Abstract
Humans can determine their self-motion from retinal image motion. The process is not fool-proof however and many conditions create incorrect percepts of self-motion. One problem scenario is when visual rotation occurs around a person's centre of mass during translation. Historically, participants experiencing this have reported impressions of traveling around a curve. It is difficult to measure this however because it requires a simultaneous estimate of both heading and curvature. We have developed a new virtual reality based response tool that allows fast and intuitive reporting of perceived path, and from which perceived heading and rotation may also be inferred. This tool consists of a line projecting from the participant into a virtual world. The line is directed with hand movement and curved using controller input. Eighteen participants viewed 165, 2 second long forward translations (1.5ms-1) over a textured plane while experiencing visual rotation of between 0 and ±7.5deg/s. They used the new tool to indicate their perceived self-motion. Individuals reported consistently curved paths despite moving straight ahead. Greater variation in perceived path across trials was observed however at higher rates of rotation (mean SD curvature (±7.5deg/s) = .055m-1) than lower rates ((±0.6deg/s) = .006m-1). Absolute mean curvature at ±7.5deg/s rotation was .099m-1 (SE = 0.018m-1), consistent with the curvature of the circular path corresponding to 7.5deg/s rotation and 1.5ms-1 translation, although there were significant individual differences. There was also a strong (r = .70) relationship between ability to determine heading and ability to determine one's rate of rotation t(16) = 3.88, p < .001. Finally, there was anecdotal evidence that memory of path decays rapidly, with participants acknowledging resorting to guesswork if they delayed for as little as two seconds when responding. This supports the use of our tool which minimizes the time required to obtain heading and rotation estimates.
Meeting abstract presented at VSS 2018