Self-motion illusions ("vection") in Virtual Environments: Do active control and user-generated motion cueing enhance visually induced vection?

The human perceptual system can be tricked into believing that one is moving, when in fact, one is not. These self-motion illusions (vection) can be exploited to convincingly simulate self-motion without the need for costly and cumbersome motion platforms. Traditionally, vection has been elicited by moving visual stimuli on custom optokinetic drums or virtual reality (VR) setups. Surprisingly, little is known about contributions of cross-modal effects on vection in contemporary, interactive VR applications. Two studies investigated the effect of active versus passive locomotion and small, actively versus passively generated physical motion cues on optic flow based vection. Twenty four participants used a joystick or gaming chair to navigate on curved (experiment 1, training) or a combination of curved and straight trajectories (experiment 2, main study) presented in an immersive, 3D VR system. The gaming chair allowed for 10 centimeter forward/backward and left/right swivel motions of the seat. Participants experienced four conditions: 1) just watching the scene (passive, no motion cueing), 2) motion cues applied to the participant’s seat (passive, motion cueing), 3) joystick locomotion (active, no motion cueing) and 4) participants using the gaming chair for locomotion (active, motion cueing). Overall, participants took 16% longer to experience vection for active compared to passive locomotion. Small, physical motion cues increased vection intensity by 22%. Trajectory curvature most consistently affected vection. Participants experienced vection 34% more intense, 20% earlier and 9% more likely during narrow turns compared to straight paths. Participants experienced vection up to 18% earlier in experiment 2 over experiment 1 possibly due to training effects. It seems that actively controlling locomotion may have distracted participants from the motion stimulus or the task of reporting vection. It became evident that smoothness, precision and ease-of-use of the interface were possible factors that affected vection. In conclusion, vection can be enhanced by using simple motion paradigms and adding curved trajectories to the simulation at minimal cost and effort. For interactive applications, prudent selection of interaction paradigms and ample training is advised.