Abstract
The rod-and-frame illusion, in which a tilted frame causes an enclosed line to appear rotated in the direction opposite the frame, is thought to be driven by two separate mechanisms: 1) a visuovestibular distortion in the observer's perception of gravitational vertical, and 2) low-level orientation contrast effects brought about by mutually inhibitory populations of neurons that encode the respective orientations of the rod and frame. However, these two mechanisms are thought to contribute to the overall perceptual phenomenon of the illusion in a graded fashion, with the visuovestibular and orientation contrast effects somewhat stronger with larger and smaller frames, respectively. In past work (e.g., Dassonville & Williamson, VSS 2010), we have devised two sensorimotor tasks to independently measure these effects. The visuovestibular effect was assessed with the saccade-to-vertical task, with participants asked to make a saccade to the topmost point on a response circle contained within the tilted frame. The orientation contrast effect was assessed with the saccade-to-rod task, with participants asked to make a saccade to the point on the response circle intersected by the rod if it were extended upward. In the current experiment, we verified the selectivity of these tasks by having the participants perform with the head tilted, with the expectation that the tilted head would attenuate vestibular cues and thereby cause an enhanced reliance on the visual orientation cues that specifically drive the visuovestibular effect (see Prinzmetal & Beck, 2001). Indeed, head tilt magnified the illusion-related bias in the saccade-to-vertical task (especially with large frames), but had no effect on the magnitude of the bias in the saccade-to-rod task. These findings confirm that these two tasks can successfully isolate the visuovestibular and orientation contrast effects of the rod-and-frame illusion, and provide further evidence that head-tilt magnifies visuovestibular distortions but leaves low-level orientation contrast effects unaffected.
Meeting abstract presented at VSS 2014