Clinical Suppression and Amblyopia

Karen Holopigian,*^: Randolph Blake,* and Mark J. Greenwaldf

In individuals with abnormal binocular vision, such as strabismics and anisometropes, it is common for all or part of one eye's view to be suppressed so binocular confusion and diplopia are eliminated. We examined the relation between the depth of suppression (the amount by which the monocular contrast increment threshold for an eye was elevated by stimulation in the contralateral eye) and the degree of amblyopia (difference in monocular contrast thresholds for the two eyes). There was a significant negative correlation between suppression and amblyopia, so that clinical suppressors with no amblyopia exhibited deep suppression (ie, large threshold elevation) while observers with amblyopia exhibited weaker or no suppression. This negative correlation was found when the two eyes viewed orthogonally oriented contours as well as identically oriented contours. These results suggest that when an eye is amblyopic there is no longer a need for strong suppression of that eye by the contralateral eye. Invest Ophthalmol Vis Sci 29:444-451,1988

Individuals with abnormal binocular vision, such in response to conflicting monocular visual input to as strabismics and anisometropes, often suppress part the two eyes, the relationship between these entities of one eye's view. This phenomenon may be either remains unclear. There is some speculation that unilateral, such that one eye is chronically sup- long-term chronic suppression is actually responsible pressed, or bilateral, with dominance and suppression for the development of amblyopia in one eye. Indeed, alternating between the two eyes. It is generally as- Sireteanu and Fronius13 and Sireteanu14 found that sumed1 that this clinical suppression is adaptive in portions of the visual field that exhibited deeper in- that it eliminates confusion (resulting from different terocular suppression (assessed with a luminance de- images falling on corresponding retinal locations) tection task) were also more amblyopic (ie, had and diplopia (arising from the left and right eye poorer monocular acuity); areas less strongly sup- images falling on noncorresponding retinal loca- pressed had better acuity. The view that suppression tions). causes amblyopia cannot be entirely correct, how- Another condition commonly associated with stra- ever, for it is known that many individuals with clini- bismus and anisometropia is amblyopia, a chronic cal suppression have equal visual acuity in the two reduction in monocular vision. Amblyopia may im- eyes. pair visual performance as measured by Snellen and To clarify the relationship between clinical sup- grating acuity,2 contrast sensitivity,3"6 vernier acuity,7 pression and amblyopia, we examined the correlation contrast matching8 and stereoacuity.9"12 between the degree of amblyopia and the depth of Although amblyopia and suppression both develop suppression in a group of strabismic and anisometropic observers. In these experiments, the depth of From the "Departments of Psychology and Neurobiology/Physi- suppression was defined as the amount by which the ology, Northwestern University, Evanston, Illinois, and the fDe- contrast increment threshold for an eye was raised by partment of Ophthalmology, Northwestern University Medical the simultaneous presentation of a stimulus to the School, Chicago, Illinois. J Present address: Department of Ophthalmology, N.Y.U. Med- contralateral eye. Amblyopia was indexed by the ical Center, New York, New York. magnitude of the difference between right and left eye This research was conducted as a partial requirement for the contrast thresholds. To our surprise, we found that PhD degree for KH. Portions of this work were presented at the the degree of amblyopia and the depth of suppression 1987 ARVO meeting, May 4-8, Sarasota, Florida. were inversely related. Clinical suppressors with Supported by NSF grant BNS 8418731 to RB, by NSF grant NS07223-05 to the Neurobiology and Physiology Department at equal monocular vision showed large amounts of in- Northwestern University, a Northwestern University dissertation terocular suppression, while those with amblyopia ex- year grant to KH and a grant from Children's Memorial Hospital. hibited a much smaller suppression effect. KH is currently supported by a grant from the RP Fighting Blind- ness to the Retina Clinic at NYU Medical Center. Materials and Methods Submitted for publication: June 4, 1987; accepted September 24, Observers 1987. Reprint requests: Department of Ophthalmology, N.Y.U. Medi- Nine individuals with clinical suppression, soli- cal Center, 550 First Avenue, New York, NY 10016. cited from the student population at Northwestern

444

Downloaded from iovs.arvojournals.org on 09/27/2021 No. 3 SUPPRESSION AND AMDLYOPIA / Holopigian er al. 445

