Kristina Knox^1,2 (), Jay Pratt¹, Jonathan S. Cant²; ¹University of Toronto, ²University of Toronto Scarborough

Our visual environment uses ensemble perception to summarize multiple sources of information into a single summary statistic, helping overcome capacity limitations. Our study explores how ensemble representations guide attention in a subsequent visual search task, and the precision of these representations relative to the mean orientation of a set of items. To explore this, we integrated an orientation-based ensemble task with a visual search task. Each trial began with an ensemble display of eight bars with varying orientations. Then, on a subsequent blank screen, the color of a fixation cross signalled which of two possible tasks —a search or an ensemble task— participants would perform. For the search task (75% of trials, orange cross), participants saw a display of randomly oriented and sized bars and identified whether a T or L appeared inside the shortest bar. Importantly, the target bar’s orientation could either exactly match the average orientation, be within the range, or be outside the range of the initial ensemble display. In the ensemble task (25% of trials, blue cross), participants selected which of two bars matched the average orientation of the initial ensemble display. The results revealed shorter response times (RTs) in the search task when the target bar’s orientation either matched the average orientation or fell within the range of the initial ensemble display, compared to the orientation that was outside of the range. These results suggest that ensemble processing guides attention a visual search task, and does so via precise representations of different summary statistical properties (i.e., the mean and range of a distribution). Together, these findings highlight the prioritization of ensemble representations in everyday behavior (i.e., visual search), and provide insight into the interaction between different perceptual and cognitive systems.