Abstract
An ellipse rotating at a fixed angular velocity appears to rotate faster as its aspect ratio increases. (Caplovitz et al, 2005;VSS, SFN). We have hypothesized that regions of high curvature act as form-based, trackable features (TFs). When trackable feature are weak, the corresponding motion signals are weak, generating the illusion. Here we examined the perceived angular velocity of rotating ellipses defined by equally spaced dots. The local motion of each dot along the contour is unambiguous. If the dots can serve as TFs, the illusory percept that is observed with continuous contours should not be observed. Methods: Subjects performed 2AFC speed discrimination judgments on pairs of rotating dotted ellipses. 12,24 and 32 dots were used to define ellipses. Control experiments adjusted the size of the dots and the overall sizes of the ellipses. Results: The illusory percept was observed for the 24 and 32 dot ellipses and not for the 12 dot condition. The size of the dots did not affect the perception of angular velocity. Changing the size of the ellipses, and thereby the spacing of the dots, in the 24 dot condition determined whether or not the illusory percept was observed. Conclusion: the spacing and not the number or size of the dots is the critical variable in determining whether the unambiguous local motion of individual dots is used to generate the percept of angular velocity. If the dots are too closely-spaced, the visual system appears to process them as continuous contours, generating the illusion.
Supported by:NIH Grant R03 MH0609660-01 to PUT and an NSF Predoctoral Fellowship to GPC