When you keep a red apple in working memory, your attention is usually—but not always—attracted by other red objects. The conditions under which the contents of visual working memory guide visual attention are still unclear. Here we tested whether attention is indeed biased toward memory-match stimuli, and, if so, whether this bias is transient or sustained. We used a new pupillometric technique, which exploits that the pupil is larger when you covertly attend to dark, compared to bright,

more »

... i. This allowed us to test whether and when attention was biased toward memory-match stimuli, by placing them either on a dark or a bright background, while measuring pupil size over time. We found that the eyes were captured by memory-match probes early on in the trial. However, the pupillary data suggested that there was no sustained attention bias toward memory-match probes later in time; rather, attention appeared to be biased away from them. Together, our results suggest that: (1) memory-match stimuli automatically capture attention; but (2) they do so only briefly; and (3) they may even become inhibited later in time.