groups, each eye was tested with 50 letters under each of the conditions (including the monocular screening tests), with the letters, task conditions and eye being varied at random for a total of 500 trials lasting for more than an hour. The children were provided with half this number of triah-i.e. 25 letters for each condition-so that the total testing time would be well under an hour. RESULTS
The data were averaged across stimulus trials to derive a measure of accuracy for each subject, expressed as percent correct. Figure 2 shows the means for the three experimental groups, with the two blurred and the nonblurred control measure from the left and right eyes plotted separately. This summary is useful for showing the similarity of the contralateral blur and the nonblur treatments, and for showing that the pattern of effects was comparable for the left and right eyes. Further, from this summary one can see that the overall performance of the children was about 10% iower than that of the adults. This decrement was probably due to the demand for sustained attention, and it is unlikely that it indicates a fu~amental diEerence in their perceptual skill These children were able to detect the letter with 90% accuracy under monocular viewing conditions. The lower baseline performance of the children is not particularly pertinent to the major hypothesis under examination-laterality of biurring-and could cloud interpretation of statistical analysis by adding an irrelevant factor to the mix. In order to keep a focus on the key issues, therefore, we have used the binocular control measure (in which the background was not blurred in either eye) to adjust for the overall differences in performance. Specifically, accuracy on the ipsilateral and contralateral test conditions was divided by the binqcular control measure, which served to equate the overall
RESEARCH
NOTE
1173
performance of the children without introducing a bias with respect to laterality. As there was no indication of differential performance by right and left eyes, these data were averaged. The major experimental question was whether viewing of a blurred image by one eye would suppress perception by that eye, and in particular, reduce its effectiveness in detecting a letter which briefly was shown to it. Mean ipsilateral and contralatera! performance by the three subject groups (using the transformed and condensed data) is plotted in Fig. 3. As conf?rmed by the statistical analysis described below, the side of blurring had a substantial impact on performance of strabismic patients, but had much less effect on the performance of _-iii normal adults and children. All three groups were in the Normal UWdren Strabww Dotie !nts LOO%range when the blur was presented to the eye that was contralateral to the Ietter, and thus the eye which FIGURE 3. General differencesin the performance of the childrenlikely from poor vigilance--was compensated by expressing ipsilateral detected the letter was watching a sharp background and contralateral performance as a function of the control condition image. Further, with normal adults and children, per- where neither eye was blurred. Additionally, for purposes of statistical formance was in the 80-100% range even when the letter evaluation and for the plot shown here, left and right eye scores were combined. The summary values enforce the conclusion that children was presented to the eye which viewed the blurred do not show any interocular suppression, normal adults show mild background. However, under these conditions, detection accuracy by the strabismic patient dropped to a very suppression, and corrected strabismics are severely deficient in their ability to identify a letter that is presented to the ipsilateral side--to low level-less than a fifth of the accuracy they achieved the eye which is viewing a blurred background image. when the background image to that eye was not blurred. The groups and laterality conditions were evaluated using a 2-way repeated measured ANOVA. This analysis of the smaller sample size, or might be a valid difference indicated a significant difference among the subject in perceptual skill given that the ipsilatera!/contralateral difference is less than 10% for children and is more than groups (&,O = 67.2, P < O.OOl), a significant difference as a function of blurring laterality (F,.30= 247.8, twice that amount for adults. P -cO.OOl), and a significant interaction between these The source of treatment effects (F2.30= 93.3, P < 0.001). DISCUSSION the effects was evaluated using a Dunnett T3 procedure, The key issues in the present experiment were whether which divided alpha according to the number of post hoc tests being performed. Counting the tests described blur can control and determine the use of an eye, and subsequently, a total of eight post hoc tests were calcu- whether that influence differs for normal adults, normal lated, with decision criteria set a P < 0.05 across the children, and corrected strabismics with alternating entire set of comparisons. Given this stringent standard fixation. A refractive error was simulated by means of an one would be unlikely to make more than one error in apparatus which presents the subject with an uncontwenty