This study of adult patients with strabismic amblyopia has demonstrated abnormalities in the VEPs recorded in response to both color and motion-onset stimuli. These abnormalities affect both the amblyopic and fellow eyes, are different for color and motion-onset stimuli, and are different in subjects with an onset of amblyopia before or after 18 months of age.
A previous study of motion-onset VEPs in persons with amblyopia described a small difference in peak time between amblyopic and fellow eyes and concluded that motion-onset responses were relatively spared in amblyopia.
13 The present study has confirmed the presence of a small difference in peak time between amblyopic and fellow eyes, but has also shown that the VEPs from both amblyopic and fellow eyes are abnormal compared with those from normal subjects, being of substantially shorter peak time and reduced amplitude. These abnormalities of motion-onset VEPs are similar in subjects with early- and late-onset amblyopia. Deficits of motion processing have been well described in various forms of amblyopia using both psychophysical methods (e.g., Refs.
16 17 18 19 ) and functional MRI
20 and have been shown to affect the fellow eye as well as the amblyopic eye.
16 18 21 Abnormalities of motion-onset VEPs affecting both eyes are thus not unexpected. However, it is surprising that the most striking abnormality is the shortening of the peak time response to stimulation of either the amblyopic or fellow eye. Possibly loss of some components of cortical processing, abnormal binocular interaction or some change in interaction between magno- and parvocellular pathways results in faster motion processing.
One previous study has examined changes in color responses in amblyopia, but differed from the present study in several important respects.
14 It used reversal, rather than color-onset VEPs and its subjects were children 5 to 10 years of age with no indication of age of onset. Many appear to have had binocular function. Despite these differences, both studies indicate a delay in the color VEP from the amblyopic eye. A study in which steady state VEP responses to achromatic checks of different sizes and temporal frequencies was examined also indicated a selective reduction in the response to the parvocellular-biased stimulus from the amblyopic eye of subjects with anisometropic amblyopia.
22
In the present study, the difference in peak time to a color stimulus between amblyopic and fellow eyes was only significant for the patients with late-onset amblyopia. This peak time was also delayed compared with normal, whereas peak times from the fellow eye and for both eyes of early-onset amblyopic patients did not differ significantly from normal. The pattern of change seen with a color stimulus was thus different from that seen for the motion-onset VEPs. In particular, there was a much greater change of the color VEP in the patients with late-onset amblyopia. Early studies of cell size change in nonhuman primates showed a reduction in parvocellular size relative to magnocellular cell size for both deprived and undeprived LGN cells after monocular visual deprivation.
10 11 A previous psychophysical study in essentially the same group of subjects with strabismic amblyopia as studied in the present study
12 demonstrated reduced color contrast sensitivity in both eyes, with the amblyopic eye being more affected than the fellow eye, and a relative reduction of color contrast sensitivity compared with luminance contrast sensitivity in both amblyopic and fellow eyes. These parameters probably reflect mainly parvo- and magnocellular functions, respectively, and indicate a reduction of parvocellular relative to magnocellular function, which was more marked in subjects with late-onset amblyopia. It is likely that the motion-onset stimulus used in the present study activates predominantly the magnocellular pathways and the color stimulus activates mainly parvocellular pathways. The above findings would thus reflect a different pattern of change in the magno- and parvocellular pathways in strabismic amblyopia.
The general picture that emerges is of a reduction of parvocellular function relative to magnocellular function in strabismic amblyopia, with the reduction being more marked in amblyopia of later onset. However, other results indicate that the difference between early- or late-onset amblyopia may be more complex. In a previous study of similar groups of patients with early- or late-onset amblyopia,
15 the peak time of the achromatic pattern onset VEP response to a 30-minute check was shortened in both eyes in patients with early-onset amblyopia. This resembles the changes found with motion-onset VEPs in patients with both early- or late-onset amblyopia in the present study. However, in the previous study, patients with late-onset amblyopia showed very different responses to the same 30-minute pattern-onset check, with a prolonged peak time from the amblyopic eye and a normal peak time from the fellow eye. This resembles the changes found here with the color VEPs in the subjects with late-onset amblyopia. The same achromatic pattern-onset stimulus thus gave very different results in the subjects with early- or late-onset amblyopia. This stimulus contained elements that would be expected to stimulate both the magno- and parvocellular pathways. In the early-onset amblyopia, the results appear to be dominated by the magnocellular pattern of change, whereas in late-onset amblyopia, they appear to be dominated by the parvocellular pattern. Abnormal visual experience starting at different ages would seem to affect the interaction between the magno- and parvocellular pathways differently.
The problems of recombining and binding information relating to the color and motion of an object are of considerable theoretical interest,
4 23 24 25 26 27 28 and under certain conditions, it is possible to produce misbinding illusions by making them combine incorrectly.
