Sensory conflict alters visual perception of action capabilities during crossing of a closing gap in virtual reality.

The somatosensory, vestibular, and visual systems contribute to multisensory integration, which facilitates locomotion around obstacles in the environment. The joystick-controlled virtual reality (VR) locomotion interface does not preserve congruent sensory input like real-walking, yet is commonly used in human behaviour research. Our purpose was to determine if collision avoidance behaviours were affected during an aperture crossing task when somatosensory and vestibular input were incongruent, and only vision was accurate. Participants included 36 young adults who completed a closing gap aperture crossing task in VR using real-walking and joystick-controlled locomotion. Participants successfully completed the task using both interfaces. Switch point between passable and impassable apertures was larger for joystick-controlled locomotion compared with real-walking, but time-to-contact (TTC) was lower for real-walking than joystick-controlled locomotion. Increased joystick-controlled locomotion switch point may be attributed to incongruency between visual and non-visual information, causing underestimation of distance travelled towards the aperture. Performance on future VR applications incorporating dynamically changing gaps can be considered successful using joystick-controlled locomotion, while taking into account a potential behaviour difference. Differences in TTC may be explained by the requirement of gait termination in real-walking but not in joystick-controlled locomotion. Future VR studies would benefit from programming acceleration and deceleration into joystick-controlled locomotion interfaces.