Fig. 1

Association between diffusion properties of 18 major tracts and covariates. a For 18 major tracts, covariates age, sex, years of education, marriage status, clinical diagnosis, APOE4 genotype, and total brain volume were tested against tract measures that are volume, volume normalized by total brain volume, tract length, axial diffusivity (AxD), radial diffusivity (RD), mean diffusivity (MD), and fractional anisotropy (FA). Rows of the heatmaps represent covariates. Individual association was −log10 transformed and displayed as heatmaps. Columns are organized by tract measure types separated by black vertical lines. For each tract measure type, individual tract associations are arranged in the following orders from left to right: forceps major, forceps minor, left anterior thalamic radiations (L-ATR), left cingulum—angular bundle (L-CAB), left cingulum—cingulate gyrus (L-CCG), left corticospinal tract (L-CST), left inferior longitudinal fasciculus (L-ILF), left superior longitudinal fasciculus parietal (L-SLFP), left superior longitudinal fasciculus temporal (L-SLFT), left uncinate fasciculus (L-UNC), right anterior thalamic radiations (R-ATR), right cingulum—angular bundle (R-CAB), right cingulum—cingulate gyrus (R-CCG), right corticospinal tract (R-CST), right inferior longitudinal fasciculus (R-ILF), right superior longitudinal fasciculus parietal (R-SLFP), right superior longitudinal fasciculus temporal (R-SLFT), right uncinate fasciculus (R-UNC). b Tract measures were adjusted for age, sex, and total brain volume before association testing, and only diagnosis associations remain strongly associated with diffusion-related measures AxD, RD, MD, and FA