fmri data preprocessing preprocessing  (MathWorks Inc)

Journal: Cerebral cortex (New York, N.Y. : 1991)

Article Title: Multiscale functional connectome abnormality predicts cognitive outcomes in subcortical ischemic vascular disease.

doi: 10.1093/cercor/bhab507

Figure Lengend Snippet: Figure 1. The schematic illustration of the main method. (A) The procedures to obtain multiscale FCNs. (i) the fMRI data from one individual is inputted and preprocessed. (ii) The application of the Schaefer’s multiscale atlases to the fMRI data. The black border lines indicate boundaries of ROIs and the colors encode the resting-state network (RSN) to which the ROI belongs. The RSNs include DMN, frontoparietal network (FP), limbic network (LIM), salience network (SAL), attention network (ATT), somatomotor network (SM), and visual network (VIS). (iii) The extraction of ROI-averaged signals. (iv) The construction of multiscale FCNs from individual fMRI data. (B) The architecture for multiscale atlas-based GCN (MAGCN). The multiscale FCNs are extracted via a series of GCNs connected by the APs. The nodal features h are integrated with skip connections and concatenations, based on which individualized diagnosis is generated.

Article Snippet: The sagittal T1-weighted images covering the whole brain were acquired by a 3Dfast spoiled gradient recalled echo sequence: TR = 5.6 ms, TE = 1.8 ms, matrix = 256 × 256, inversion time = 450 ms, flip angle = 15◦, slice thickness/gap = 1/0 mm, number of slices = 156, gap = 0, and FOV = 256 × 256 mm2. fMRI Data Preprocessing We adopt the standardized pipeline from the public available toolbox Data Processing Assistant for Resting-State fMRI (Yan and Zang 2010) in Matlab (Mathworks.

Techniques: Extraction, Biomarker Discovery, Generated