Abstract
Is depth perception from binocular disparities—stereopsis—slow or fast? If stereopsis is slow, as current dogma suggests, then its utility for everyday visual function is limited. Previous studies showed that disparity is quickly extracted from a stimulus but that rapidly alternating disparity is hard to see. However the processing speed of stereopsis after presentation of stereo images has never been measured. To measure this latency and compare it to the latency of brightness processing we used a deadline procedure. Observers responded in a stereo-based or luminance-based 2-alternative forced choice task before a given deadline that varied from block to block. In the stereo-based conditions one disk with 5 arcmin of crossed disparity was displayed on one side of fixation and a disk with uncrossed disparity was displayed on the other side. In the luminance condition one disk was black and the other one gray. We then estimated the difference in latency for the two different signals by fitting Speed-Accuracy Tradeoff Functions (SATFs) to the data. The two SATFs typically deviated from chance levels of accuracy at the same time. Thus, stereoscopic depth becomes available at the same time as apparent brightness in perceptual tasks. Stereopsis could therefore be more useful than previously thought. In addition, because binocular disparities are available quickly, they may participate in other visual processing tasks. Such broad utility would help explain why disparity-sensitive cells are ubiquitous in visual cortex.