A1186
Title: Accurate individualized functional brain connectivity and topography via ICA with empirical population priors
Authors: Amanda Mejia - Indiana University (United States) [presenting]
Abstract: Independent component analysis (ICA) is often applied to functional MRI data to estimate functional topography and connectivity (FC). However, subject-level ICA results are typically too noisy due to high noise levels to be practically useful. Hierarchical Bayesian ICA models leverage information shared across subjects to improve estimation efficiency, but they have several limitations. Functional connectivity template ICA, a single-subject Bayesian ICA model, is proposed using empirical population priors on spatial topology and functional connectivity between spatial components. These priors can be derived from large fMRI databases or holdout data. Compared with hierarchical models, the proposed approach is computationally convenient, allows for more complex model formulations, and is potentially clinically applicable. This approach is validated through simulation studies and data from the Human Connectome Project and the Midnight Scan Club.
