Reversed Depth Illusion in Random-dot Stereograms Becomes More Visible When the Stereograms Are More Dynamic in Both Central and Peripheral Vision

Poster Presentation 26.459: Saturday, May 18, 2024, 2:45 – 6:45 pm, Pavilion
Session: 3D Perception: Depth cue integration, neural mechanisms

Li Zhaoping1 (); 1University of Tuebingen, Max Planck Institute for Biological Cybernetics

In anticorrelated random-dot stereograms (RDSs), a black dot in one eye corresponds to a white dot in the other eye to depict visual depth of object surfaces. Neurons in the primary visual cortex (V1) respond as if the depth order between surfaces is reversed from that defined by binocular disparities (Cummings & Parker 1997). The resulting reversed depth illusion is visible in peripheral but not in central vision (Zhaoping & Ackermann 2018). A Central-peripheral Dichotomy (CPD) theory explains this lack of central illusion as follows. Feedback from higher to lower visual areas to aid recognition mainly targets the central, rather than peripheral, visual fields. Such feedback verifies feedforward sensory signals, and thereby vetoes misleading V1 signals that cause the illusion. However, this illusion becomes visible in central vision using dynamic RDSs in which the random set of dots was replaced every 10 milliseconds by another randomly generated set while keeping the scene and stimulus design unchanged (Zhaoping ECVP 2021). The subsequent dots backward mask the previous dots, thereby compromising the feedback process that normally vetoes the illusion. In this study, we test whether this illusion also becomes more likely perceived in peripheral vision (at about 10 degrees eccentricity in the lower visual field) when the RDSs are made dynamic, and examine whether this illusion is more visible when the RDS becomes more dynamic, i.e., when the duration (e.g., 10 or 20 milliseconds) of each random set of dots becomes shorter. We report that, while this illusion remains more visible in peripheral vision, this visibility is enhanced in both central and peripheral vision when the RDSs are more dynamic. These findings suggest that the feedback verification is also present, albeit weaker, in peripheral vision. Whether backward masking enhances the reversed depth illusion more in central than peripheral vision will be reported.