Abstract
Purpose: Simultaneous contrast demonstrates that a uniform grey is seen as darker on a white background than on a black background. The phenomenon can be easily understood in terms of centre-surround mechanisms known to exist throughout the visual system. The sizes of these mechanisms are known to increase with eccentricity. Therefore, we ask whether the strength of the simultaneous contrast illusion changes with stimulus size and eccentricity.
Method: Stimuli comprised five square regions stacked to form a vertical rectangle. Region 2 (the standard) was of fixed luminance (.5) and subjects adjusted the luminance or region 4 (the match) to match that of region 2. Regions 1 and 5 contained luminance gradients from 0 to 1 [/] and region 3 contained a luminance gradient going from 1 to 0 [\]. Therefore, the cross-section of a stimulus going from top to bottom would be: [/—\—/]; i.e., region 2 was surrounded by white and region 4 by black. Stimuli were presented at 9 logarithmically spaced sizes ranging from 102 to 1152 pixels at eccentricities of 0, 1, 2, 4 to 8° in the right visual field.
Results: Under free viewing conditions (stimuli presented nominally at 0°) matches were veridical (.5) at small sizes and decreased (e.g., to .30) as stimulus size increased; i.e., the standard illusion was seen at large stimulus sizes. Similar results were found when subjects maintained fixation on a small dot and stimuli were presented at 0 to 8°. Unlike standard size scaling results, however, the functions relating stimulus size to match luminance were not shifted versions of each other on a log size axis. The curves at 0 to 4° were essentially superimposed and that at 8° was shifted rightward.
Conclusions: Although simultaneous contrast is size dependent it does not show typical eccentricity dependence.
This research was supported by NSERC grants to Rick Gurnsey.