Caiden Ross-Moody¹ (), Andrew Leber¹; ¹The Ohio State University, Columbus, United States

Our world is filled with visual regularities, as we often navigate the same environments each day. Much research has investigated how we behaviorally exploit these regularities, but questions remain about how we learn to prioritize such regularities. Zhao, Al-Aidroos & Turk-Browne (2013) proposed we are spontaneously biased toward regularities, enabling us to learn whether such regularities can benefit behavior. However, as in our familiar everyday environments, many regularities prove to be uninformative (cf. learned irrelevance). Under these circumstances, we should prioritize novel information. Infants demonstrate preferences for either familiar or novel stimuli depending on their habituation to the former. Here, we investigate if adults flexibly prioritize repeated vs. novel stimuli, based on whether each type of information has the potential to afford a behavioral advantage. In Experiment 1, 50 participants completed a visual search task for a T target among L distractors, with the display split into left and right sides. 8-item configurations in each side repeated throughout, except for occasional (17%) ‘test’ trials for which either the left- or right-side configuration was novel. Critically, unlike standard contextual cueing experiments, no configuration predicted the target location, although participants would presumably attend these configurations to determine their predictiveness. Results showed a regularity bias: critical trial RTs were faster to targets in the repeated than novel display half. In Experiment 2, we attempted to more strongly habituate participants to the regularities by using only one repeated configuration on each display side. Here, we expected participants to learn that repeated configurations clearly conferred no behavioral advantage. Now, results revealed a novelty bias: participants were faster to identify targets in the novel half of the display. Altogether, these results confirm that human adults deploy attention flexibly to either statistical regularities or novel information, depending on the potential behavioral utility of this information.