The coarse vs. fine dichotomy in stereopsis: A matter of scale

Abstract

Is disparity processing subserved by a single mechanism that spans all disparities or by two distinct mechanisms that operate over different disparity ranges? Data from depth pedestal experiments have proven divided on this issue, likely due to the wide range of stimuli and tasks used, and to variables such as stimulus size and retinal eccentricity. In an effort to resolve this issue we exploit a result from Wilcox and Hess (1995) who showed that the coarse (2nd-order) system dominates processing for diplopic targets. Further, 2nd-order stereopsis depends critically on scale, as the upper limits for stereopsis increase with increasing stimulus width. Here we used a 1AFC method of constant stimuli to measure discrimination thresholds for a test bar relative to a reference bar positioned at depth pedestals ranging from 0 to 1 deg (n=5). The stimulus parameters were selected so that the target and reference would become diplopic within the test range. Not surprisingly, discrimination performance was degraded as pedestal disparity was increased. This occurred over a large range of disparities within the fusable range. However, outside the fusable range we find a dramatically different pattern of results; discrimination performance plateaus, remaining constant in spite of increasing diplopia. Our data echo those of Ogle (1952) and others, but with important differences that are tied to the test disparity relative to the stimulus width. We argue that these data reflect a neural dichotomy between coarse and fine processing that is tightly tied to scale rather than absolute disparity.