Abstract
Local pairing of opposite motion directions results in a strong inhibition at the motion integration stage and canceling of the global motion percept. However, when motion vectors with directions as far apart as 120 are locally paired, the global motion percept is in the vector average direction and little inhibition has been demonstrated. By measuring the strength of the global directional signal, we tested the integration dynamics of locally paired non-opposite motion vectors. The stimuli consisted of two successive brief motion intervals. The first interval, the duration of which varied, contained two populations of moving dots, one moving in the +60 direction, the other in −60 . In the “transparent” condition, dots were randomly located, generating a percept of two transparent surfaces moving in different directions. In the “locally paired dot” condition, each dot from one population was paired at close proximity with one in the other population, and observers perceived one surface moving along the vector average direction of the two components. In the second interval, the detection threshold was measured for the vector sum direction (0 ) of the two components that appeared during the first interval. In all conditions, the perception of the global direction during the second interval was biased in the direction opposite to the vector sum of the two components in the first interval. When the duration of the first interval was short (120 ms), the amount of the directional bias was significantly smaller in the locally paired dot condition than in the transparent condition, whereas such clear tendency was not observed in the longer duration (750 ms). The results imply that the global integration of locally paired non-opposite motion signals consists of two stages: an early phase where global motion signal is diminished by a transient inhibitory mechanism, followed by a sustained phase where the inhibition is replaced by vector averaging mechanisms.
This work was supported by NIH/NEI grant R01 EY 013758-01. This work was supported by NIH/NEI grant R01 EY 013758-01. This work was supported by NIH/NEI grant R01 EY 013758-01.