University, were paid an hourly wage to serve in this alignment. For this alignment procedure, which pre- experiment. Informed consent was obtained after the ceded each testing session, precision and care were nature of the procedure had been explained fully. All stressed. observers exhibited a suppression scotoma and were Contrast thresholds: For all observers, monocular strabismic and/or anisometropic. To be classified as a contrast thresholds were assessed for both horizontal clinical suppressor, the observer had to exhibit sup- and vertical sinusoidal gratings of 3.3 c/deg. In the pression with the Bagolini striated glass test. This depth of suppression experiments, described in the method is the least dissociating (and therefore closest next section, it was important to have a large range of to normal viewing conditions) of the standard clinical contrasts available above threshold in order to deter- tests for suppression.15 The test was administered mine the contrast increment threshold during sup- with the room lights on, so items in the room were pression. Therefore, a spatial frequency close to the clearly visible to the two eyes, thus minimizing disso- peak of the contrast sensitivity function was desir- ciation. Although the existence of a suppression sco- able. Since a 3.3 c/deg grating is close to the peak and toma with the Bagolini lenses was the criteria for the provides four complete light and dark cycles across classification of suppression, for purposes of compar- the 1.2 degree field, this spatial frequency was used ison, suppression was also assessed with the Worth 4 for both the contrast threshold and the depth of sup- Dot test. At the viewing distance of 33 cm, each dot pression measurements. subtended 1 degree, and the entire field subtended 6 Contrast thresholds were measured using a two-al- degrees. Both the Bagolini and the Worth 4 Dot tests ternative temporal forced-choice staircase procedure. were administered under normal viewing conditions The staircase estimated the 71% correct detection for these observers. In addition, all observers under- level using a rule which incremented the contrast fol- went a comprehensive ophthalmologic evaluation, lowing each incorrect response and decremented the including refraction, ophthalmoscopy and clinical contrast following each two correct responses. Con- assessment of visual acuity, ocular alignment, stere- trast was initially changed in 3 dB steps, but subse- opsis, binocularity, and binocular and monocular quently was changed in 1 dB steps following two fixation patterns. The observers' visual characteristics staircase reversals. Each staircase was terminated are listed in Table 1. after 12 reversals, and the last five reversals were All psychophysical testing was conducted at North- averaged to yield an estimate of the contrast thresh- western University. For purposes of comparison, we old. Two independent staircases were randomly in- also tested four observers with good visual acuity and terleaved in one experimental run. At least four normal binocular vision. Three of the normal ob- thresholds were averaged for each data point. servers were paid for their participation and were The observer adapted to the prevailing light level completely naive as to the purposes of the experi- for 1 min before the start of each staircase run. The ment. All clinical and normal observers wore their observer triggered each trial, which consisted of two 1 best refractive correction during psychophysical second intervals, denoted by tones. During one ran- testing. domly selected interval, a sinusoidal grating pattern was ramped on for 1 second (rise time and fall time Apparatus and Procedure both equal to 250 msec); during the other interval no The observer stereoscopically viewed two identical grating was presented. The observer's task was to sig- CRT screens situated side by side. The CRT screens nal to the computer which interval contained the were at a viewing distance of 96 cm and had an aver- grating; feedback was provided. age luminance of 30.6 c/m2. For all testing sessions, Contrast thresholds were assessed separately for the the face of each CRT was masked to a circular region two eyes, with presentation order randomized. These 1.2 degrees in diameter. These small fields were used monocular threshold measurements were performed so that suppression measures would be obtained from under two conditions. In one condition the non- a retinal area completely suppressed and so that tested eye viewed a uncontoured raster, and in the wholesale binocular rivalry (ie, not piecemeal) would other condition the non-tested eye was occluded with be experienced by normal observers tested under a black eyepiece. For all control observers and seven comparable conditions. of the nine clinical suppressors, these two testing pro- At the start of the first testing session, each observer cedures yielded equivalent thresholds. In two clinical was taught to adjust the stereoscope mirrors to opti- observers, thresholds were considerably lower when cally align the dichoptically viewed CRTs, using the contralateral eye was occluded (rather than view- nonius lines to gauge alignment. During this process ing the uncontoured raster) and for these two individ- fusion was disrupted using a cover/uncover maneu- uals all subsequent thresholds were measured in this ver to eliminate any fusional maintenance of ocular manner.

