Signa hdxt mr

All participants were scanned using a 3.0T MR scanner (Signa HDxt MR, GE Healthcare, Milwaukee, WI). During scanning, foam pads and earbuds were used to reduce head motion and scanner noise respectively. Participants were required to keep still with their eyes closed. Diffusion-weighted images were obtained using a single-shot echo-planar imaging sequence according to the following parameters: repetition time (TR) = 13,000 ms; echo time (TE) = 85.9 ms; number of excitations (NEX) = 1, field of view (FOV) = 256 × 256 mm²; matrix size = 128 × 128; slice thickness = 3 mm; 32 non-collinear diffusion directions with a b-value of 1,000 s/mm² and one additional volume without diffusion weighting (b = 0 s/mm²) were acquired; and 50 transverse slices without gaps, covering the entire brain. We also acquired high-resolution 3D brain anatomical images using a T1-weighted BRAVO sequence according to the following parameters: TR = 6.8 ms, TE =2.5 ms, flip angle = 9°, slice gap = 0 mm, turnover time (TI) = 1,100 ms, NEX = 1, FOV = 256 × 256 mm², matrix size = 256 × 256, and 192 contiguous sagittal slices with slice thickness = 1 mm.

All the subjects were scanned using a 3 T MR scanner (Signa HDxt MR, GE Healthcare). A preprocessing approach similar to that in our previous studies was used^{30 (link),31 (link)}. Details of MRI data acquisition and preprocessing are described in the Supplementary Information.
To generate the whole-brain functional connectome, we used a previously established functional parcellation of the cerebral cortex and striatum to decompose the whole brain FC into 7 resting-state networks, namely VN, SMN, Limbic, ECN, DAN, SN and DMN (Supplementary Fig. S1)^{32 (link),33 (link)}. Across all the 7 sub-networks, there are 132 separated anatomical regions of interest (ROIs), which were then used to represent nodes in FC networks. The functional connection between any two nodes i and j was defined as the Fisher-z transformed Pearson product-moment correlation of the averaged blood oxygen level-dependent (BOLD) time courses within these regions. The averaged BOLD time series of an ROI was obtained by averaging the time series of all the voxels in this ROI. The Pearson’s correlation coefficients were then calculated between each pair of ROIs. To improve the normality of the correlation coefficients, Fisher’s r-to-z transformation was performed to convert the correlation coefficients to z-values (see Supplementary Information).