After removing a virtual reality (VR) headset we can be surprised to find we are facing an unexpected direction. In 3 previous experiments we investigated if people could keep track of objects in the real world while in VR. Participants’ task was to point to previously seen targets in the real-world lab while in a virtual replication of the lab. The VR lab could be aligned or misaligned to the real-world lab (pitched up or down). Pointing end points were based primarily on the locations of objects in VR suggesting that people might not update the egocentric reference frame when the environment changes.
In the current study, we investigated if people could keep track of the real-world door while moving in VR. Participants completed 3 conditions: (a) baseline condition where the real and VR labs were aligned, (b) infinite floor condition, and (c) rotation condition with the VR lab yawed by 45°. Participants’ task was to approach and hit randomly positioned spheres in VR and then point to the real-world door.
Performance was accurate in the baseline (1.08° error) and increased when only an infinite floor was visible (12.54° error). In the rotation condition pointing errors were highly affected by the location of the virtual door causing the largest error (37.28°).
While immersed in VR, participants appear to rely primarily on allocentric information to inform them about the real world. This suggests that people might struggle to accurately update egocentric information in the face of conflicting allocentric information.