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

Ensemble processing plays an important role in our daily lives by condensing abundant visual information in our environment into statistical representations. Our study examined how these statistical representations are prioritized in the attentional system by asking whether ensemble representations, such as the average orientation of a set of items, can guide attention in a subsequent task. To explore this, we integrated an orientation-based ensemble-processing task with a visual search task. On each trial, participants were shown an initial display of eight bars of varying orientations. The subsequent task—either a search or an average task—was signalled by the colour of the fixation cross. When the cross turned orange (25% of trials), participants engaged in the search task. They had to locate and click on the shortest bar among six others displayed around the fixation point. Importantly, in half of these search displays, the target bar matched the average orientation of the initial eight-bar display. When the fixation cross turned blue (75% of trials), participants performed the average task. This task involved a display of two bars to the left and right of the fixation point, and participants had to determine which of these two bars corresponded to the average orientation of the initial eight-bar display. In both tasks, participants were instructed to respond as quickly and accurately as possible. The results revealed shorter response times (RTs) in the search task when the target bar matched the average orientation of the initial eight-bar display compared to when the orientation of the target bar did not match the average orientation of the initial display. On a local level, this finding indicates that ensemble representations guide attention in subsequent tasks. On a global level, this means that the representation of an item never explicitly perceived can guide attention and subsequent behaviour.