Abstract
Orientation judgments are made within a reference frame that is greatly dependent on the vestibular system, but the visual system is also able to extract contextual cues from a viewed scene (e.g., vertical door frames, horizontal desktops). This becomes dramatically apparent when prominent cues are misleading, as in the case of the rod-and-frame illusion (RFI, Asch & Witkin, 1948), where a tilted frame causes a distortion in the egocentric reference frame, such that an enclosed rod is perceived as being rotated in the opposite direction. Past studies of the RFI have documented the manner in which contextual cues are incorporated into an observer's reference frame, but the time course of this effect remains unclear. To characterize this time course, we used a variation of the RFI which included a temporal mismatch between the onset of a tilted frame (±15°) and the occurrence of a briefly flashed rod (5ms duration, presented in a range of times from 200ms before to 200ms after frame onset). In otherwise complete darkness, participants compared the tilt of the rod to subjective vertical. There was no effect of the frame when the rod was presented well before (200ms) frame onset, and the effect reached a maximum plateau for rods presented simultaneous with frame onset or later. In between these extremes, the effect grew in size, becoming significant even for rods extinguished ~128ms before frame onset. An effect of the frame that precedes the frame's actual onset can be explained by a difference in the processing delays for 1) the orientation judgment task and 2) the context-driven distortion of the egocentric reference frame. The pattern of findings reported here indicate that the orientation judgment takes approximately 130 ms longer than the delay in the egocentric distortion.
Meeting abstract presented at VSS 2018