Meeting Banner
Abstract #3585

Whole Brain Parcellation Based on Group-ICA of Tractography Connectivity Maps Shows Differences in Schizophrenia Subjects and Healthy Controls

Lei Wu1, Vince D. Calhoun1, Rex Jung2, Arvind Caprihan1

1The Mind Research Network, Albuquerque, NM, United States; 2Neurosurgery, University of New Mexico, Albuquerque, NM, United States

Whole brain structural connectivity maps based on diffusion imaging data were calculated for a large group of healthy controls and schizophrenia subjects. Connectivity maps were obtained based on probtrackx/FSL algorithm. We develop methods based on group-ICA to parcellate the whole brain into regions characterized by strong connectivity within each region. These regions automatically split the white matter into established white matter tracts and the gray matter into functional modules. This brain parcellation is then used to look for connectivity and the fractional anisotropy differences between HC and SZ groups.

Keywords

able adjusted anisotropy anterior applications applied atlas audience axis back better brain capability characterize clinical cluster clustering clusters collected combine components conjugate connecting connections connectivity contain controls corpus corrected corresponds cortical covariate decreased developing diffusion discriminating distant distinct dominated driven either example excluded extend extracted fibers focus forceps fractional frontiers functional future gives gradient gray healthy hemisphere human identify indicates inferior interactive larger look major manageable mapping maps matrix meaningful meeting might mind modal model mutual nature network networks neurosurgery next occipital overlap paper previously principal priori probabilistic radiation reconstruction reduced reported resolution resting reviewed sampling scanner schizophrenia seem segmentation segmented separate space specified spinal statistical streamline stronger structural subject subjects subset substantially target tend thalamus tighter tract tracts trio typical unlike useful weaker white whole