Keynotes | K&Ts | GACs | Talks | Posters | Search

Poster D in Poster Session D: Wednesday, August 5, 2:00 – 3:45 pm, Kimmel Center, Shorin & Rosenthal Rooms

Interactions of Hippocampal Category Learning with Prior Knowledge

Xiaozhi Yang¹, Dhairyya Singh¹, Sudeep Bhatia¹, Anna C Schapiro¹; ¹University of Pennsylvania

Presenter: Xiaozhi Yang

There is increasing evidence that the hippocampus plays a central role in rapid novel category learning, where prior knowledge is unrelated to the new category structure. In natural settings, however, category learning often builds on rich, structured semantic spaces shaped by prior experience. Here we investigated how humans integrate new category information with existing knowledge during learning and provide a hypothesis for how the hippocampus builds from prior knowledge to support this integration. We created categories of known stimuli (e.g. daily foods) that were either congruent or incongruent with the similarity structure of their word embeddings. In simulations using a neural network model of the hippocampus, we demonstrated how different pathways within the hippocampus differentially benefit from congruence. Preliminary results from human behavior were consistent with the model’s account. Together, the work advances our understanding of the neural computations that may support the interaction between prior knowledge and new learning.

Topic Area: Memory, Learning & Knowledge Structures