Perception of binocular 3-D motion: visual direction is more important than binocular disparity

When an object moves in three dimensions (3-D), a combination of binocular disparity and lateral motion (change in visual direction) can, in principle, be used to determine its direction. We know from previous research that observers can be very poor at judging 3-D direction (e.g. Harris, ECVP, 2000). However, we do not know whether the misperception is due to errors in the processing of visual direction or of binocular disparity. Here we tested how well observers can detect different trajectory angles when disparity, or visual direction, are varied separately. We compared two conditions. In the first, the distance moved in depth by a target (with respect to a stationary reference) was held constant at 27.3 min arc (13.2cm in depth). Trajectory angle was varied by changing the extent of lateral target motion (hence changing the final visual direction of the target), to produce trajectories ranging from straight ahead to 20deg to the left or right of the nose. In the second condition, the lateral distance moved was held constant at 1.2cm to the left or right of straight ahead, and the change in depth of the target was varied to create the same range of trajectory angles. Stimuli were presented stereoscopically using stereo shutter goggles running at 120Hz. When depth was held constant, observers were highly inaccurate at perceiving the trajectory angle, but there was a monotonic relationship between physical and perceived angle: wider physical angles were perceived as wider. However, when visual direction was held constant, many observers perceived the whole range of angles as being very similar. The results suggest that observers may be basing their responses primarily on the visual direction of the target, rather than on the relative extent laterally and in depth. Although in principle disparity and lateral position specify 3-D trajectory angle, in practise the human brain may use very much simpler strategies