Poster Session C, Sunday, March 24, 5:00 – 7:00 pm, Pacific Concourse
Influence of genetic relatedness on fMRI activation pattern similarity during the HCP working memory task
Joset A. Etzel1, Maria Z. Gehred1, Arpana Agrawal1, Todd S. Braver1; 1Washington University in St. Louis
Previous studies have investigated genetic influences on fMRI activity patterns with twin designs, but have not typically considered task context effects. Here we leverage the large sample and family structure of the Human Connectome Project (HCP) dataset to examine individual differences and genetic influences on activation pattern similarity during the Working Memory (N-back) task, specifically contrasting load-based coding (0-back, 2-back) against the coding of picture category (Face, Place), in two ROIs: the FrontoParietal and Visual Communities defined by the Gordon parcellation. HCP participants were chosen to form 105 MZ (monozygotic) twin pairs, 78 DZ (dizygotic) twin pairs, 99 non-twin sibling pairs, and 100 unrelated pairs. The sixteen condition-wise correlations of the pairs’ activation patterns were computed for each Community, with Load and Category coding quantified from the resulting similarity matrices. As expected, there was strong regional specificity: Load was selectively encoded in FrontoParietal, and Category in Visual. Likewise, there was a strong influence of genetic relatedness (MZ > DZ and siblings > unrelated), but only for Load in FrontoParietal and Category in Visual. In unrelated people, similarity was greater in Visual than FrontoParietal, suggesting its organization may exhibit more inter-individual consistency. Strikingly, higher Load quantification scores predicted better task behavioral performance (d’), but only in related individuals, supporting their functional importance. These results suggest that robust genetic influences can be detected in fMRI activity patterns, but that task context critically determines their anatomical locus. Additionally, they highlight the power of pattern similarity analyses for detecting key frontoparietal cortex coding features.
Topic Area: METHODS: Neuroimaging