Abstract
We have described Tilt-induced Motion (TIM), a variant of the barber-pole illusion, in which a drifting grating is viewed through an obliquely-oriented aperture, inducing an illusory vertical motion in the aperture, even though the aperture is in fact stationary (Caplovitz, Paymer, & Tse, VSS 2007). The effect appears to be strongest when the terminator motion vector has an upward component, and the motion direction orthogonal to the orientation of the grating has a downward component. We raised the hypothesis that conflicting terminator and component motion sources can lead to a misattribution of motion information to the stationary contours of the aperture. We further hypothesized that the direction of the illusory motion is influenced by both local and non-local configural cues.
In the current set of experiments, we tested the first of these hypotheses by removing terminator motion from the stimulus. Here we replaced the drifting grating with a drifting random dot field, thereby eliminating terminator motion entirely. We found that in this case, for certain directions of motion, the TIM illusion persists - thus it does not arise solely from a conflict between terminator and component motion sources. In another set of experiments, we tested the second hypothesis by measuring the perceived direction of TIM under different stimulus configurations. We will discuss the influence that orientation of the short edges of the aperture (which were always vertically oriented in the original experiments) has on the perceived direction of TIM. In addition, TIM can be influenced by using obliquely oriented grey “tracks” that seem to guide the illusory motion. Because there is no real motion along these tracks, they serve as non-local configural cues.