Abstract
In the classic ventriloquism effect the perceived location of auditory and visual stimuli are shifted towards each other. Actually cues from almost any modality can be affected by information from other modalities, and the combined perception is often predictable from statistically optimal integration theory in which cues are weighted according to their reliability. Surprisingly, there have been no reports of the perceived location of touches being influenced by the optimal integration of touch with another sense. We therefore measured the perceived location of touch-light pairs. In experiment 1 supra-threshold tactile stimuli were presented either simultaneously with a light (required for the ventriloquism effect) or offset by 500ms (outside the window in which ventriloquism occurs). Subjects fixated a point positioned between two tactile stimuli separated by 7cm. Lights were presented in the same place as the touch or shifted along the arm by 3.5cm. Participants reported whether the light appeared left or right of the touch. If ventriloquism occurred, then subjects should be worse at discriminating the relative positions of the touch and light since they would be pulled closer together. When the lights and touches were presented asynchronously, participants correctly identified their relative positions. However, when they were presented synchronously, participants made more errors in identifying their relative positions suggesting that the stimuli were shifted towards each other or ‘ventriloquised’. In experiment 2 subjects reported the perceived position of a visual, tactile, or visual-tactile pair relative to a visual reference line. While the perceived position of the visuo-tactile pairs was influenced by both component stimuli, surprisingly, the variance of the multimodal responses did not decrease as would be predicted if they were integrated statistically optimally. These results are generally compatible with visuo-tactile ventriloquism. Reasons for the disassociation between the predictability of the position and the variance will be discussed.
This study was funded by a Natural Sciences and Engineering Research Council (NSERC) of Canada to LRH. Vanessa Harrar holds an NSERC graduate scholarship.