Figure 1.

Image processing. (A) Imaging protocols included structural images (T1WI and T2-FLAIR images) and dynamic imaging. The time bar shows dynamic imaging intervals during the 25-min acquisition time. (B, C) Examples show the two regions of interest (ROIs) in the white matter (WM) (B), namely, the frontal and parietal WM, and three ROIs in the gray matter (C), namely, the caudate head, putamen, and thalamus. (D) An example shows the perturbation of ROIs. The center and radius of the initial ROI (in red) randomly oscillated within three pixels to obtain another ROI (in blue) that was near to the initially examined ROI (in red). (E) The internal carotid artery was selected and set as the region of plasma in Patlak analysis. (F, G) The K_trans of each pixel was plotted as a histogram. Red bars indicated negative K_trans values and were considered as noise (F). The noise was supposed to exist with the same distribution on the positive side. By subtracting noise (red bars) from the original histogram (blue bars), tissue permeability was observed (black bars in G). x-axis, K_trans ; y-axis, the fraction of voxels. F, frontal; P, parietal; C, caudate head; P, putamen; T, thalamus; Gd-DOTA, gadoterate meglumine; T1WI, T1-weighted image; T2-FLAIR, T2-weighted fluid-attenuated inversion recovery.