Abstract
Previous studies have found that adapting to orthogonal transparent motion patterns yields an integrated motion aftereffect (MAE) opposite to the average direction of the transparent motion. Here we demonstrate a new phenomenon of segmented MAE opposite in direction to just one of the transparent components, and investigate the mechanisms that underlie the segmented and integrated MAEs. The 24-degree-wide adapting stimulus contained 396 randomly-oriented, equally-spaced Gabor elements. Half of the Gabors were assigned a global-motion direction of X−45° (pattern 1) and the other half X+45° (pattern 2), where X is the integrated direction of the two pattern vectors. After 45s of adaptation, observers reported the MAE direction when tested with static Gabor elements shown at different locations. In Experiment 1, when test Gabors were presented at locations from both adapting patterns, an integrated MAE was found, with the aftereffect direction opposite to the integrated direction (i.e., X−180°). However, when all test elements were at locations from one adapting pattern (e.g., pattern 1), a segmented MAE was obtained, with the aftereffect direction opposite to that pattern's adapting direction (e.g., X+135°). MAE changed from segmented to integrated when test orientations were orthogonal to adapting orientations (Experiment 2), and when test Gabors were presented at “phantom” locations, at which no adapting elements had been presented (Experiment 3). The results of Experiment 1 imply that both segmented and integrated motions are represented during transparent motion adaptation, as different test stimuli can reveal either form of MAE. The reemergence of integrated MAE in Experiment 2 demonstrates a dominant role of integration processing when local adaptation was weakened. The “phantom integrated MAE” found when local adaptation was eliminated in Experiment 3 provides further support for this interpretation. Our findings demonstrate the interaction between two distinct motion-adaptation mechanisms, which may be crucial when adapting in dynamic environments.
This research was supported by NSF grant BCS-0843880 and UCLA Faculty Research Grant.