Abstract
Computational studies suggest that the visual system uses its knowledge of the recent motion history of objects to help constrain the interpretation of depth-order from motion (DFM). However, it is unclear whether or how the visual system makes use of such information. We characterized the influence of prior temporal information on DFM perception using a priming experiment. Each stimulus consisted of a central strip of random dots moving in depth relative to a larger, outer reference surface, also comprised of random dots. Depending on the stimulus, the central surface appeared to be in the same, nearer or farther depth plane relative to the reference surface. During each trial, a prime stimulus and a target stimulus were presented sequentially for 100 and 50 ms, respectively. The two stimuli were separated by an inter-stimulus interval of 33, 66, 100, 133, or 166 ms, depending on the trial. Subjects reported the perceived depth-order of the target stimulus. We found that the depth-order of the prime stimulus had a significant influence on DFM perception (2-way ANOVA, depth-order of prime x SOA; p <0.05 for both factors and their interaction). As expected, when both stimuli contained the same depth-order information, the perceived depth-order of the target stimulus was consistent with the depth-order expected from the depth-order information in the stimuli. When the target stimulus contained no depth-order information, but the prime stimulus did, subjects perceived the target stimulus to have the same depth-order as the prime, indicating that the prior depth-order information from the prime influenced the depth-order interpretation of the subsequent target stimulus. When the target stimulus supported the opposite DFM percept as the prime, the reported depth-order was intermediate between those expected from either stimulus alone. Together, our results characterize some key properties of temporal summation of DFM information.
Supported by Medical College of Georgia.