Abstract
The visual system endows an image region that owns the boundary contour as the modal/occluding surface, and in binocular rivalry, such a region tends to dominate over another with a weaker boundary contour. In this study, we investigated the proposal that the process of representing the dominant surface begins and spreads from its boundary contours. We used a rivalry stimulus with a common vertical grating in both half-images, and an additional quasi-rectangular area (3 lateral widths: 1.50, 2.00, 2.66deg) with horizontal grating placed in the center of the vertical grating background in one half-image (SF=4cpd; contrast=29.1%; luminance=40.3cd/m2). Since the monocular area with the horizontal grating surface owns the boundary contours, it tends to dominate almost continuously. To test the prediction that the global dominance of the horizontal grating is reached by the spreading of the horizontal grating surface from the monocular boundary contours at both (vertical) sides of the quasi-rectangle, we varied the rivalry display duration (30–500msec) for each rectangular width. Confirming the prediction, at short duration (lt;250msec), the observer perceived the horizontal grating pattern at the right and left sides of the quasi-rectangular area, with the central area having no horizontal grating (filled instead with vertical grating, piecemeal, and/or checkerboard percepts). As the stimulus duration increased, the horizontal grating percept increased from both ends of the quasi-rectangle, indicating a spread toward the center. To measure the spread of the horizontal grating surface, observers rated the relative size of the center of the quasi-rectangle (gap) that had no horizontal grating percept. We found that the reported gap size decreased as the stimulus duration increased up to approximately 250msec. From these data, we estimated the average spreading speed of the horizontal grating surface from the two ends of the quasi-rectangle, i.e., from the monocular boundary contours, to be about 1.5–3 deg/sec.
Supported by a grant from NIH (R01 EY 015804)