PSYCHOPHYSICS OF SUPPRESSION

RICHARD HARRAD Bristol

SUMMARY placed over one eye to measure the extent of the

In amblyopia the visual input from the amblyopic eye is area of suppression. The subject was instructed to suppressed. The difference in contrast sensitivity look at a distant light and the onset of diplopia was between the two eyes can lead to suppression of the taken to indicate that the boundary of the scotoma amblyopic eye through dichoptic masking in aniso­ had been reached. Like Travers he also found that metropic amblyopia and small angle strabismus. Bino­ the area of suppression included the point of cular cells in the visual cortex may also be instrumental confusion and the point of diplopia and that it was in bringing about suppression in small angle strabismus larger in dense amblyopia. However, he also found through fusional suppression and disparity-dependent that in most patients he studied, the suppression suppression. In large angle strabismus without amblyo­ scotoma stopped abruptly at the vertical meridian pia suppression is very powerful and is probably a passing through the fovea. J ampolsky quotes his own modified form of binocular rivalry suppression. Which experience and that of Posner and Schlossman4 that particular mechanism is operating is dependent on the over-correction of a strabismus by as little as 2 or 3 pathogenesis of the amblyopia, the depth of the prism dioptres will cause diplopia in most cases. He amblyopia and, in strabismic amblyopia, the angle of reported that there was an area of 'hemiretinal strabismus. In any given subject more than one suppression' in the deviating eye which extended mechanism may be operating in the different areas of between the diplopic point and a vertical line running the field. through the fovea. He made the assumption that the region where a target initiated suppression was the The child with strabismus and amblyopia does not same as the area of suppression; however, this does suffer from double vision because the visual input of ! not necessarily follow. He had located the area the deviating eye is suppressed. Von Graefe was the where suppression may be initiated but it is possible first to demonstrate a suppression scotoma in an that the entire visual field of the deviating eye was amblyopic eye. He placed a red filter over the good suppressed when there was no diplopia. This is the eye and with the amblyopic eye fixing a light in the S conclusion reached by Pratt-Johnson and Tillson, centre of a screen performed perimetry in the usual who presented fusible targets on a modified Lees way. The presence of a suppression scotoma in the screen. They found that when strabismic patients amblyopic eye was indicated by the test object were overcorrected with prisms, most developed turning red. Travers2 put a mirror in front of one diplopia, and they refer to the critical nature of the eye and presented targets on separate screens to the ocular alignment about the vertical midline as the two eyes independently. He found that the region of hemiretinal trigger. They concluded that since suppression in the deviating eye was in the area of diplopia was present throughout the binocular visual the retina where the image presented at the fovea field after the midline had been transgressed, the of the fixing eye falls, i.e. the diplopic point, and that entirety of this area must have been suppressed the area of visual field suppressed was larger in dense previously. This does not necessarily follow; it may amblyopia. In some patients Travers also found be that diplopia is simply not perceived very readily suppression at the fovea of the deviating eye (the in the periphery of the visual field. These authors and point of confusion). These findings have been widely 6 Sireteanu et al. found central suppression sur­ quoted and form the basis of many accounts of rounded by peripheral fusion in microtropic subjects. suppression in strabismus. 7 Sireteanu and Fronius put forward an interesting Jampolskf used prisms of increasing strengths argument to explain these findings. They suggest that the progressive enlargement of receptive field size Correspondence to: R. Harrad, Bristol Eye Hospital, Lower Maudlin Street, Bristol BSl 2LX, UK. from the fovea to the retinal periphery means that a