Downloaded from iovs.arvojournals.org on 09/27/2021 o

Table 1. Observer characteristics

A. Strabismic suppressors B. Anisometropic suppressors

BG CD AS MM WV PD LP KJ MT z m Age/gender 26/M 20/M 24/F 20/F 28/M 28/M 24/F 24/F 20/F Treatment history: Occlusion at 10 yr Surgery at 3 yr (OD) Surgery at 5 yr (OS), Surgery at 2 yr Occlusion at 8 Surgery (OS) at 1 Occlusion Occlusion at Occlusion from 8- > at 7 yrs (OD) yr 9yr 12 yr < Refraction: OD -1.50 +4.75 -2.75 +3.75 + 2.00 X 100 -2.75 -2.50 -11.00 -0.25 -1.25 m OS -1.50 + 1.50 -3.50 +4.25 + 0.75 X 60 -2.75 -2.75 + 0.50 X 85 -4.00 -3.25 +4.00+ 1.25X90 Snellen acuity: OD 20/30+ 20/30 - 2 20/20- 20/20 20/20+ 20/20- 15/100 20/30- 20/20 o with correction OS 20/20 20/15 20/25 20/20 20/16 20/20 20/20 20/20+ 20/80 Fixation: OD Near central Central Central Central Central Not tested Central Central Not tested > OS Near central Central Central Central Central Not tested Central Central Not tested O Ocular alignment:* Near: RET s 2 PD RET = 8 PD, LHT = 2 PD AET = 10 PD AET = 20 PD, RHT = 3-5 PD RET = 10 PD AET = 20-25 PD Orthot Ortho Ortho 8 4-6 PD with ACT 16 PD with ACT RHT= 5PD ET = 35 PD with ACT RHT = 5 PD Far: RET s2 PD RET = 8 PD AET = 10 PD AET = 20 PD, RHT = 3-5 PD RET= 6PD AET= 15 PD OrthoJ Ortho LET = 2 PD 4-6 PD with ACT LHT = 2 PD RHT= 3PD ET = 40 PD with ACT RHT = 10 PD <*^~•. i^i

Binocular status: > Worth 4 Dot: Scotoma OD Scotoma OD Rapid alternating Alternating suppression Scotoma OD Alternating Scotoma Scotoma OS suppression suppression OD n between OD and m OS r> Bagolini lens: Central scotoma OD Central scotoma OD Rapid alternating Central scotoma OD with Central Alternating Scotoma Scotoma Scotoma OS m suppression peripheral ARC scotoma OD suppression OD OD between OD and q OS n Titmus stereo: 100" 3000" or poorert 140" Poorer than 3000" 3000" Poorer than 3000" 400" 50" 50"

CD * Except as indicated cover test deviation and alternate cover measurements were identical. £ Response to Worth 4 Dot was difficult to interpret. Co t Gave one positive response, one negative response to the Titmus Fly (3000 sec) in separate testing Abbreviations: OD = right eye, OS = left eye, PD = prism diopters, ET = esodeviation, RET = right sessions. esotropia, LET = left esotropia, ACT = alternating esotropia, RHT = right hypertropia, LHT = left X By history intermittently esotropic. Frequently manifested in childhood, currently only with fatigue hypertropia, ACT = alternate cover test, Ortho = orthophoric. or ethanol ingestion.

Downloaded from iovs.arvojournals.org on 09/27/2021 No. 3 SUPPRESSION AND AMDLYOPIA / Holopigian er al. 447

