Abstract
It is well-known that the coherent displacement of a sector of a random-dot cinematogram results in the perception of motion in the direction of the displacement (the short-range motion effect; Braddick, 1974; Lappin & Bell, 1976), and that motion in the direction opposite to the displacement, so-called reverse-phi motion, is perceived when the displacement is accompanied by the reversal of luminance polarity (Anstis, 1970; Sato, 1989). These motion phenomena have commonly been attributed to the extraction of motion energy, with the further implication that figure segregation and the perception of shape-from-motion likewise depends on motion energy extraction (Dosher, Landy & Sperling, 1989). Our psychophysical and computational results cast doubt on the role of motion energy extraction in this paradigm. We have confirmed, first of all, that the discrimination of both motion direction and figure shape are better in the standard than the reverse-contrast conditions. This asymmetry is inconsistent with the elaborated Reichardt detector (van Santen & Sperling, 1985), which predicts equally strong and coherent motion in the standard and reverse-phi directions. In contrast, the counterchange detector, which signals motion through the detection of oppositely-signed changes in contrast for pairs of edge detectors (Hock, Schöner & Gilroy, 2009), is sensitive to the asymmetry in standard and reverse-phi motion. Both trial-by-trial and full-experiment simulations indicate that a motion architecture based on the counterchange principle is sufficient to plausibly account for the observed asymmetries in both motion direction and shape discrimination, including the perception of motion in the reverse-phi direction. In conclusion, counterchange detection rather than motion energy extraction is the likely basis for the short-range motion effect, and for figure segregation and the recovery of shape from coherent motion.
Meeting abstract presented at VSS 2013