Abstract

Successive presentations of two different stimulus configurations were used to generate a compelling but ambiguous apparent-motion sequence, such that the likelihood of the perceived motion being resolved as clockwise or anticlockwise was equal. In the first stimulus configuration, four small squares were arranged in a square formation (5 deg × 5 deg) about a central fixation mark; in the second, they were arranged in a diamond formation. Twelve naïve observers reported perceived direction of motion in discrete presentations (< 1 s), comprising three stimulus configurations (square - diamond - square; stimulus configuration duration 100 ms; ISI 200 ms). We were surprised to find that ten of the observers exhibited a strong bias for motion in one or other direction. This bias to report a particular direction persisted even when observers were informed of the inherent ambiguity of the stimulus. Next, in extended presentations (2 min) five observers indicated each occurrence of a perceived switch in direction. Despite individual variation, all observers reported switches in direction; an analysis of the interswitch durations conformed to a gamma distribution. These observations support the notion that uninterrupted presentations are required for perceptual switching to occur in ambiguous displays (Leopold et al, 2002 Nature Neuroscience 5 605 - 609).