Depth of suppression: Contrast increment thresh- CQ •D olds were determined using a two-alternative spatial \j c forced-choice procedure. For the initial baseline o ai measures, the observer viewed a horizontal, 3.3 c/deg CO 0) sinusoidal grating of 10% contrast with one eye, while w Q. the other eye received no stimulation. The observer a 3 controlled the trial presentation and was told to trig- CO ger trials only when the horizontal grating was clearly and completely visible. Each trial involved the abrupt *- 0 a. MB MW QH KH addition of a contrast increment to either the top or a> a ORTHO the bottom half of the grating, with the incremented -5 half determined randomly. The contrast increment Fig. 1. Histogram showing average depth of suppression in dB for occurred almost immediately (25 msec) following the four normal observers. Depth of suppression is the ratio of the observer's trigger press and lasted 200 msec. When contrast increment threshold for dichoptic viewing conditions rela- the contrast increment was removed, the grating re- tive to the monocular baseline. A suppression effect of 6 dB means that, on average, twice as much contrast increment must be added turned to its original contrast. The contrast incre- to the suppressed stimulus for it to be detected binocularly as when ment was positioned spatially so that its boundary it was viewed monocularly. The conditions shown are: (1) ortho- corresponded to a zero crossing and produced no lu- gonally oriented gratings with the tested eye suppressed (ORTHO); minance edge or change in overall luminance. (2) identically oriented gratings (IDENT); and (3) orthogonally oriented gratings with the tested eye dominant (DOM). The stand- Following each trial, the observer signaled which ing contrast of the gratings was 10% and the spatial frequency was half of the grating received the increment; feedback 3.3 c/deg. Each value is the mean of four to six threshold estimates. was again provided. Over trials, the contrast incre- Standard error bars are shown. ment was varied in 2 and then 1 dB steps, again following a staircase rule which converged on the 71% ings were identified by high contrast markers (black correct level. Two independent staircases were ran- spots 3.6 min of arc) located 21 min above the center domly interleaved during a single run. The monocu- of the circular fieldi n the left eye and 21 min below in lar contrast increment threshold was used as the base- the right eye. Observers were instructed to trigger line for measuring suppression. For the normal ob- trials only when the appropriate marker (ie, the top servers, the right eye was tested for this baseline one) was visible and the other marker was suppressed. measure, and for each clinical suppressor the eye All observers, both control and clinical, were tested which was usually suppressed when viewing with the with both orthogonally and identically oriented grat- Bagolini lenses was tested. ings. When the gratings were orthogonal, all ob- For the suppression measures, gratings were pre- servers experienced suppression. When viewing iden- sented simultaneously to both eyes. When suppres- tically oriented gratings, however, the control ob- sion was measured for orthogonal gratings, (ORTHO servers did not experience suppression—both condition) the tested eye still viewed a horizontal gratings were seen continuously, as noted by the con- grating, but the contralateral eye now viewed a 3.3 tinuous presence of both markers. Since neither grat- c/deg vertical grating of 10% contrast. For this condi- ing was ever suppressed, this provides one test of fu- tion, observers were told to trigger trials when they sion in normal observers. The clinical observers, on saw only the vertical grating, with no trace of the the other hand, reliably suppressed one eye's view. horizontal grating. The contrast increment, however, For the amblyopic suppressors, the marker viewed by was still added to the top or bottom of the horizontal the poorer eye was suppressed, while visibility of the grating, so the staircase now assessed the amount of markers fluctuated between the eyes for the alternat- contrast increment needed for the suppressed eye to ing suppressors. detect the increment. All other aspects of the staircase procedure remained the same. Results Orthogonal gratings are used to measure binocular Our index of the depth of suppression was the ratio rivalry suppression. For clinical suppressors, how- of the monocular baseline increment threshold rela- ever, the visual stimuli for suppression under normal tive to the increment threshold obtained under binoc- viewing conditions are presumed to be identical or ular conditions. It is important to note that the depth nearly identical contours seen by the two eyes.16 We of suppression is not a measure of amblyopia, since therefore also measured the depth of suppression by amblyopia is reflected as overall higher monocular presenting both the eyes with horizontal gratings increment thresholds. (IDENT condition), and added the contrast incre- Figure 1 shows the results obtained for the four ment to the suppressed horizontal grating. The grat- normal observers. When the gratings were identically

Downloaded from iovs.arvojournals.org on 09/27/2021 448 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / March 1988 Vol. 29

