Direction discrimination of cyclopean (stereoscopic) and luminance motion

This study compared direction discrimination of cyclopean (stereoscopic) and luminance motion involving stimuli equated for effective strength. The stimuli were random-walk cinematogram (RWC) displays whose signal and noise discs were created from binocular disparity differences embedded in a dynamic random-dot stereogram or from luminance differences. Experiment 1 measured global motion detection thresholds for cyclopean and luminance stimuli by manipulating the proportion of signal to noise discs. Detection thresholds for cyclopean motion were about 25% whereas detection thresholds for luminance motion were 5%, thus five times more cyclopean motion events than luminance events were necessary to elicit threshold responding. Experiment 2 measured thresholds for discriminating the direction of cyclopean and luminance motion under conditions of equal stimulus strength by presenting the motion displays at equal multiples of detection threshold. Direction discrimination thresholds (ranging from about 5-30 deg, depending upon conditions) were similar for cyclopean and luminance motion, thus the precision with which the pooling of local motion events in one direction can be discriminated from the pooling of events in a slightly different direction is the same for cyclopean and luminance stimuli. The finding that cyclopean motion information is pooled is consistent with the idea that the direction of cyclopean motion is coded in the responses of a population of directionally selective mechanisms.