Eye (1996) 10,270-273 © 1996 Royal College of Ophthalmologists PSYCHOPHYSICS OF SUPPRESSION 271 small misalignment of the eyes will lead to loss of the amblyopes there was no masking of the normal eye capacity for constructive binocular interactions at the by the amblyopic eye although the amblyopic eye fovea where receptive field sizes are small, but that in continued to be masked by the normal eye. In pure the retinal periphery there will still be the possibility strabismic amblyopia with moderate amblyopia and of facilitatory interactions and stereopsis since the in alternate fixation, suppression was more profound physical separation of the cortical representation of than would be expected on the basis of dichoptic these areas is very small. In order to abolish diplopia, masking alone. This is probably because dichoptic suppression must therefore occur only in the central masking cannot operate once all retinal correspon­ area around the fovea. dence is lost. This loss of low contrast information There are certain difficulties inherent in the study due to dichoptic masking is similar to the process of suppression in strabismus. Firstly, in order to be whereby a progressive reduction of the contrast of able to distinguish between the stimuli presented to one of a pair of stereo half-images leads to the the two eyes they need to be different in some way. reduction of stereo acuity and eventually suppression 3 !; However, Jampolsky and Schor have shown that of the half image having the lower contrast.13-15 It is similar stimuli are more likely to produce suppres­ probably this process, in the presence of reduced sion than dissimilar stimuli, so any difference contrast sensitivity in anisometropic amblyopia, that between the stimuli which enables the subject to is responsible for the poor stereo acuity in these J6 distinguish them will tend to break down the patients. However, in subjects with deep strabismic suppression which is being measured. An analogy amblyopia or alternate fixation, suppression cannot can be made here with Heisenberg's Uncertainty be accounted for by dichoptic masking. Principle. Workers who have used orthogonal Binocular rivalry occurs when each eye views two patches of grating, or red and green stimuli viewed images that are so dissimilar that they cannot be J7 19 through red and green goggles, are likely to have fused. - The observer experiences alternating underestimated the extent to which suppression dominance and suppression of each binocular occurs. In any given experiment the conclusion that image. This is a normal physiological process which is reached applies only to the immediate area of the is able to bring about perceptual suppression of the visual field containing the suppressed stimulus; no input of one eye. Harrad and Hess20 measured the firm conclusions can be drawn about what is threshold for the onset of rivalry between gratings happening elsewhere in the visual field where presented independently to the two eyes which suppression may or may not be occurring. The differed only in orientation. A grating of a given above considerations go some way to explaining orientation was presented to one eye and a grating of why the findings of different workers with regard to a different orientation was presented to the other eye the area of suppression in strabismus show such and the contrast of this second grating was increased differences. from threshold until an episode of rivalry suppres­ Dichoptic masking is the physiological process sion occurred. They asked whether a raised threshold where a stimulus of a given contrast presented to one for binocular rivalry could account for suppression in eye can prevent the detection of a lower contrast but amblyopia. They found that threshold elevation was otherwise identical stimulus presented to the other seen for gratings of similar orientation, i.e. less than 9 IO eye. , This binocular inhibitory process is known to 30° of orientation difference, in anisometropic u be present in some amblyopic subjects. Harrad and amblyopia and in normal SUbjects. This was simply Hess12 put forward the following hypothesis: since a measure of the orientational tuning of dichoptic there is a difference in the thresholds for contrast masking. Threshold elevation also occurred for the detection between the two eyes in amblyopic amblyopic eye of mixed strabismic anisometropic subjects, the normal eye will always receive input subjects, while the normal eyes of these subjects of a higher supra-threshold contrast and therefore showed a very small threshold elevation with no the input to this eye should always mask the input to orientational tuning. These findings supported the the amblyopic eye. If this is the case then dichoptic view that the stimuli which are most likely to produce masking in the presence of a contrast threshold suppression, i.e. in the case of gratings those of difference could account for suppression in amblyo­ similar orientation, are not those which produce pia. They looked at a range of amblyopic subjects binocular rivalry. The similarity between the orienta­ with different degrees of severity and types of tional tuning in anisometropic amblyopia, mixed amblyopia. They found that dichoptic masking strabismic and anisometropic amblyopia and in could account for suppression in some anisometropic normal subjects suggests that a form of orientational amblyopes and also mixed strabismic and anisome­ tuned contrast masking is producing suppression in tropic amblyopes at low spatial frequency once the these patients. One alternately fixating amblyopic difference in contrast threshold is accounted for. At subject showed a very high threshold elevation in high spatial frequencies in the mixed strab/aniso either eye in the absence of any orientational tuning. 272 R. HARRAD

