Abstract
Two possible mechanisms have been proposed for binocular slant detection in the visual cortex: (1) binocular differences in orientation (Sanada & Ohzawa 2006) and (2) variations in binocular position differences across the surface (Bridge & Cumming 2001). The possible contributions of these two forms of binocular disparity have proved difficult to disentangle since the two are confounded on slanted surfaces. We used reverse correlation to find binocular classification images for a task in which observers judged the slant of Gabor stimuli slanted around a horizontal axis. On each trial, the Gabor was slanted so that the top was either nearer or further away, and the observer's task was to judge the direction of slant. Oriented, bandpass noise, which was uncorrelated between the left and right eyes' views, was added to the stimuli. The cyclopean orientation of the stimuli was varied to prevent classification by monocular orientation. 4 observers completed a total of 15900 trials. The classification images show vertically oriented structures in both left and right views, rather than a difference in orientation between the eyes. We also found evidence of phase difference in the classification images between left and right views, consistent with the use of position or phase disparities. This provides psychophysical evidence against the use of orientation disparity in the perception of binocular slant.
Meeting abstract presented at VSS 2017