Poster B97, Sunday, March 26, 8:00 – 10:00 am, Pacific Concourse
Normal variation in relational memory and pattern separation can be predicted by white matter connectivity
Chi Ngo1, Kylie Alm1, Athanasia Metoki1, Nora Newcombe1, Ingrid Olson1; 1Temple University
To remember the last time you visited San Francisco, you need to discern between memories of similar conferences (pattern separation) as well as remember the relations between different aspects of the conference (relational memory), such as whom you saw and at which venue. Here we asked whether performance variation in pattern separation and relational memory depends on variation inherent in white matter. Healthy young adults learned 64 associations in an AB-AC structure (AB was shown in one context, and AC was shown in another context). Associations were tested by four-alternative-forced-choice. Half of the test trials taxed relational memory the other half taxed pattern separation. Probabilistic tractography on diffusion-weighted imaging data was used to examine structural connectivity between gray matter regions linked to the monkey and fMRI literatures to episodic memory. Consistent with past literature, fornix microstructure significantly predicted relational memory performance, whereas the left uncinate fasciculus microstructure significantly predicted pattern separation performance. Overall performance on both tasks significantly correlated with white matter microstructure connecting the left hippocampus and medial prefrontal cortex. Microstructure of control white matter tracts did not predict any memory measures. These findings indicate that even in healthy young adults, facility with specific aspects of episodic memory can be predicted by examining white matter in a fronto-temporal memory network.
Topic Area: LONG-TERM MEMORY: Episodic