Evaluation of the impact of a virtual self-avatar on gait parameters

Abstract

Studies in the field of gait rehabilitation frequently use treadmills rather than over-ground walking, in part because treadmills require less space and allow for the capture of multiple gait cycles within a limited motion capture volume. Walking on a treadmill alters some kinematic and spatiotemporal parameters, which may be in part a result of the absence of optical flow. To overcome this limitation, some studies have used immersive virtual environments (VEs) to create an optical flow during treadmill gait. Using a head-mounted device, this has the effect of also hiding a user's body from himself. A real-time representation of the user can be displayed in the VEs, in the form of an animated self-avatar. Little is known on the effect of viewing one's self-avatar on gait patterns. This study aims to quantify the differences between walking in a VE with and without a real-time self-avatar. Eleven healthy participants walked in 3 different conditions. In the 1st condition, participants walked over-ground in a hallway. In the 2nd and 3rd conditions (randomized), participants walked in a virtual replica of the hallway, respectively with and without a real-time self-avatar. Through all conditions, participants wore inertial sensors on their pelvis and lower limbs. Participants’ mean cadence and gait cycle duration were computed for each condition. Conditions were compared using paired t -tests. The 2nd and 3rd conditions were compared. The addition of the self-avatar significantly reduced cadence (−1.9 step/min) and lead to a longer gait cycle duration (+0.04 s), even though they walked at the same speed on the treadmill. These results show that the addition of a self-avatar alters spatiotemporal gait patterns. Different potential explanations for these changes will be discussed. Future work will include analysis of the kinematic data for all conditions.