Head position affects touch localization on the torso in different directions depending whether the head returns to centre or remains eccentric during testing. These observations have been taken to suggest that touch is coded in either gaze-centered or body-centered reference frames depending on the task (Pritchett et al., 2012). Here we tested localization of tactile stimuli applied to the back using the same paradigms to indicate gaze-centered or body-centered coding. Tactile stimulation was applied using a horizontal array of eight tactors mounted on a belt, centred on the backbone. Participants reported the perceived location of touches on a visual scale presented on a screen. The head was held eccentrically during the stimulation and either returned to centre or remained eccentric during response. When the head was returned to centre, as for the touches on the front of the torso, perceived touch locations on the back were shifted in the head’s direction, with largest effects on the side to which the head was turned. When the head remained eccentric, perceived touch location was unaffected by head position. Returning the head to centre evokes a gaze-centred strategy with errors related to errors in coding gaze. Leaving the head eccentric is associated with a body-centred coding strategy. The lack of errors on the back, compared to on the front of the torso under this condition may indicate a body-centred representation of the back that is more stable than the front – perhaps related to the availability of the backbone as a prominent body landmark.

Touch localization; reference frames; body representation