Working memory maintains perceptual biases during binocular rivalry

Reducing the availability of working memory, by loading it in an unrelated concurrent task, impairs the ability to maintain perceptual selectivity and ignore distracters in visual attention tasks. This suggests a role for working memory in the top-down control of attention. Here, we show that working memory may serve a broader function in the control of visual competition, by examining its role in binocular rivalry. Participants reported their percepts while viewing a typical binocular rivalry stimulus during the retention interval of a working memory task. The working memory task required participants to rehearse a set of auditorilly presented digits, which could be in either random order (high load) or fixed ascending order (low load). We found that loading working memory in this way strongly influenced the dynamics of perceptual alternations in binocular rivalry. Specifically, when binocular rivalry was viewed under high working memory load, dominance phases were shorter and the duration of the initial mixed percept was longer. A control experiment and additional analyses ruled out alternative explanations for our findings, such as reduced sampling duration or a criterion shift under high working memory load. There has been longstanding controversy over whether top-down signals play a role in binocular rivalry. Our findings demonstrate that one type of top-down signal, associated with working memory, plays an important role in maintaining perceptual biases in binocular rivalry. Moreover, taken together with earlier findings showing that working memory maintains selectivity in visual attention tasks, they suggest that working memory has a general influence on competitive interactions in vision, serving the function of maintaining perceptual biases while incoming information works constantly to destabilize them.