Abstract
Purpose: The perceived position of the stationary Gaussian window of a Gabor stimulus is shifted in the direction of motion of the Gabor's carrier grating. This illusory position shift increases as a function of carrier speed and reaches an asymptote. We tested whether the asymptotic value of this illusory shift is determined by the precision of relative position processing. Methods: To test this hypothesis, we measured the illusory vertical position shift between two horizontal 1-deg Gabor patches (4 cpd, 50% contrast, 18 arc-min SD) that horizontally flanked a central fixation point with separations of 120 or 480 arc-min. On each trial, the two Gabor carrier gratings moved in opposite directions at equal speed, ranging from 0.0625 to 4 deg/s. Gabor patches were presented for a duration of 453 msec after which observers (N=4) reported whether the left or right Gabor patch was higher. Relative position thresholds were determined also for Gabor patches with stationary carriers, presented in random spatial phases. All data were obtained using the method of constant stimuli. Results: The illusory position shift averaged across observers increased monotonically as a function of carrier speed, reaching asymptotic values of approximately 2 and 6 arc-min per Gabor for gap sizes of 120 and 480 arc-min, respectively. The average relative position thresholds for stationary Gabors were approximately 1.5 and 3.5 arc-min. Conclusions: The results are consistent with the hypothesis that the illusory shift of a Gabor patch with a moving carrier grating results from an inappropriate application of the position-change that is determined by the motion-processing system to the entire patch. However, the magnitude of this inappropriate position change is limited by the tolerance of relative position processing. Our data suggest that the tolerance for motion-induced position change is approximately two times the relative-position threshold of the Gabor stimuli.
Supported by R01 EY05068, R01 EY12810, and R01 MH 49892