Influence of active control on motion-induced position shift in a 3-dimensional setting

Abstract

It is well known that the perceived location of an object can be considerably influenced by motion signals. The position of a stationary Gaussian window with a locally drifting sinusoidal grating - a Gabor patch, appears to shift in the direction of motion of the carrier stimulus. This robust, illusory phenomenon is referred to as motion-induced position shift (MIPS). It has also been theorized that actions are less prone to visual illusions than are perceptions because the same visual information is coded through anatomically distinct pathways and in a different way for perception as for action. Recent studies have used the motion-induced position shift illusion to explore the effect of active control on the perceived physical position of a target. The results indicated that action systems not only are unable to counteract this form of visual illusion but indeed that the illusionary effect was consistently larger under active conditions. These studies, like much research to date has typically presented the stimuli in a frontoparallel, 2-dimensional plane. To pursue the possibility of creating a more ecologically valid stimulus presentation, this study had the goal to explore the influence of motion-induced position shift during active perceptual tasks in a 3- dimensional setting. For this purpose, a novel game experiment was created using Unity3D to measure and analyse the effect of action under globally moving but locally static or drifting visual stimuli. In one task, participants were required to follow with a drifting Gabor patch a curving line along the floor of a virtual tunnel. The other task entailed participants steering the Gabor to collide with block objects lying along the floor of the tunnel. Clear evidence suggesting an effect of embedded motion on global positional error was recorded in both tasks. The current study suggests that the influence of the illusory position shift due to motion extends to action performed in a 3D setting. Mislocalization of position due to motion is not merely an academic oddity, but can also have real life consequences. Every day, people are required to make judgements about what actions to execute based on what they see and safety research recognizes the import of the influence perceptual visual biases can have on this choice of action in human-machine interactions. This research also adds to the understanding on how these interactions can be designed such that unsafe outcomes can be predicted, avoided or eliminated.