Tracking the Transition from Stimulus-Specific Object Representations to Category-Level Abstractions During Visual Search

Poster Session A - Saturday, April 13, 2024, 2:30 – 4:30 pm EDT, Sheraton Hall ABC

Ryan S. Williams¹ (ryanscott.williams@mail.utoronto.ca), Joseph M. Saito², Keisuke Fukuda², Susanne Ferber¹; ¹University of Toronto, ²University of Toronto Mississauga

Search performance becomes more efficient with repeated exposure to the same target. Here, we were interested in 1) whether individuals similarly benefit from category-based repetitions of target stimuli when such information is sufficient for visual search and 2) whether this benefit corresponds to a gradual change in the neural representation of such stimuli from stimulus-specific exemplars to category-level abstractions. We developed a category-repeating search paradigm where, on each trial, individuals were cued to encode greyscale images of real-world objects chosen from eight unique categories. Participants were required to search for the cued targets within four-item search arrays. During exemplar search blocks, nontarget items were chosen from the same category as the target, whereas during category search blocks, all nontarget items were categorically distinct from the target. Importantly, for both conditions, the basic-level category of the target images (e.g., birds) was always held constant for five to seven consecutive presentations, while the specific exemplars that defined the targets never repeated within a given run (e.g., robin, finch, sparrow, etc.). As expected, search performance was better overall in the category-search condition relative to the exemplar-search condition. Importantly, this category-search benefit increased linearly with each consecutive category repetition. This behavioral finding was mirrored in our electroencephalographic data, where we used the contralateral delay activity to quantify the stimulus-specific maintenance of the target: less exemplar-specific information was encoded with each category repetition during category search. Thus, we provide a neural index that tracks the movement of target representations from stimulus-specific exemplars to categorical abstractions.

Topic Area: ATTENTION: Nonspatial