Schedule of Events | Symposia

Memory reactivation during sleep facilitates abstraction of category structure

Poster Session C - Sunday, March 30, 2025, 5:00 – 7:00 pm EDT, Back Bay Ballroom/Republic Ballroom

Sarah Solomon^1,2 (ssolomon4@binghamton.edu), Siri Krishnamurthy², Anna Schapiro²; ¹Binghamton University, ²University of Pennsylvania

Successful learning involves the generalization of details into an abstracted format that can be applied to new situations, and sleep has been hypothesized to play a role in this abstraction process. We conducted a test of this idea in the semantic learning domain. Participants learned about novel insect categories with different structures (different patterns of feature co-occurrence) in a missing feature task. The categories all had non-overlapping sets of surface-level features, but the structure of the initial categories was either congruent or incongruent with the final target category. Structure knowledge of the final category was tested using a visual feature sorting task. Experiment 1 tested for immediate structure transfer to the final category with no delay. Experiment 2 tested for structure transfer after a period of awake rest, a nap, or a nap during which the memory of a recently learned category was reactivated. Memories were reactivated during sleep using targeted memory reactivation (TMR): sound cues that were paired with a category during learning were replayed during sleep. Across the two experiments, we only observed successful transfer learning when the initial congruent category was reactivated during sleep using TMR. The finding that category structure transfer is possible at all suggests that learned category structure can indeed become abstracted away from surface-level features. The finding that this transfer only occurs after related memories are reactivated with TMR suggests that disentangling structure from surface-level details is not automatic but rather may rely on reactivation-related processes during sleep.

Topic Area: LONG-TERM MEMORY: Semantic