these three conditions for normal observers, F (2,6) = 7.2, P < 0.05. A Neuman-Keuls post-hoc test re- C 20 o vealed that the magnitude of suppression for the ORTHO condition was significantly different than the magnitude for the IDENT and DOM conditions r— CO while the latter two did not differ. 1 Figure 2 shows the results for the clinical suppres- ; sors. Two findings are noteworthy. First, the depth of a. suppression varied greatly among these observers, 0) Q ranging from 0 to greater than 20 dB. Second, for any LP CD BG AS MM MT KJ WV PD LP CD BG AS MM MT KJ WV PD one clinical suppressor, the depth of suppression for ORTHO IDENT orthogonal gratings closely corresponded to the depth Fig. 2. Histogram showing average depth of suppression in dB for of suppression for identical gratings. (The same pat- clinical suppressors. Depth of suppression is the ratio of the con- tern of results was obtained when the observers trast increment threshold for dichoptic viewing conditions relative to the monocular baseline. The conditions shown are: (1) ortho- viewed two vertically oriented gratings.) The Pearson gonally oriented gratings with the tested eye suppressed (ORTHO); correlation coefficient for the two measures of sup- and (2) identically oriented gratings with the tested eye suppressed pression was 0.94, and a repeated measure analysis of (IDENT). The standing contrast of the gratings was 10% (20% for variance indicated no significant differences between observer LP) and the spatial frequency was 3.3 c/deg. Each value is the depth of suppression for orthogonally vs. identi- the mean of four to six threshold estimates. Standard error bars are shown. cally oriented gratings, F (1,8) = 0.002, P > 0.05. Why are there such large differences in the magnitude of the suppression effect across clinical ob- oriented (IDENT), there was no suppression effect, servers? One possibility is that the depth of suppres- and thresholds were equivalent to those for the mon- sion is related to the degree of amblyopia. To exam- ocular baseline condition. It is interesting that this ine this possibility we compared depth of suppression condition did not produce the dichoptic masking ob- for both orthogonally and identically oriented grat- served by Legge.17 He found that the threshold for a ings to the degree of amblyopia for these same stim- sinusoidal grating was greatly elevated when a supra- uli. Amblyopia was indexed as the ratio of the con- threshold grating of the same orientation and spatial trast thresholds for the poorer eye to that of the better frequency was simultaneously presented to the con- eye for a 3.3 c/deg grating of the appropriate orienta- tralateral eye. In the Legge experiment, however, tion. both the test and the mask gratings were briefly For both orthogonally and identically oriented flashed for 200 msec, while in the present experiment gratings, the correlation between depth of suppres- the grating viewed by the contralateral eye was sion and degree of amblyopia in the clinical suppres- present continuously and only the contrast increment sors was negative. Observers with severe amblyopia to the ipsilateral eye was flashed. showed almost no suppression, while observers with For the normal observers, the magnitude of the small amounts of amblyopia demonstrated large suppression effect for orthogonal gratings (ORTHO) amounts of suppression. The correlation (Pearson r) ranged from 4.2 dB to 7.1 dB, with an average of 5.6 for these two indices was —0.674 for orthogonal ori- dB. Increment thresholds for the horizontal grating ented gratings and —0.670 for identically oriented were also measured under conditions of dominance gratings. Both of these correlations are significantly (DOM condition), where trials were triggered only different from zero (df =7,P< 0.05). when the horizontal grating was clearly visible and The ratio of Snellen acuities between amblyopic the vertical was suppressed.* (Note that increment and nonamblyopic eyes was also compared to the thresholds were assessed in the same eye for all these depth of suppression. This index of amblyopia conditions; only the instructions concerning when to showed similar negative correlation with depth of trigger trials changed.) These dominance thresholds suppression. The Pearson r values were -0.698 for were equivalent to the monocular baseline thresh- orthogonal gratings and -0.645 for identically ori- olds. A repeated measure analysis of variance indi- ented gratings. Only the first correlation is signifi- cated that there were significant differences among cantly different from zero, indicating that Snellen acuity reduction is not as good a predictor of depth of suppression as is contrast threshold elevation. These negative correlations are shown graphically as scat- * For the clinical suppressors, this condition could not be tested, because most had no periods of complete dominance by the habit- terplots in Figure 3. The top scatterplot shows the ually suppressed eye. relationship between the depth of suppression and

Downloaded from iovs.arvojournals.org on 09/27/2021 No. 3 SUPPRESSION AND AMBLYOPIA / Holopigion er ol. 449

the grating sensitivity difference for orthogonal grat- Table 2. Summary data ings and the second scatterplot shows the relationship between the depth of suppression and the interocular Orthogonally oriented gratings acuity ratios, also for orthogonal gratings. Depth of Contrast Acuity We also examined the relationship between depth Observer supp ratio ratio Stereoacuity of suppression and stereopsis. There was no signifi- LP -0.40 20.40 6.70 400 cant correlation between stereopsis and the depth of CD 3.90 7.80 2.00 3000 suppression for either orthogonal gratings (r = 0.454) BG 9.75 8.80 .50 100 AS 12.30 4.70 .25 140 or identical gratings (0.528). MM 20.50 1.30 .00 3000 Our results are summarized in Table 2, which lists MT 5.87 3.00 i1.00 50 depth of suppression ratios, contrast threshold ratios, KJ 3.30 3.16 .50 50 WV 11.53 2.25 .25 3000 Snellen acuity ratios and stereoacuities for the clinical PD 17.06 1.13 .00 3000 suppressors. Identically oriented gratings

