Abstract
Tactile stimuli are initially represented with respect to their location on the skin. However, to direct gaze towards the source of a tactile stimulus, its location in external visual space must be determined. That is, body posture has to be taken into account to enable tactile-visual interaction. The ideal observer derives the external location of a tactile stimulus as the vector sum of the location of the touch on the hand and the location of the hand in space. To determine whether humans apply this ideal strategy, participants performed three tasks: 1) Somatic: localization of a tactile stimulus on the hand, 2) Proprioceptive: localization of the hand in space, and 3) Cue integration: localization of a tactile stimulus in space. Somatic stimulation was a gentle buzz in one of 9 locations on the back of the hand. For each stimulus, the hand was passively moved to a new location under computer control. Participants indicated which of two sequentially presented locations (a standard and a comparison, in random order) was further to the right (2IFC, with feedback). For the standard, the hand was centered and the tactile stimulus, if any, was the central buzzer location. For the comparison, hand and/or buzz locations varied randomly over trials. Test locations were 6-24 mm left or right of the standard. In the cue-integration task, a factorial combination of hand and buzz location was used. JNDs in the cue-integration task were slightly worse than predicted from the two single-cue tasks. Analysis of PSEs in the cue-integration task suggests that participants weighted the cues unequally in the vector sum, even though this strategy is disadvantageous. Our results provide the first formal evidence that somatic and proprioceptive cues are combined incorrectly for estimating tactile location in visual space, which is the basis of tactile-visual interaction.
Meeting abstract presented at VSS 2016