Abstract
High-level figural processing constrains the interpretation of ambiguous motion under certain conditions (Tse, Cavanagh, & Nakayama, 1998). Transformational Apparent Motion (TAM) occurs when one figure is replaced by another that overlaps its location, causing observers to perceive a shape change of the original object. TAM occurs whether objects are defined by luminance (first-order TAM) or texture (second-order TAM; Hartstein & Tse, VSS 2016). Although the previous experiment found a weaker effect for texture-defined objects, we found no difference in the strength of the illusion between first and second-order TAM when using a dynamic white noise background, indicating strong dependence on high-level figural processing rather than low-level motion energy. While fixating a central point (monitored by an eyetracker), two horizontally or vertically aligned squares were presented to observers. At variable stimulus onset asynchronies, a rectangle appeared connecting the squares. Concurrently, a cue-square appeared adjacent to one of the original stimuli, biasing the direction of perceived motion toward that side (as in Tse, 2006). Stimuli were black on a gray background (first-order TAM) or static white noise on a dynamic white noise background (second-order TAM). The dynamic white noise background allowed shape information to persist between the appearance of the initial stimuli and the rectangle that produced motion perception. Participants indicated whether they perceived motion and in what direction. A compelling TAM illusion was perceived in both conditions, with no difference between first and second-order displays. TAM suggests a figural processing and matching step in the high-level motion pathway that precedes perception of motion. The current result suggests that this figural parsing is sufficient to produce an equally strong illusion without first-order motion signals. Further research explores whether endogenous attention alone is sufficient to provide the figural parsing information necessary for TAM without exogenous cues.
Meeting abstract presented at VSS 2017