Abstract
After adapting to transparent bidirectional motion at different depth planes, one perceives transparent motion aftereffects (MAE) with a test pattern containing the same two depth planes (Verstraten et al., 1994). We investigated whether this motion adaptation is specific to absolute disparity or to the relative depth order of surfaces. Observers adapted to two superimposed populations of dots that were moving in opposite directions (left and right) in different depth planes (approximately −11 and +11 arc min). In experiment 1, the MAE test contained two planes of randomly moving dots presented at various absolute disparities (−29 and −6; −23 and 0; −11 and 11; 0 and 23; 6 and 29 arc min). Observers' task was to report the MAE direction in either the front or back surface, which was cued by a tone. The results indicated that the MAE was influenced by the depth order of surfaces. When the target surface was in front, observers more often reported the MAE in the direction opposite to that of the front adapting surface than the back. Even with the maximum depth difference tested between adapting and target surfaces (18 arc min), this effect was observed in 78% of trials. In experiment 2, we used a more objective method to quantify the MAE dependent on the surface depth order. During the test, the target surface consisted of dots coherently moving orthogonal to or slightly tilted towards either of the adapted motion directions. Observers were asked to report the direction of the tilt. Replicating experiment 1, the tilt of the motion in the front surface was biased towards the MAE of the front adapted surface. When the target surface was presented exactly in-between the two adapting planes (zero disparity), the amount of biased tilt was 1 – 3 deg. MAEs appear to be sensitive to the contingency between depth orders of adapting and test surfaces. The results imply that some motion information is coded with respect to depth-order of surface planes.