| The current studies examined the use and capacity of memory for rejected distractors during visual search. In each experiment, participants' eye movements were recorded while they searched for a target. In Study 1, we examined the extent to which people use memory for rejected distractors to avoid reinspecting them in four experiments (search for a toy among other toys, an O among Q-like distractors, Waldo in Where's Waldo scenes, or a tank in fully realistic landscapes). In each experiment, we compared distractor refixation rates when fixated objects (Experiments 1 and 2) or display locations (Experiments 3 and 4) were explicitly "marked" (e.g., by changing each fixated object's color) to both a no-marking version of the same task and to a memoryless random-sampling model. Overall, we found that distractor refixation rates were low across all four experiments, tended to be comparable with and without an explicit marker, and tended to be lower than the memoryless model. In Study 2, we examined the capacity of distractor memory by following each trial in a search for a right- or left-gapped gray box among top- and bottom-gapped gray boxes with a spatial memory probe. A red box was presented at either a previously fixated location or a non-fixated location, and participants rated their confidence that the target appeared at the probed location on a 9-point Likert scale. Probes at fixated locations could appear up to 13 objects back in the participant's fixation history. In Experiment 6, probes appeared on an object-based representation of search items' locations; in Experiment 7, probes appeared on a blank background. In both experiments, participants were able to discriminate among fixated and non-fixated locations throughout most of the probed fixation history. In addition, control experiments in each study showed that participants were able to avoid fixating marked yet-to-be fixated items and that participants' search behavior was not affected by the memory test in Study 2. Taken together, these studies suggest that people have a relatively high-capacity memory for rejected distractors that is used to improve search efficiency by preventing distractor refixations. Implications for visual search theories are discussed. |
