In order to investigate how cyclopean motion is coded by the visual system, the points of subjective equality (PSEs) were measured for (i) speed, (ii) spatial frequency (SF), and (iii) temporal frequency (TF) as a function of peak-to-trough disparity amplitude for cyclopean corrugations. Two panels (3.0 deg × 7.0 deg) of dynamic random-dot stereograms were located 0.5 deg on either side of a central fixation spot. Each panel contained a horizontally oriented sinusoidal cyclopean corrugation whose SF, TF, and disparity amplitude were under experimental control. On each trial, the cyclopean corrugations were displaced vertically in opposite directions. Subjects judged which panel contained the higher SF, TF, or speed depending on condition. The reference stimulus was a sinusoidal corrugation with SF=0.4 cycles deg−1, TF=0.8 Hz, speed of 2.0 deg s−1, and peak-to-trough disparity amplitude of 8 min arc around fixation. We found that, as the peak-to-trough disparity amplitude of the test stimulus increased from 2 min arc to 32 min arc, the PSE for speed decreased from 2.21 deg s−1 to 1.67 deg s−1, compared to a reference speed of 2.00 deg s−1. However, across the same levels of disparity amplitude, the PSE for SF remained constant and the PSE for TF varied but with no consistent pattern. Thus, perceived speed increases with increased disparity amplitude. As all levels of disparity amplitude were above threshold, cyclopean speed cannot be detected by a purely ‘feature-tracking’ mechanism. These metamers and the poor TF matching performance suggest that cyclopean speed is coded by a sparse number of temporal mechanisms.
Salient objects dominate the central fixation bias when orienting toward images
