Meeting Banner
Abstract #3284

Extracting Connectomic Profiles from Group Resting State fMRI Data Using Dictionary Learning

Kaiming LI1, Xiaoping P. Hu1

1Emory University, Atlanta, GA, United States

This paper describes a new framework to characterize the connectomic profiles for distinct functional regions on the cortical surface. Unlike existing group ICA approaches that heavily rely on spatial smoothing and registration techniques, this framework employs two measures, cortical parcellation by BOLD signal homogeneity and over complete dictionary learning, to account for the well-known anatomical variability across individuals. Our results show that the resultant connectomic profiles are robust and can be used for the identification of both distinct functional regions and functional networks, facilitating building statistical models for these profiles and pinpointing disrupted regions in pathological/psychiatric brain disorder datasets.

Abstract PDF Presentation Video

Keywords

according account accurately achieve affinity alleviate almost although anatomical another anterior applications approaches arrows bilateral biomedical bold brain building central characterization characterize clinical closely clustering complete component connectivity constraints contains correspondence corresponds cortex cortical cortices creates cuts datasets depict derived described dictionary difficulty disorder disrupted distinct drawn effort element entire examples excluded extracting facilitating factorization feature five form framework functional generated generation highlighted homogeneity identification imperfect imposed impossible includes increasing independent indicate individuals inferior influence intelligence intensity journal learning like literature loadings lobules machine matrix maximal measures messages mode models moderate negative network networks normalized onto orange parietal particular passing pathological pattern patterns performance pinpointing pipeline poor positive preprocessed probabilistic profile profiles project propagation psychiatric quantitative recent recognized regional registration released rely respectively resting resultant science segmentation since smith smoothing sparsity spatial steps strong strongest studies subject subjects summation superior surfaces taking technology template temporal thank towards transactions unlike variability white