A1283
Title: Sparse independent component analysis with an application to cortical surface fMRI data in autism
Authors: Benjamin Risk - Emory University (United States) [presenting]
Zihang Wang - Emory University (United States)
Irina Gaynanova - University of Michigan (United States)
Aleksandr Aravkin - University of Washington (United States)
Abstract: Independent component analysis (ICA) is widely used to estimate spatial resting-state networks and their time courses in neuroimaging studies. It is thought that independent components correspond to sparse patterns of coactivating brain locations. Previous approaches for introducing sparsity to ICA replace the non-smooth objective function with smooth approximations, resulting in components that do not achieve exact zeros. A novel sparse ICA method is proposed that enables sparse estimation of independent source components by solving a non-smooth non-convex optimization problem via the relax-and-split framework. The proposed Sparse ICA method balances statistical independence and sparsity simultaneously and is computationally fast. In simulations, improved estimation accuracy is demonstrated for both source signals and signal time courses compared to existing approaches. The sparse ICA is applied to cortical surface resting-state fMRI in school-aged autistic children. The analysis reveals differences in brain activity between certain regions in autistic children compared to children without autism. Sparse ICA selects coactivating locations, which is argued to be more interpretable than dense components from popular approaches. Sparse ICA is fast and easy to apply to big data.
