Two Stages for Depth Integration of Motion Parallax

Abstract

Motion parallax has been shown to be a principal cue for depth perception under monocular viewing. The simulated depth of stimuli in previous studies has been constant in both magnitude and direction. In the present study we addressed the question how the visual system detects parallactic depth change. To answer this we investigated the temporal characteristics of parallactic depth change and the effect of a motion signal on them. The stimulus consisted of four bands of 15-cycle sinusoidal gratings and parallactic depth was simulated between each band. In experiment 1, we measured the amount of perceived depth change with different frequencies (0.125 to 10 Hz) of simulated depth change and with different velocities (2.5 to 40 cm s−1) of head movements. The result showed the perceived depth change decreased with frequency of depth change, and it increased with head velocity when the frequency was constant. In experiment 2, we measured the motion threshold with different velocities of head movement. The result showed the threshold was constant across different head velocities. In experiment 3, we measured the amount of perceived depth using apparent motion stimuli with the head moving. The result showed depth decreased with SOA of apparent motion stimuli, but there was no effect of different head velocities. The results of these three experiments indicate that parallactic depth change is determined by the duration of simulated depth, which corresponds to the integration time of motion, as well as by the extent of head movement. We conclude that parallactic depth is integrated in two stages: first, integration of motion and, second, integration of motion parallax.