Abstract
If a moving object is viewed while the input to one eye is delayed (a temporal disparity), a sensation of depth results (the Pulfrich effect). Under continuous illumination, this is not surprising, since images which arrive simultaneously do have a spatial disparity. However, the Pulfrich effect occurs also during stroboscopic illumination. This is puzzling, because there is no spatial disparity between those matches with the smallest temporal disparity. Modern models postulate that joint encoding of motion and disparity in early processing leads to an equivalence between temporal and spatial disparity. Our recent finding that most disparity-selective neurons in primate V1 do not show this joint encoding led us to reconsider the role of spatial disparities. If matches more separated in time are also allowed to influence perception, then the stroboscopic Pulfrich effect can be explained without invoking joint encoding. This explanation implies that, for short delays, the perceived depth should be smaller in stroboscopic displays than in continuous equivalents. Although this is at odds with previous reports, those may have been affected by eye movements — smooth pursuit can convert temporal disparity into spatial disparity. We performed two different experiments which eliminate this problem. 1. We used brief presentations (160ms). 2. We presented the moving target (for up to 2 s) against a stationary strobed background (with the same temporal disparity). This ensures that tracking eye movements induce the same spatial disparity in both target and background, leaving their relative disparity unchanged. In both cases the depth induced by stroboscopic displays was smaller than for continuous displays, provided the interocular delay is small. This is predicted by our model and in conflict with the joint-encoding hypothesis. We conclude that joint encoding of motion and disparity may not be the main explanation of the Pulfrich effect.