LP 1.60 20.40 6.70 400 CD 3.30 9.54 2.00 3000 BG 8.00 6.73 .50 100 CD AS 16.40 5.82 .25 140 MM 22.60 2.60 .00 3000 0 MT 2.30 10.96 t.00 50 O KJ 4.10 4.52 .50 50 c WV 11.10 1.57 .25 3000 PD 14.80 3.35 .00 3000 Q5 3= T3 Discussion Q) Across the clinical suppressors tested, there was a significant negative correlation between depth of suppression and the degree of amblyopia for sinusoidal gratings with both orthogonally and identically oriented patterns, and a significant negative correlation between depth of suppression and Snellen acuity 7 for orthogonally oriented gratings. In other words, much stronger suppression effects were present in g 6 clinical observers with equal visual acuity than in to those with amblyopia. Before the implications of 5 these results can be considered, certain procedural O questions need to be addressed. 4 Is it possible that the eyes of the clinical observers o 3 became misaligned in the course of a trial, so that the 2 suppressed eye was directed at some point off the a> 2 CRT screen? We reject this hypothesis for several reasons. Great care was taken to ensure that eyes were 1 properly aligned at the start of a testing session. Ob- servers were asked to recheck alignment periodically during testing by alternately occluding each eye's view, and to realign the CRT screens, if necessary. Depth of suppression (dB) Furthermore, the magnitude of the suppression effect Fig. 3. Top, scatterplot demonstrating the relationship between for a given observer was remarkably consistent, the depth of suppression (in dB) and the contrast sensitivity differ- across the conditions tested here and several other ence (in dB) for orthogonal gratings of 3.3 c/deg. The data shown sets of experiments,18 making it difficult to attribute are for the nine clinical suppressors tested in this study. Bottom, the results to occasional eye misalignment. Finally, scatterplot demonstrating the relationship between the depth of when these observers were asked to report on the suppression (in dB) and the interocular acuity ratio (where a ratio of 1.0 indicates no difference between the eyes). The data shown are visibility fluctuations of the patterns, all observers re- for the nine clinical suppressors tested in this study. ported some periods in which the grating presented to

Downloaded from iovs.arvojournals.org on 09/27/2021 450 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Morch 1988 Vol. 29

