Attention to one component of bivectorial transparent motion strongly inhibits the processing of the unattended component

Abstract

Adaptation to bivectorial motion gives rise to unidirectional motion aftereffect (MAE) in the direction opposite to the vector sum of the components. Attention to one component changes the direction of the MAE toward that due to the attended component. A fundamental question about this effect is whether it is due to the enhancement of the attended component, or the inhibition of the unattended component, or both. We designed an experiment to test explicitly the possible inhibition of the unattended component. In each trial, stimuli during adaptation consisted of two populations of dots, moving with the same speed in different directions. Dots in population A moved coherently in one direction. Dots in population B moved randomly most of the time; however, occasionally a proportion of B dots moved coherently for a brief duration as a burst. The proportion of burst dots was varied across blocks of trials. Burst direction was ±90 degrees with respect to A's direction, or ±30 degrees. Observers reported the burst's motion direction in a 2AFC paradigm. MAE strength was measured by the nulling technique during the test phase of each trial. We found that, independently of burst direction, the strength of the MAE is significantly reduced when observers attend to the bursts compared to the passive viewing condition. This reduction is present even when the proportion of burst dots is zero or below detection threshold. In the passive condition the mere presence of the burst had very small effect on MAE strength. These results show that attention to one component of bivectorial motion, in addition to its own enhancement, also results in strong inhibition of the unattended component. Therefore attention-induced changes of the MAE direction are combined effects of enhancement of the attended and inhibition of unattended component.