Abstract
Classical or long-range apparent motion is sometimes considered as a first-order motion (Cavanagh & Mather, 1989), but sometimes as an attention-based phenomenon (Horowitz & Treisman, 1994). This confusion partly resulted from the lack of detailed knowledge about spatio-temporal characteristics of classical apparent motion. Thus, in this study, we investigated its spatio-temporal characteristics especially with respect to horizontal and vertical meridians. The stimulus consisted of two white discs (3.8-deg diameter) successively presented against dark gray background. Combinations of motion-directions (vertical/horizontal) and within/across hemifields (right/left or upper/lower) resulted in four conditions. All motions were presented along the horizontal or vertical meridian near the fixation point. The spatial distance was varied in 3 to 6 steps depending on the condition between 15 and 135 deg. The duration was 33ms for all stimuli, and ISI was varied between 0 to 533ms in 7 steps. The observer’s task was to judge whether motion was perceived. For within-hemifield motions, motion was perceived nearly 100% from 0 to 300ms ISIs and declined beyond this point regardless of spatial distances and motion directions. These results might suggest that these motions are mediated by local-motion detectors, but it should be noted that the largest distance here was 60 deg, much larger than ordinary receptive-field sizes. As for inter-hemifield motions, the percentage declines at shorter ISIs and it peaked around 100ms for all stimuli except for the smallest distance where no decline on the shorter side was acknowledged. In summary, these results suggest that classical/long-range apparent motion is mediated by at least two separate mechanisms; most likely a first-order-like mechanism for within-stimuli, and an attention-based mechanism, or another passive mechanism that is susceptible to ISI for across-stimuli. In addition, it is quite intriguing that almost identical inter-hemifield effects were found for upper/lower as well as right/left hemifield.
Meeting abstract presented at VSS 2013