experiments in which all these comparisons were stricted view of a binocular scene-an animated cartoon repeated. displayed on a color television monitor-in which the Comparison of the subject groups with respect to view of one eye can be blurred with a lens while leaving their contralateral performance did not show any the image to the other eye unaffected. For convenience significant differences. However, when the groups were we describe the cartoon display as a “background” compared on their ipsilatera! performance, the differ- image, by which we mean that it provides a spatial as ences suggested in Fig. 3 were confirmed by the post hoc well as temporal context for the briefly presented letters analysis. The strabismie patients were significantly differ- which probe the subject’s use of each eye. ent from normal adults and from chldren in their ability The subject’s access to information presented to the to detect the letter when the blur was provided to the eye-and more specifically, to the fovea-was evaluated ipsilateral eye. The children did not differ from normal by briefly presenting a sharply focused letter and asking adults in this regard. that the subject identity it. The expectation was that an Additional tests of the children and normal adultseye which had been watching a sharply focused backwere calculated to determine if there was any difference ground would be able to detect the letter with high between their ipsilateral and contralateral performaccuracy, and this was the case for a!! subject groups ance. The null hypothesis could not be rejected for the whether or not the image to the contralateral eye was children, but it was rejected for the adult subjects. It is blurred. However, blurring the background in the eye clear, therefore, that normal adults were impaired in which was required to identify the letter did affect the their ability to detect a briefly exposed letter when that performance of two out of the three groups. Under these eye was also being shown a blurred background. The ipsilateral blur conditions, adults with normal vision lack of significance for the children might be a function showed a mild but significant impairment in accuracy,
1174
RESEARCH NOTE
and strabismic patients were extremely impaired. By contrast, children showed no deficit in accuracy for identifying the letters whether the blur was presented to the ipsilateral or to the contralateral eye. It is often suggested that clinical suppression is derived from normal binocular suppre~ion (e.g. Fahle, 1983). The clinical suppression of strabismic patients is generally assumed to be adaptive, in that it eliminates the confusion of different image elements falling on corresponding retinal locations. Thomas (1978) argues that amblyopia may be caused by lack of image quality in the deviated eye, as the dominant eye will determine the plane of fixation. By default, the images which fall upon the fovea of the deviated eye will likely be out of focus. When the strabismus is not corrected early in childhood, it will generally eliminate the coordinated use of eyes for purposes of stereopsis, and often the dominance of one eye will lead to permanent clinical suppressioni.e. amblyopia. We included children in the present study to see whether they were especially sensitive to the effects of blur. We thought it possible that their ability to detect the well-focused letter would be affected by whether the eye was viewing a defocused background image, causing them to show a greater than normal performance penalty when using that eye. To the contrary, the children showed no decrement-they detected the letter as well when it was presented to the eye which had been watching the blur as to the one which was watching a sharp background image. This may reflect enhanced sensitivity to onset transients or to some other-as yet unspecified-kind of attentional flexibility. It does not support the view that children are abnormally affected by image quality with respect to the control of attention to a given eye. The corrected strabismic patients, on the other hand, were almost totally unable to attend to info~ation which was presented to the eye that was viewing a degraded image. Based on the work of Sireteanu (1982) we have reason to believe that the quality of monocular vision is normal in such patients. This investigator evaluated acuity thresholds as a function of retinal and found symmetric-i.e. normaleccentricity, psychophysical curves for strabismic patients who were classified as having aftemating fixation. Hamasaki and Flynn (1981) say that there is a lower speed of response to a stimulus presented to an amblyopic eye of a strabismic patient, but no difference between the eyes of controls or of strabismic patients who do not suffer from amblyopia. There is firm evidence, however, that such patients will suppress under conditions in which normal. observers do not. Holopigian, Blake and Greenwald (1988) evaluated suppression by presenting different spatial frequency gratings to the two eyes. Due to the effects of binocular rivalry, normals saw one of the gratings and not the other when the orientation of the gratings was orthogonal-this result also being reported by strabismics with alternating fixation. However, the patients also supp~ss~ one eye even when the gratings to the two eyes were of identical orientation and frequency (the