To summarise, Harrad and Hess20 found that the found that suppression in his group of subjects was suppression was relatively weak and orientationally stimulated by a fixed disparity rather than suppres­ tuned in anisometropic amblyopia, more powerful sion being confined to a particular area of the visual but still orientationally tuned in the presence of field. He found that uncrossed disparity led to strabismus, and very powerful and lacking in suppression in his esotropic subjects and crossed orientational tuning in alternate fixation. The lack disparity produced suppression in his exotropic of orientational tuning in alternate fixation is to be subject. The subjects with normal retinal correspon­ expected from the results of animal experiments dence suppressed at zero disparity, while those with which show that such subjects are likely to have very abnormal retinal correspondence suppressed at the !> few binocularly driven visual cortical cells.21 Blake1 abnormally corresponding point where the disparity has proposed that powerful suppression seen in was equal to the angle of the strabismus. There was alternately fixating strabismics is necessary to inhibit therefore no evidence of selective suppression of all the large monocular pool of cells driven by the or part of one hemifield. He suggested that although contralateral eye. Holopigian et al.22 measured the suppression may be disparity dependent in small depth of suppression in a group of amblyopic angle strabismus, this is unlikely to be the mechanism subjects and found an inverse correlation between in large angle strabismus. the depth of suppression and the depth of amblyopia. Observers with normal stereopsis suppress some of Their results also show that suppression was most the information contained in each stereo half-image­ powerful in alternately fixating strabismics, inter­ a phenomenon called 'fusional suppression'. Where mediate in strabismic amblyopia and weakest of all in the two stereoscopic half-images consist of a vernier anisometropic amblyopia. offset target presented to one eye and aligned Many authors23,24 have suggested that binocular vertical lines presented to the other eye, the observer rivalry is the basis of suppression in amblyopia. is unable to make accurate positional judgements However, there are a number of objections to this about the vernier component when the two half­ proposition: (i) the characteristic alternation of images are fused. McKee and Harrad28 predicted binocular rivalry is not seen in strabismic suppres­ that since they lack a functional stereo system at the sion; (ii) the wavelength suppression characteristics fovea, stereoanomalous observers would have better of binocular rivalry differ from those of strabismic access to monocular positional information than 5 suppression,2 in that normal observers exhibit normal observers. This prediction was confirmed; wavelength-specific loss of sensitivity during the stereoanomalous observers were better at making suppression phase of rivalry affecting the region real judgements for the stereo half-images than between 400 and 470 nm but this wavelength-specific normal observers; the competition between units en­ loss is not seen in strabismic suppression; (iii) the coding conflicting information in the subjects seemed strength of suppression in strabismus has been found to be resolved by suppressing signals from the to be much stronger than rivalry suppression in weaker eye. This fusional suppression was present 1 normal subjects;2 and (iv) the visual stimuli, such as for both vertical and horizontal disparities and for gratings of different orientation, that lead to bino­ this reason it was felt to be a function of the fusion cular rivalry have tended to stimulate rivalry rather mechanism rather than the stereo mechanism. They than suppression in amblyopia.8 However, it is suggested that the physiology amounted to suppres­ probable that at the onset of strabismus in children sion of small units coding for position by large units over 6 months of age diplopia and binocular rivalry detecting large disparities, thus leading to loss of occur and it is likely that suppression in large angle positional information. strabismus and in those subjects that can alternately It seems that there are a combination of processes fix with either eye is a modified form of binocular operating to produce suppression in amblyopia. rivalry. Which particular process is operating depends upon 6 Richards2 described stereoanomalous subjects the pathogenesis of the amblyopia, the depth of who were unable