Abstract
Displaying a succession of independent Glass patterns (dipoles of same contrast polarity, oriented radially, concentrically or linearly) conveys a strong sense of global motion, even though the stimulus itself contains no coherent motion and direction of perceived motion is always ambiguous (Ross, Badcock & Hayes, Current Biology, 2000). Perception of these stimuli could be explained by form integration along dipoles orientation by early visual areas (Smith, Bair & Movshon, J. Neuroscience, 2002) and a subsequent integration along flow trajectories by global motion areas (Krekelberg et al, Nature, 2003). According to this hypothesis, if dots in a single dipole had opposite contrast polarity, perception of motion should be disrupted. We measured thresholds for perception of motion in sequences of independent Glass patterns, where each dipole was made of a black and a white dot, as a function of noise (randomly oriented dipoles). We found that, over a wide range of stimulus parameters such as dot distance, frame duration and dot number, motion sensitivity in this stimulus is similar to sensitivity in glass patterns with both dots of same contrast polarity. We also measured direction discrimination thresholds as a function of signal to noise ratio and found that, in all subjects, there is always perception of motion in the direction from the black to the white dot of the dipole. This effect is stronger for frame durations of about 30 msec and 0.5 deg dot spacing but is still present at larger distances and shorter and longer frame durations. Direction discrimination thresholds are much lower than global motion detection thresholds, suggesting two stages of analysis: an early local motion mechanism which integrates signals over a single dipole and a subsequent level of global motion analysis.