stimuli being distinguishable on the basis of sma]] identifying icons). These investigators also timed the speed of eye alternation under conditions similar to ours, in that the subject was asked to identify a stimulus which was briefly shown. They found that most of these individuals averaged 1000 msec or longer of viewing time for a given eye, and thus had a low accuracy in reporting any stimulus which was flashed to the nonattended eye. The subjects with the fastest rate of alternation averaged 500 msec for each viewing state. These results indicate that switching time between the two monocular pools is relatively slow, and is unaffected by the stimulus onset transients which can trigger attentional shift in normals. For the patients in which strabismus was corrected in late childhood. it is generally said that these individuals do not have binocular vision-meaning that stereopsis has been lost. It is implicit, furthermore, that form vision is accomplished by monocular systems, those which serve one eye when the use of the other is absent or weak, or by switching back and forth between the monocular systems of each eye for subjects such as our own who have alternating fixation. Are these the pr~on~ious monocular systems that are used by normal subjects? We think not, as physiological studies in animals suggest that conditions which work to the detriment of stereopsis result in changes of connection by cells which were initially binocular-rendering them monocular in their responsiveness. In other words, we think it likely that the cell population whose activity provides for conscious percept in the normal viewer provides that experience for the corrected strabismic as we&--the difference being that the option for influence by either eye has been lost. Binocular responsive cells in the normal viewer have become monocular responsive cells in the late-corrected strabismic, with some of the options for stimulus cross-influen~ being greatly diminished or lost. In particular, from the results of the present study, we would contend that the normal binocuiar system allows for abrupt onset cues to disengage suppression which may be in place. We suggest that the binocular population has the capacity to respond selectively to monocular input which contains de~nitive images, and to suppress monocular inputs which are blurred. However, the onset transient which is produced by the presentation of a sharp image can destabilize that suppression, allowing the image to be registered with the binocular population. Thus we view the suppression as being conditional-it precludes influence by the defocused background image, but may not completely preclude other stimulus attributes such as the rapid onset of a sharpIy defined letter. Relative to children, normal adults showed greater suppression. One might speculate that this difference is due to maturation of rivalry mechanisms. In adults the presentation of a sharp background image to one eye and a blurred image to the other might induce rivalry, resulting in control of vision by the eye which is shown the clear image, Under these conditions~ the central area which must detect the target woukl be inhibited, and the
RESEARCH
quick flash of the letter would have to overcome this inhibition in order to be identified. Children, however, may be developmentally immature with respect to rivalry (Cobb, Morton & Ettlinger, 1967; Lawwill 8c Biersdorf, 1968; Wright, Ary, Shors & Erikson, 1986), and thus may synthesize the disparate images in a way which does not
reduce
their
attention
to the fovea1
zone.
REFERENCE Cobb, W. A., Morton, H. B. & Ettlinger, G. (1967). Cerebra) potentials evoked by pattern reversal and their suppression in visual rivalry. Nature, 216. 1123-I 125.
Fahle, M. (1983). Non-fusable stimuli and role of binocular inhibition in normal and pathological vision, especially strabismus. Documenfa
NOTE
1175
Holopigian, K., Blake, R. & Greenwald, M. 1. (1988). Clinical suppression and amblyopia. investigative ~phthaImofo~y and Visual Science, 29, 444-451.
Lawwill, T. & Biersdorf, W. R. (1968). Bionocular rivalry and visual
evoked responses.investigative Ophthalmology, 7, 378-385. Moses,R. A. (1981). Adler’s physiology of the eye. New York: Mosby. Parks. M. M. (1984). Amblyopic ocular motility and strabismus. In Duane, T. (Ed.), Clinical ophthalmology (Vol. 1, Chap. 14). Hagerstown, Md: Harper & Row. Sireteanu, R. (1982). Binocular vision in strabismic humans with alternating fixation. Vision Research, 22, 889-896. Thomas, J. (1978). Normal and amblyopic contrast sensitivity functions in central and peripheral retinas. Inuestigatioe ~phtha~moIogy and Visual Science,
17, 746-753.
Wright, IL W., Ary, J. P.. Shors, T. J. & Eriksen, K. J. (1986). Suppression and the pattern visual evoked potential. Journal of Pediatric Ophthalmology & Strabismus 23, 252-251.
Ophthalmologica, 55, 323-340.
Hamasaki, D. 1. & Flynn, J. T. (1981). Amblyopic eyes have longer reaction times. lnuestigative Ophthalmology and Visaal Science, 21, 846-853.
Ack~owledge~enf-This psychology Foundation.
study was supported in part by the Neuro-