the usually suppressed eye became visible, indicating triggered and remained present for only 200 msecs. In that both eyes were viewing the CRT screens simulta- other experiments on these same observers we mea- neously. sured the durations of dominance and suppression Another possibility is that the initial baseline in- under identical stimulus conditions. For the clinical crement thresholds were artificially elevated for some suppressor with the most rapid alternations in view- of the clinical suppressors, due to stimulation of the ing state, the average duration of any one viewing dominant eye even when it was occluded. Suppose, state was greater than 500 msec. For other observers, for example, that the dominant eye received some viewing states averaged 1000 msec or longer. The stray light from the CRTs while the monocular length of these suppression durations makes it im- threshold was being measured in the contralateral probable that the suppression state changed during eye. This stray light might have provided sufficient the course of the trial. We conclude that the differ- stimulation to produce suppression of the tested eye. ences among observers were not due to any method- This would have elevated the monocular baseline, ological difficulties and now turn to the task of inter- and since the depth of suppression was computed as a preting these findings. ratio relative to this baseline, a reduction in the over- Do anisometropes differ from strabismics with re- all suppression effect would have resulted. To exclude spect to suppression? Our anisometropic observers, this possibility, monocular increment thresholds on average, showed weaker suppression effects than were remeasured under three conditions for one ob- the strabismic observers. Conceivably this was be- server who had a small suppression effect (LP). The cause the stimulus for suppression in anisometropia nontested eye either viewed a uniformly luminous —blur—is less effective than is the misalignment that raster, viewed a dark occluder, or was carefully occurs in strabismus. More likely, it was due to the patched to exclude all light from the eye. There was fact that the anisometropes in this study were more no significant difference in the monocular thresholds amblyopic than the strabismics. A larger group of obtained in these three conditions, F (1,2) = 2.37, P anisometropes, with a wider range of amblyopia, > 0.05, confirming that stray light entering the domi- needs to be tested to determine whether suppression nant eye did not affect the monocular threshold. in anisometropia varies over the same range as sup- Is it possible that some clinical suppressors were pression in strabismus. simply more careful than others to trigger trials only Previous investigators have reported that the depth during periods of complete suppression, thereby pro- of suppression measured with a luminance increment ducing apparent differences in the magnitude of the task and the degree of amblyopia represented in suppression effect? If some trials were triggered when terms of acuity difference between the eyes vary simi- the contralateral eye was not completely dominant, larly across the visual field in clinical suppressors, the magnitude of the suppression effect would be un- such that areas of deep suppression are also deeply derestimated. This seems unlikely for several reasons. amblyopic.1314 Can these findings be reconciled with For one, the clinical suppressors with deeper ambly- our results? We believe so. The tasks employed by opia showed the least suppression, yet deeply ambly- Sireteanu and associates were different from ours: opic eyes are suppressed for long periods of time dur- they measured the depth of suppression using a lumi- ing binocular viewing. Hence, for amblyopic sup- nance increment, not a contrast increment, and con- pressors, the likelihood of a trial's being inadvertently trasted it to grating acuity, not contrast sensitivity. triggered during a nonsuppression state is actually They compared different points in the visual field of lower than it is for other observers. Careless triggering individual observers, not corresponding (foveal) of trials would thereby produce a greater suppression points in different observers, as we did. The conflict effect for amblyopic suppressors, not the smaller one may thus be more apparent than real. In addition, actually observed. Also, the magnitude of suppres- while Sireteanu does not present data for all of the sion was very uniform among the four normal ob- observers she tested, interobserver comparisons, servers and comparable to the magnitude of suppres- where possible, suggest that the depth of suppression sion found in other experiments on binocular ri- she found in the foveal region was much greater in valry.19"21 As three of the four normals were observers with alternating strabismus and equal vi- inexperienced psychophysical observers, they had no sual acuity14 than in strabismic and anisometropic more experience with the task than did the clinical observers with unilateral suppression and ambly- suppressors. opia.13 This is entirely consistent with our observa- A finalconcer n is whether the suppressed eye could tions. have become dominant before the contrast incre- What, then, can one conclude concerning the rela- ment was extinguished, thus allowing detection. The tionship between amblyopia and suppression? The contrast increment occurred 25 msecs after a trial was impaired monocular functioning of the amblyope's

Downloaded from iovs.arvojournals.org on 09/27/2021 No. 3 SUPPRESSION AND AMDLYOPIA / Holopigion er ol. 451

