Technology has come a long way since Stratton (1889) who built a rudimentary yet innovative setup allowing him to watch his body from the viewpoint of an other person (third person perspective or 3PP) while he was walking. More recently, behavioral studies have shown that bodily processing can be manipulated using variations of a Strattonian 3PP setup, but in standing or sitting observers (Ehrsson, 2007; Lenggenhager et al., 2007, 2008). Here, we present a novel set-up (and experimental data) allowing the presentation of stimuli during locomotion (as well as afterwards) from a 3PP or the habitual-first person perspective (1PP) of the participant.

The present experiment modified a virtual reality (VR) set-up (Salamin et al., 2006) in order to investigate the range and malleability of peripersonal space following prolonged exposure to either 3PP or 1PP, while participants wore the experimental setup and walked freely for 15 minutes. The device consisted of a rigid backpack, a camera and a head mounted device (HMD) to which the video captured from the camera was relayed in real time. In order to estimate the range of peripersonal space (defined here as grasping space; i.e. Grüsser and Landis, 1991) we asked the participants immediately after 1PP or 3PP exposure to judge the interceptability of a ball projected on the HMD (ball catching task). On the HMD, a ball originated from a fixed point and travelled towards the participant at different trajectories (final horizontal distance from the participant was 20cm, 60cm, or 150cm). Participants were asked to respond if they could catch the ball or not through a serial response box. Trajectories and perspectival conditions were counterbalanced across participants.

Analysis of the mean number of “caught ball� responses of 11 participants across different conditions showed an expected main effect between different stimuli (20cm, 60cm, 150cm; p<.001) with decreasing number of catch responses for increasing final distances. Yet, we also found a significant perspective x stimulus interaction (p<.001) based on 1PP-3PP differences only for the intermediate final distance of 60cm. These data reveal a functional extension of peripersonal grasping space based on perspectival cues during the free walking phase which was unrelated to the task. We discuss our data with respect to the more well know phenomenon of prism adaptation, multisensory and sensorimotor coding in peripersonal space, and observations in neurophysiology and neuropsychology.