The representation and retrieval of general versus specific category knowledge

Poster Session C - Sunday, April 14, 2024, 5:00 – 7:00 pm EDT, Sheraton Hall ABC

Marlie C. Tandoc¹ (tandoc@sas.upenn.edu), Sarah H. Solomon¹, Jacob A. Parker¹, Alex Gordienko¹, Anna C. Schapiro¹; ¹University of Pennsylvania

Building useful knowledge requires representing both shared and unique features of the elements of our environment (most dogs bark; your friend’s dog bites), and we are able to flexibly retrieve the kind of feature most relevant to our current goals (particular caution around your friend’s dog). How does the brain accomplish these feats of representation and retrieval? We use high-resolution fMRI to identify how shared and unique features are represented after category learning and during different memory-based decisions. Participants learn categories of flowers, where each flower has some petals shared with same-category exemplars and one petal that is unique to that exemplar. We assess how shared and unique petals are neurally represented immediately post-learning as well as during a retrieval task in which participants are cued with a category or exemplar label and indicate whether a flower presented at a delay matches the label. Data collection is ongoing (N = 23 of 30), with preliminary analyses suggesting that distinct regions are recruited during successful retrieval of category information (anterior hippocampus, medial prefrontal cortex) versus exemplar information (lateral parietal cortex). These findings are consistent with recent work uncovering prototype and exemplar representations in these networks (Bowman et al., 2020, eLife). Additional analyses will characterize the representational similarity of both item and feature-level representations, and how these representations are recruited during retrieval processes at different levels of abstraction. The approach will help us better understand how the brain builds knowledge structures that support flexible memory-based decisions.

Topic Area: LONG-TERM MEMORY: Semantic