Abstract
The ability to localize tactile stimuli on the skin varies with stimulus properties (e.g., strength). We tested whether humans systematically incorporate stimulus properties and prior knowledge when localizing tactile stimuli relative to a visual reference. On each trial, a stimulator in one of five horizontally aligned locations (10 mm spacing) gently buzzed the back of the right hand. Stimulators were hidden from view by a white foam block. Participants fixated a mark projected onto the block (horizontally centered on the right hand, vertically aligned with the stimulators). Consequently, participants received constant visual feedback about the hand's position and shape, features typically misestimated in the absence of vision. After each tactile stimulus a different "ruler" (horizontal row of 5 mm wide colored patterns) was projected onto the block at the height of the stimulators. Participants indicated the perceived location of the stimulus by selecting a pattern. We are interested in localization bias, so no feedback was given. In separate sessions, the stimulator locations were distributed around the center, left or right side of the hand; the fixation mark was always projected on the center of the hand. In two practice blocks, the session-specific spatial distribution of stimulation was learned at the beginning of each session. Two strengths of vibrotactile stimulation were used, chosen randomly. Participants showed a localization bias towards the center of the locations tested in each session. This bias was much stronger for the weaker stimulus strength. There was an additional bias toward the center of the hand (i.e., the fixation mark). Performance was consistent with a Bayesian model of localization combining a prior expectation based on the distribution of stimuli in a session with a likelihood whose variance depended on stimulus strength as well as a bias towards center of the hand (visual fixation).
Meeting abstract presented at VSS 2016