Abstract
Boundary contour information that defines surface regions can affect binocular rivalry (BR). A vertical grating disc surrounded by a horizontal grating half-image competing with a horizontal grating half-image leads to a high predominance for seeing the disc. This is attributable to either the monocular boundary contour (MBC) defining the disc, or the feature difference (vertical vs. horizontal) between the center (disc) and surround regions. Experiment-1 investigated this by measuring contrast threshold of a Gabor probe in three conditions with variously specified MBC. (1) MBC-condition: one half-image had a vertical MBC disc (1.5deg) created by a 90deg phase-shift relative to the surrounding vertical grating and the other half-image had a homogeneous vertical grating (2.2cpd, 1.5log%, 60cd/m2). (2) Ring-condition: a white ring (121.2cd/m2, 0.045deg thick), instead of the phase-shift, defined the MBC. (3) MBC+Ring condition: the same white ring encircled the phase-shifted disc. For all three conditions we found thresholds were lower on the half-image with the MBC than on the homogeneous grating half-image, indicating binocular suppression. Critically, since the Ring-condition carries only MBC, this suggests that MBC alone can lead to dominance. Experiment-2 measured BR predominance of a vertical grating disc surrounded by horizontal grating in one half-image and a horizontal grating disk created by phase-shifting it relative to the surrounding horizontal grating in the other half-image. Confirming our previous findings (Xu et al, 2006), the predominance for seeing the horizontal grating disc increased with increasing phase-shift that strengthened the boundary contour of the horizontal disc and enhanced the center-surround difference. We then added a white ring (35 cd/m2, 0.04deg thick) to encircle the phase-shifted horizontal disc. We found that with the ring, the predominance for seeing the horizontal grating remained unchanged with increasing phase-shift. This suggests that the center-surround factor contributes less to BR than the boundary contour factor.
Supported by a grant from the NIH (R01 EY015804).