Schedule of Events | Search Abstracts | Symposia | Invited Symposia | Poster Sessions | Data Blitz Sessions

Poster B157

Overlap among neural representations of similar memories triggers repulsion in verbal recall

Poster Session B - Sunday, April 14, 2024, 8:00 – 10:00 am EDT, Sheraton Hall ABC

Anisha Babu1 (, Zhifang Ye1, Brice Kuhl1; 1University of Oregon

Recent fMRI and behavioral studies have shown that highly similar memories can trigger a “repulsion” effect wherein subtle differences between memories become exaggerated. These repulsion effects are thought to be a direct reaction to representational overlap (interference) during learning. Here, we tested this idea by comparing fMRI-based measures of representational structure during the encoding of highly similar natural scene images to verbal recall measures of representational structure after extensive learning. Participants (N=19) learned 18 face-scene associations across six fMRI encoding runs. In order to create interference, scenes were grouped into three categories (e.g., pools, libraries, stadiums) with six exemplars per category. After scanning, participants were cued with each face to recall and type a description of the corresponding scene. For fMRI analyses, we focused on scene-selective cortex (PPA) and hippocampus, and measured representational structure within and across each of the visual categories during encoding. For verbal recall, we used a Natural Language Model to convert each memory into a semantic embedding. Within PPA, there was robust coding of category-level information during encoding, which was positively related to category-level representations during subsequent verbal recall. However, scene-specific (within-category) representational structure in PPA was inversely related to representational structure during recall—consistent with repulsion. In the hippocampus, category representations were weaker than in PPA, but within-category representational structure was more positively related to the structure of verbal recall. These findings provide novel evidence that representational overlap during memory encoding (here, in PPA) triggers repulsion of corresponding behavioral expressions of memory.

Topic Area: LONG-TERM MEMORY: Episodic


CNS Account Login


April 13–16  |  2024