Abstract
It has been shown previously that stimulus driven attention can modify the temporal dynamics of binocular rivalry (Ooi & He, 1999). In addition, observers can voluntary control binocular rivalry dominance to some extent (van Ee, 2005). Here we investigate whether the task in which the observer is engaged in could influence binocular rivalry dominance.
To explore this question, we presented binocular rivalrous images composed of eight Gabor patches all oriented to the left in one eye and to the right in the other. On alternating trials, observers either reported the perceived orientation of the Gabors or searched for the location of a Gabor whose contrast was lower. The first task is the traditional binocular rivalry task to measure the dynamics of rivalry. The second task was introduced as an attempt to influence this dynamics. The experiment consisted in four blocks of trials, and unknown to the observers, the monocular target Gabor was always at the same orientation in the middle two blocks. Therefore, should the observer spend more time interpreting the stimulus in the biased orientation, she would increase her chances to detect the target. Thus, if the visual system can determine the usefulness of scene interpretation during rivalry, one would expect a bias in rivalry dominance in favor of the orientation associated with the target. We found that the biased orientation was indeed more often dominant in the first few seconds compared to baseline. This result is remarkable when one remembers that rivalry dominance was measured in absence of the target.
Consistent with this result, the biased orientation sustained over the last block of trials where the target could take either orientation with equal probability. Therefore, our results illustrate task-driven effects on perception, namely that perception is partially determined by how useful the current interpretation is for the task.