Abstract
Aim: Do fast translating motions leave neural ‘streaks’? We test this by measuring sensitivity to static oriented probes while fast ‘streaky’ motions engage in binocular rivalry. Methods: Rivalry stimuli were orthogonally translating fields of Gaussian blobs, each with 80 high contrast blobs (40 dark, 40 light). Blob size was 0.22° (defined as 4×SD). Fast and slow speeds were compared: 8.6°/s (4 dot-widths per 100ms, therefore ‘streaky’) and 2.15°/s (0.5 dot-widths per 100ms). To measure rivalry suppression depth, monocular probes were presented to the dominant or suppressed eye, using a temporal Gaussian profile to avoid transients (FWHM=60ms). Probes were noise patterns, either orientation-filtered (SD=7.5deg) or iso-oriented, and spatially lowpass filtered to match the motion stimuli. On each trial, the task was to indicate whether the randomly chosen probe was oriented or not. In separate blocks, the oriented probe was parallel to or orthogonal with the motion in the probed eye. Probe contrast was varied adaptively using Quest to find detection thresholds. Results: For fast motion, suppression depth (one minus the ratio of dominance-to-suppression thresholds) was strongly orientation-dependent. Oriented probes aligned with the motion ‘streaks’ were significantly more deeply suppressed than were orthogonal probes. There was no orientation dependency at slow speeds. Overall, suppression depths for static oriented probes in motion rivalry were typical of other rivalry suppression studies (∼0.5 log unit), except for the parallel probe in the fast condition, where suppression was twice as great (∼1.0 log unit). Conclusion: The high-speed orientation dependency suggests fast motions do leave behind neural steaks, due to temporal integration. Thus, the high-speed condition would have involved rivalry between orthogonal motions and orthogonal orientations. The extra suppression in the fast parallel condition is probably due to the combined effects of suppression across both dimensions. Thus, rivalry suppression appears to sum across orientation and motion domains.