When does a moving object influence the perception of heading?
33.439, Sunday, 18-May, 8:30 am - 12:30 pm, Banyan Breezeway
Oliver W. Layton1,2, Brett R. Fajen1; 1Cognitive Science, Rensselaer Polytechnic Institute, 2Center for Computational Neuroscience and Neural Technology, Boston University
Unlike in a rigid environment, the presence of an independently moving object may bias human heading judgments during self-motion. Existing research suggests that an object moving laterally with respect to the observer’s translation biases heading only when it crosses the observer’s focus of expansion (FoE) (Warren & Saunders, 1995; Royden & Hildreth, 1996), which specifies the direction of travel in the absence of rotation (Gibson, 1950). Bias has only been demonstrated in dot-defined environments that lack depth variation (e.g. frontal wall). We sought to characterize the range of conditions in which moving objects impact heading perception. We investigated how proximity to the FoE and approach trajectory relative to heading of a moving object impacted heading perception in environments consisting of a dot- or texture-defined fronto-parallel plane, a ground plane, or both. Twelve subjects viewed simulated self-motion (±3°, ±9°, ±15°) on a rear projection screen (100°x80°) and indicated their perceived heading by adjusting the position of a post-motion probe. The object (initially 11°x11°) was positioned to occlude the FoE for much (Near condition) or none (Far condition) of the trial. Consistent with Warren & Saunders, we found bias in the direction opposite of the object motion when the object moved toward and crossed the path (~3°; Near condition). Unexpectedly, bias was observed when the object moved close to, but did not cross, the observer’s path (~3.5°; Far condition). Objects initially positioned close to or far from the path that moved away did not yield bias. Our results show that moving objects bias heading even in environments with depth variation and even when the moving object approaches but does not occlude the FoE. The bias is consistent with models that propose a broad spatial pooling of units tuned to motion in area MT of primate visual cortex (Layton et al. 2012).