29 In general, these psychophysical studies in normal human subjects have provided evidence that the perception of the movement of an object is delayed in relation to perception of its color and contour. It is intriguing that the delays reported in some of those studies are similar to the difference in peak time between the motion-onset and color VEPs observed in our normal subjects. In our patients with strabismic amblyopia, this difference in peak time was reduced by up to half in amblyopic and fellow eyes, because the peak time of the motion-onset VEP was reduced and that of the color VEP increased. If the published psychophysical results and the present VEP data relate to the same underlying processes, it would be predicted that the difference in perception times for color and motion would be substantially reduced in both eyes of patients with strabismic amblyopia. It would be of considerable interest to study the questions of binding and misbinding of color and motion in these amblyopic patients, both for the understanding of the basic mechanisms involved and because it may demonstrate previously unsuspected abnormalities of visual perception, particularly in the fellow eyes, and may explain some the instabilities of the visual image described in amblyopia.
30 31
Changes in the fellow eyes of patients with strabismic amblyopia are of particular importance, because in most instances that is the eye that subjects rely on for vision in everyday life. Early studies in nonhuman primates described expansion of fellow eye ocular dominance columns in area 17 after early visual deprivation,
32 33 and studies of cell sizes in the LGN showed changes in cells in the undeprived laminae which were different with early- or late-onset deprivation and different in the magno- and parvocellular laminae.
10 11 Previous studies of fellow eye function have reported both increases
15 34 and decreases in certain visual functions.
16 21 35 36 37 38 39 40 41 42 43 The present study and a previous psychophysical study
12 have shown that the change in the fellow eye is not a simple increase or decrease in function, but a relative decrease in parvocellular relative to magnocellular function. Although this is evident mainly as a decrease in parvocellular sensitivity, the difference found between eyes in luminance contrast sensitivity is as much due to an increase in fellow eye sensitivity as a decrease in the amblyopic eye,
12 and the reduction in the difference between color and motion VEP peak time is as much due to shortening of the motion-onset peak time as the increase in color VEP peak time. The alterations in fellow eye function in amblyopia are complex, and the changes found are likely to depend on the spatial and particularly temporal characteristics of the stimulus used and whether it activates mainly magno- or parvocellular pathways. It is of interest that the two studies showing supranormal function of the fellow eye used rapidly presented, computer-generated stimuli that are more likely to activate magnocellular pathways.
12 34 It is not yet known whether this change in magno- to parvocellular balance in the fellow eye affects everyday visual perception in patients with amblyopia. They may have previously unsuspected visual difficulties, or indeed, abilities.
One of the surprising aspects of the present results is the greater relative sensitivity of the parvocellular pathways to abnormal visual experience with later onset of amblyopia. This may not represent a true increase in sensitivity as, although the visual acuity of the early- or late-onset groups were similar in this and the previous study,
12 it may be that more abnormal visual experience was needed to produce this change of acuity in the late-onset groups. However, it is well recognized in animal studies that sensitivity of some visual functions to abnormal visual experience increases with age.
44 In anatomic studies of the primate LGN, the change in size of undeprived parvocellular cells produced by 2 months of monocular lid closure increased between 3 and 6 months of age (Fig. 13 in Headon et al.
10 ). There is evidence that parvocellular pathways continue to show developmental change at a later age than for the magnocellular system.
45 46 47 Against this background, it would not be surprising if there were a true increase in the sensitivity of the parvocellular system to abnormal visual deprivation as it matures.
The differential effects of abnormal visual experience on the magno- and parvocellular pathways may have implications for the treatment of amblyopia. Not only do the magno- and parvocellular pathways respond differently to the initial abnormal experience, but LeVay et al.
32 illustrate an instance in which closure of one eye from birth in a nonhuman primate was followed by reopening of the eye and closure of the other eye at 3 weeks of age—an experimental model of patching (Figs 38 and 39 in LeVay et al.
32 ). That animal showed re-expansion of the ocular dominance columns in layer IVCβ, but no re-expansion of the columns in the immediately overlying IVCα. What effect this dissociation of magno- and parvocellular change may have on visual function has never been examined, but such residual disorganization of the visual pathways after amblyopia and patching may be a factor in those children whose acuity does not improve despite good compliance.
This study has shown that the changes that occur in the central visual pathways in strabismic amblyopia are complex. Not only are different changes found in parallel components of the pathway, but the interactions between the components are different, and both the changes and the interactions differ according to the age of onset of the amblyopia. Although patching is frequently effective in improving visual acuity, it does not selectively treat these different components of the visual pathway. In view of the complexity of the changes, it is perhaps not surprising that the response to patching varies greatly between children of a similar age, that some continuing deficit in acuity is common and that some children fail to respond to occlusion, even with good compliance. Understanding these complexities should lead to novel and selective forms of treatment that may need to be tailored according to the characteristics of the individual child.