Abstract
Form vision is impaired in the fovea of an amblyopic eye. One hypothesis for the neural mechanism underlying amblyopia is that it results from a cessation of the development of the visual system, thus explaining why the amblyopic fovea shares many characteristics of the normal periphery. Consistent with this view, it is believed that crowding in the amblyopic fovea is qualitatively identical to that in the normal periphery. Here we reexamined this hypothesis using the first- and second-order classification image techniques described in Nandy & Tjan (2007, Journal of Vision). Two strabismic amblyopes with comparable visual acuities, GK & RH, participated in the study; both had extensive experience in psychophysical experiments. They were tested at their “fovea” for the non-amblyopic and amblyopic eye separately, with the target letters (“o”, “x”) presented with or without flankers. For both subjects, contrast thresholds were significantly higher for the flanked than the unflanked conditions, indicative of crowding. As is the case for peripheral crowding, the first order classification images at the target location showed no distortions except for a reduction in image contrast. The similarity with peripheral crowding broke down for RH but not for GK when we obtained the classification images at the flanker locations and in the second-order analysis. Specifically, RH did not show any systematic link between flanker features and response errors, which is prominent in peripheral crowding. RH also showed fewer changes in feature utilization compared to GK and normal periphery, even though the effect of crowding in terms of threshold elevation was comparable between the two subjects. For RH, unlike crowding in normal periphery, the flankers behaved as if they induced a strong stochastic internal noise that masked the target.
NIH EY016391 (BST), UH GEAR grant (SC).