Abstract
Fang & He (2004) used stereoptical depth cues to make the upper and lower thirds of a rotating structure-from-motion cylinder appear to rotate in a single direction. Although the middle section of the cylinder had zero disparity, its rotation direction was stabilized, appearing to rotate in the same direction as the two ends. Adaptation to the stabilized middle section generated a motion aftereffect, inducing a subsequently viewed ambiguous cylinder to appear to rotate in the opposite direction. We asked whether stabilization of the ambiguous middle section occurred merely because it was situated between two rotating objects, or if the shape of the stimulus implied that the three sections were part of the same object. To answer this question, we used more complex shapes than a cylinder to more strongly imply that the middle section fits between the two ends. During adaptation observers viewed one of three rotating shapes (a cylinder, sphere, or cylinder that tapered towards the top). The upper and lower third of each adapting shape consisted of dots drifting in just one direction as if that part of the shape were opaque. In the middle third dots drifted in both directions as if scattered across a transparent object. After 60 seconds the upper and lower thirds disappeared and the middle section continued rotating for 15 seconds, during which observers indicated the apparent rotation direction with key presses. Contrary to our expectations, adaptation was stronger (as indicated by longer rotation duration in the direction opposite to adapting rotation) for the cylinder than for either of the two other shapes. One possible explanation for these results is that the cylinder spanned a larger visual angle than the other shapes so it may have had more stimulus energy than the other shapes.
Meeting abstract presented at VSS 2013