IDENTIFYING ALTERED BRAIN NETWORKS USING DIFFUSION TENSOR IMAGING IN ELDERLY SUBJECTS WITH SUBJECTIVE COGNITIVE IMPAIRMENT

Abstract

Although subjective cognitive impairment (SCI) is widely considered to be an early sign of the prodromal phase of dementia, little is known about how brain networks in SCI elders differ from those in healthy elderly people (NC) and affect their cognitive decline. Therefore, we investigated the changes in brain network organization between SCI elders and NC using network-based statistics (NBS) analysis. We recruited a total of 46 subjects (28 SCI and 18 NC) from Korea University Guro Hospital and performed MRI scanning and neurocognitive tests on all subjects. We used FSL Toolkit (v5.0.9) and Trackvis (v0.6.0.1) for preprocessing and tractography. We coregistered diffusion-weighted images (DWI) with T-1weighted images to define the region of interests (ROIs). We obtained a 90 by 90 connectivity matrix indicating the probabilistic white matter connections in 78 cortical and 12 subcortical regions. Using Brain Connectivity Toolbox, we quantified the local and global properties of the network. We then implemented NBS to identify the different subnetworks in SCI elders and NC. Between SCI and NC, we found a distinct brain network consists of the left lingual gyrus, left supra-orbital frontal gyrus, left fusiform, left middle cingulum, right posterior cingulum, left cuneus, right middle cingulum, and right precuneus. The left posterior cingulum and left precuneus, which were identified as the hub nodes, were affected the most and had abnormal functional connectivity with other edges.