Visual responses to changing size and to sideways motion for different directions of motion in depth: linearization of visual responses.

This psychophysical study explored one possible basis for visually judging the direction of motion in depth. We propose that the changing-size channels precisely compute the algebraic difference between the velocities of opposite edges of a target, thus extracting the velocity component Vz along a line through the eye independent of the trajectory of the target, so that (with added “jitter”) this computation is accurately independent of the component of motion Vx parallel to the frontoparallel plane over a wide range of Vx:Vz ratios. We have no evidence for a complementary channel that computes Vx independently of Vz over any comparable range of Vx: Vz ratios. Our evidence shows that the oscillations of the edges of our stimulus square were equivalent to the oscillation of the square along one of 11 trajectories in depth. All 11 trajectories had the same value of Vz, but the 11 trajectories had different Vx values. In separate experiments, subjects adapted to each trajectory and we measured threshold elevations for two test oscillations, one equivalent to pure z-direction motion and the other equivalent to pure x-direction motion. To a first approximation, threshold elevations for the z-direction test were the same for all 11 trajectories, with the greatest departure from constancy (30%) when two edges of the adapting stimulus were stationary (i.e., equivalent to trajectories that just grazed the eye). Adding an 8-Hz “jitter” oscillation to the 2-Hz adapting oscillation “linearized” the visual response so that threshold elevations were rendered accurately constant (± 5%) for all trajectories tested.