*2.5. Di*ff*usion Feature Space*

All diffusion features were calculated using DSI Studio (http://dsi-studio.labsolver.org) on processed and normalized diffusion-weighted images. The diffusion information was reconstructed in two fashions using diffusion tensor [47] and generalized q-space imaging [48]. Diffusion tensor imaging (DTI) determines three primary diffusion directions (and magnitudes) using a tensor, from which the standard diffusion metrics fractional anisotropy (FA) and mean diffusivity (MD) were determined. Generalized q-ball imaging (GQI) is a model-free method that calculates the orientation distribution of the density of diffusing water. Using GQI, the non-standard diffusion metrics quantitative anisotropy (QA) and generalized fractional anisotropy (GFA) were determined. A diffusion sampling length ratio of 1.25 was used. The b-table was checked by an automatic quality control routine to ensure its accuracy [49]. Diffusion feature (FA, MD, QA, and GFA) maps were extracted for each subject from normalized diffusion images.