to make judgements of depth for amblyopia, the angle of the strabismus where either crossed or uncrossed disparity. Schor8 showed present, and whether or not there is the capacity that suppression was most pronounced for gratings of for alternate fixation. Dichoptic masking probably a similar orientation and that vertically orientated largely accounts for suppression in anisometropic gratings provoked suppression more readily than amblyopia and in small angle strabismus. Disparity­ horizontally orientated gratings, suggesting that specific suppression or fusional suppression may also horizontal disparities are more likely to be a stimulus be important in those patients with a small angle for suppression than vertical ones. Schor27 measured strabismus in the absence of amblyopia. In large the area of suppression in the deviating eye in terms angle strabismus where normal binocular interac­ of the position of targets presented to the fixating eye tions are not possible a powerful form of binocular in a group of subjects with small angle strabismus. He rivalry suppression occurs. The differences between PSYCHOPHYSICS OF SUPPRESSION 273 these processes in the normal population and in 14.Legge GE, Gu Y. Stereopsis and contrast. Vision Res amblyopic subjects can be accounted for by modi­ 1989;29:989-1004. 15.Schor CM, Heckmann T. Interocular difference in fications in the underlying neural circuitry which take contrast and spatial frequency: effects on stereopsis place during visual development. and fusion. Vision Res 1989;29:837-47. 16.Holopigian K, Blake R, Greenwald Ml. Selective Key words: Amblyopia, Binocular rivalry, Strabismus, Suppres­ losses in binocular vision in anisometropic amblyopia. sion. Vision Res 1986;26:621-7. REFERENCES 17.Wheatstone C. Contributions to the physiology of vision.Part the first: On some remarkable, and hitherto 1. von Graefe A. Das Sehen der Schielenden. Eine unobserved, phenomena of binocular vision.Phil Trans ophthalmologisch-physiologische Studie. Wiesbaden, R Soc Lond 1838;128:371-94. 1896. 18.Blake R.A neural theory of binocular rivalry.Psychol 2. Travers T. Suppression of vision in squint and its Rev 1989;96:145-67. association with retinal correspondence and amblyo­ 19.Fox R. Binocular rivalry. In: Regan D, editor. Vision pia. Br 1 OphthalmoI1938;22:577-604. and visual dysfunction. vol 9. London: Macmillan, 3.lampolsky A. Characteristics of suppression in strabis­ 1991:93-11 O. mus. AMA Arch Ophthalmol 1955;54:683-96. 20. Harrad RA. Hess RF. A model for suppression in 4. Posner A, Schlossman A. Relation of diplopia to amblyopia. Transactions of the 20th Meeting of the binocular vision in concomitant strabismus. AMA European Strabismological Association 1992:113-8. Arch OphthalmoI1951;45:615-22. 21.Crawford MLJ, Smith EL III, Harwerth RS, von 5. Pratt-lohnson lA, Tillson G. Suppression in strabis­ Noorden GK.Stereoblind monkeys have few binocular mus: an update. Br 1 Ophthalmol 1984;68:174-8. neurons. Invest Ophthalmol Vis Sci 1984;25:779-81. 6. Sireteanu R, Fronius M, Singer W. Binocular interac­ 22. Holopigian K, Blake R. Greenwald Ml. Clinical tion in the peripheral visual field of humans with suppression and amblyopia. Invest Ophthalmol Vis strabismic and anisometropic amblyopia. Vision Res Sci 1988;29:444-51. 1981;21:1 065-74. 7. Sireteanu R, Fronius M.Naso-temporal asymmetries in 23.Fahle M. Non-fusible stimuli and the role of binocular human amblyopia: consequence of long-term inter­ inhibition in normal and pathologic vision, especially ocular suppression. Vision Res 1981;21:lO55-63. strabismus. Doc Ophthalmol 1983;55:323-40. 8. Schor CM. Visual stimuli for strabismic suppression. 24.von Noorden GK.Binocular vision and ocular motility: Perception 1977;6:583-93. theory and management of strabismus. St Louis: CV 9. Abadi RV. Induction masking: a study of some Mosby. 1990. inhibitory interactions during dichoptic viewing.Vision 25.Smith EL III, Levi DM. Manny RE, Harwerth RS, Res 1976;16:269-75. White 1M. The relationship between binocular rivalry 10. Legge GE.Spatial frequency masking in human vision: and strabismic suppression.Invest Ophthalmol Vis Sci binocular interactions.1 Opt Soc Am 1979;69:838-74. 1985;26:80-7. 11. Levi DM, Harwerth RS, Smith EL III. Humans 26.Richards W. Anomalous stereoscopic depth percep­ deprived of normal binocular vision have binocular tion.l Opt Soc Am 1971;61:4lO-4. interactions tuned to size and orientation. Science 27.Schor C. Zero retinal image disparity: a stimulus for 1979;206:852-4. suppression in small angle strabismus.Doc Ophthalmol 12. Harrad RA, Hess RF.Binocular integration of contrast 1978;46:149-60. information in amblyopia.Vision Res 1992;32:2135-50. 28.McKee SP, Harrad RA. Fusional suppression in 13.Halpern DL, Blake R. How contrast affects stereo­ normal and stereoanomalous observers. Vision Res acuity.Perception 1988;17:483-96. 1993;33:1645-58.