Abstract
In a typical binocular rivalry (BR) display, orthogonal grating disks (vertical vs horizontal) at corresponding retinal areas induce a strong alternation wherein each disk has ∼ 50% predominance. But by placing these grating disks on, say, a vertical grating background in each half-image (thus blending the vertical grating disk into the background), the alternation is vastly reduced as the horizontal grating disk claims ∼ 90% predominance. We proposed that the perceptual difference between this, monocular-boundary-contour (MBC)-display, and the typical-BR-display reveals a preference for boundary contours (He & Ooi 02). Only a monocular boundary contour (rim of horizontal grating disk) exists in the MBC-display, hence its predominance, whereas binocular boundary contours (rims of the two grating disks) exist in the typical-BR-display. To further test the proposal, Exp 1 tracked the BR dynamics of a modified MBC-display with a monocular grating disk on two types of random-dots background conditions. i) Fully correlated condition: the two half-images had the same random-dots pattern; ii) Partially correlated condition: the random-dots pattern in one half-image was flipped around the vertical and horizontal axes. It was predicted that the fully correlated condition would have more BR alternations as stronger binocular boundary contours were formed between the correlated random-dots around the rim of the grating disk and its corresponding area in the other half-image. Confirming this, we found more alternations in the fully correlated condition, while the grating disk in the partially correlated condition has a higher predominance. Besides BR's varied dynamics, the sensitivity of the suppressed area is expected to decrease. Thus Exp 2 used a probe to measure the sensitivity reduction in the MBC-display. We found a suppression threshold similar to that in other BR displays, suggesting a common (suppression) fate beyond the boundary contour processing stage.