fovea evidently reduces the requirement for suppres- Acknowledgment sion to eliminate diplopia for binocular viewing con- We thank William Seiple for comments on an ear- ditions. We believe that suppression is a beneficial lier version of this paper. response which is used by the visual system as eco- nomically as possible. Our results should not be in- References terpreted to mean that amblyopia is a similarly pur- 1. Burian HM and von Noorden G: Binocular Vision and Ocular poseful response. Suppression alone is quite sufficient Motility. St. Louis, The C. V. Mosby Company, 1974. 2. Greenwald MJ and Parks M: Amblyopia. In Clinical Ophthal- to eliminate diplopia in the nonamblyopic individual mology, Vol. 1, Duane TD and Jaeger EA, editors. Philadel- without requiring a sacrifice of monocular acuity. phia, Harper and Row, Chapter 10, 1983, pp. 1-16. Neither our data nor those of other investigators 3. Gstalder RJ and Green DG: Laser interferometric acuity in demonstrate a direct causal correlation between am- amblyopia. J Pediatr Ophthalmol 8:251, 1971. blyopia and suppression. We suspect their mecha- 4. Levi DM and HarwerthRS: Spatio-temporal interactions in anisometropic and strabismic amblyopia. Invest Ophthalmol nisms are quite distinct. 16:90, 1977. Are binocular rivalry suppression and clinical sup- 5. Hess RF and Howell ER: The threshold contrast sensitivity pression in fact equivalent? Since both forms of sup- function in strabismic amblyopia: Evidence for a two-type pression apparently serve to protect the visual system classification. Vision Res 17:1049, 1977. from non-fusable binocular input, some authors have 6. Bradley A and Freeman RD: Contrast sensitivity in anisome- postulated that they are products of the same mecha- tropic amblyopia. Invest Ophthalmol Vis Sci 21:467, 1981. 22 7. Levi DM and Klein S: Hyperacuity and amblyopia. Nature nism. In our study, however, binocular rivalry sup- 289:268, 1982. pression and clinical suppression showed a different 8. Hess RF and Bradley A: Contrast perception above threshold pattern of results. In normal observers, identically is only minimally impaired in human amblyopia. Nature oriented dichoptically viewed gratings produce no ev- 287:463, 1980. 9. Cooper J and Feldman J: Random-dot stereogram perfor- idence for transient suppression while orthogonally mance by strabismic, amblyopic and ocular-pathology patients oriented gratings evoked alternating suppression and in an operant-discrimination task. Am J Optom Physiol Opt elevated increment thresholds during suppression. 55:599, 1978. This elevation in threshold was relatively uniform 10. Henson DB and Williams DE: Depth perception in stra- among the normal observers (on the order of 0.30 bismus. Br J Ophthalmol 64:349, 1980. 11. Schor CM, Bridgeman B, and Tyler CW: Spatial characteristics log-units) and never approached the larger magni- of static and dynamic stereoacuity in strabismus. Invest Oph- tudes (1.0 log-unit and greater) observed in some thalmol Vis Sci 24:1572, 1983. clinical suppressors. For the clinical suppressors, the 12. Holopigian K, Blake R, and Greenwald MJ: Selective losses in magnitude of suppression varied greatly among ob- binocular vision in anisometropic amblyopes. Vision Res servers, yet was remarkably consistent for any indi- 26:621, 1986. 13. Sireteanu R and Fronius M: Naso-temporal asymmetries in vidual observer, regardless of the orientations of the human amblyopia: Consequences of long-term interocular dichoptic stimuli. In this respect, the suppression that suppression. Vision Res 21:1055, 1981. occurs when clinical suppressors view orthogonally 14. Sireteanu R: Binocular vision in strabismic humans with alter- oriented patterns does not resemble binocular rivalry nating fixation. Vision Res 22:889, 1982. suppression in normal observers but is nearly identi- 15. Campos EC: Binocularity in comitant strabismus: Binocular visual fieldsstudies . Doc Ophthalmol 53:249, 1982. cal to the type of suppression the clinical observers 16. Schor C: Visual stimuli for strabismic suppression. Perception experience when viewing identical contours. 6:583, 1977. A difference between binocular rivalry suppression 17. Legge GE: Spatial frequency masking in human vision: binoc- in normal observers and clinical suppression was also ular interactions. J Opt Soc Am 69:838, 1979. noted by Smith et al.23 They found that normal ob- 18. Holopigian K, Blake R, and Greenwald MJ: The depth of suppression in amblyopes and nonamblyopes. ARVO Ab- servers exhibited a wavelength-specific abnormality stracts. Invest Ophthalmol Vis Sci 28 (Suppl):101, 1987. in the spectral increment sensitivity function during 19. Wales R and Fox R: Increment detection thresholds during binocular rivalry suppression that was not present binocular rivalry and suppression. Percept Psychophys 8:90, during normal viewing or binocular rivalry domi- 1970. nance states. Clinical suppressors did not show this 20. Fox R and Check R: Independence between binocular rivalry suppression duration and magnitude of suppression. J Exp wavelength-selective loss, regardless of whether the Psychol 93:283, 1972. test gratings were orthogonally or identically ori- 21. Blake R and Fox R: Binocular rivalry suppression: Insensitive ented. These differences between binocular rivalry to spatial frequency and orientation change. Vision Res suppression and clinical suppression seem funda- 14:687, 1974. mental, and pose a real challenge to theorists who 22. Fahle M: Non-fusable stimuli and role of binocular inhibition in normal and pathological vision, especially strabismus. Doc postulate that their mechanisms are identical. Ophthalmol 55:323, 1983. 23. Smith EL, Levi DM, Manny RE, Harwerth RS, and White JM: Key words: suppression, amblyopia, strabismus, anisome- The relationship between binocular rivalry and strabismic tropia, binocular rivalry suppression. Invest Ophthalmol Vis Sci 26:80, 1985.

Downloaded from iovs.arvojournals.org on 09/27/2021