Abstract
The motion-induced position shift (MIPS) (DeValois & DeValois, 1991) and the flash-drag effect (FDE) (Whitney & Cavanagh, 2000) are examples of illusory position shifts induced by visual motion information. Previous studies have revealed that several different kinds of motion induce these illusions. However, their selectivity for binocular disparity is largely unknown. The disparity tuning of the MIPS and FDE, or the absence of it, would be indicative of the degree of contribution from processing stages in which depth relationship is not made explicit. In the MIPS experiment, a sinusoid with a disparity was windowed by a static Gaussian envelope with zero disparity. The background was filled with static random noise. Subjects correctly reported whether the grating was in front or behind. Two such Gabor patches with the same disparity were presented 7 deg above and below the fixation point and moved horizontally in opposing directions. We determined the position of subjective alignment as the magnitude of the MIPS. We found that the MIPS occurred even when the grating had a different disparity than the envelope’s. From the results of the current and previous studies on the MIPS, we argue that the MIPS involves several distinct visual processing stages. In the FDE experiment, two random-dot patterns having the same disparity were presented in upper and lower portions of the visual field and moved horizontally in opposing directions. Two Gaussian blobs with zero disparity were flashed outside these moving random-dot patterns. The FDE, quantified as the position of subjective alignment, also occurred with crossed disparities, but there was a tendency that the magnitude of the FDE reduced with uncrossed disparities. From these results, we will propose possible neural mechanisms that underlie these illusions. In particular, we will discuss the degree of functional independence between motion-based computation of position and disparity-based depth ordering.
Meeting abstract presented at VSS 2012