The imaging data were mainly processed with Analysis of Functional Neuroimages (AFNI) (Cox, 1996). For each participant, the first four time points were discarded to account for the approach to steady state in the BOLD signal. The raw data were corrected for temporal shifts between slices, corrected for head motion, spatially smoothed with a Gaussian kernel (full width at half maximum = 4 mm) and temporally normalized (for each voxel, the signal of each image was divided by the temporally averaged signal.). We totally scanned 17 HU and 16 CN participants and discarded the data of participants whose head moved more than 1.0 mm in any dimension through the resting-state run. Data of 14 HU and 13 CN participants met the movement criterion and appeared in this study. To further reduce the effect of motion and obtain low frequency fluctuation, we regressed the motion data out of the time series and then preformed band pass temporal filtering (0.01Hz to 0.08 Hz) on the residual signals (Auer, 2008; Birn et al., 2006). These preprocessed time series were used in the subsequent seed-base regions of interest (ROI) correlation analysis.
