Abstract
We studied changes in illusory contour clarity as a function of retinal eccentricity. We used an adaptive staircase to identify the point of subjective equality (PSE) for pairs of illusory contours at differing eccentricities. In two experiments, subjects were shown two illusory rectangles with circular inducing elements and a fixation cross. While fixating, they judged the relative clarity of the nearest illusory edges. The fixation point appeared at one of five horizontal locations ranging from the midpoint between the two illusory rectangles to 14 degrees to the left (which was inside the left rectangle). In Experiment 1, the vertical distance between the inducing elements of the left rectangle changed from trial to trial. It decreased when the left illusory edge was judged weaker than the right and increased otherwise. In Experiment 2, the size of the inducing elements for the left rectangle changed; the diameter was increased when the left illusory edge was judged weaker than the right, and decreased otherwise.
For ten subjects, in each experiment, the average PSE determined by varying inducing element size and spacing was significantly affected by the location of the fixation point. In general, the nearer the illusory edge was to the point of fixation, the larger the gap needed to match a more distant illusory edge, and the smaller the inducing elements needed to match a more distant illusory edge.
These results are consistent with decreases in strength of contour interpolation with retinal eccentricity. These findings may be modeled in terms of a modified ratio rule for contour strength, or an eccentricity dependent grouping-kernel size.
The effect of some spatial variables appears independent of retinal size in free viewing of illusory contours. It remains to be determined how retinally dependant grouping will be integrated over time to yield retinally independent grouping.