Detecting changes in one's own velocity from the optic flow.

Experiments were designed to establish whether we can use the optic flow to detect changes in our own velocity. Subjects were presented with simulations of forward motion across a flat surface. They were asked to respond as quickly as possible to a step increase in simulated ego-velocity. The smallest change for which subjects could respond within 500 ms was determined. At realistic simulated speeds of locomotion, the simulated ego-velocity had to increase by about 50%. The threshold for detecting changes in simulated ego-velocity was hardly better than the threshold for detecting other changes in the acceleration of the dots on the screen. It made little difference whether the surface across which the subject appeared to move was built up of dots, lines, or triangles; neither did it matter whether subjects saw the same image with both eyes, or whether the simulation was presented in stereoscopic depth. The results show that we are very poor at detecting changes in our own velocity on the basis of visual input alone.