The Desikan–Killiany atlas was also used to calculate mean ASL perfusion for each lobe. Thus, SPM5 tissue segmentation (CT99021 clinical trial Ashburner and Friston 2008; Ashburner 2009) was

applied to the T1-MPRAGE data acquired during the same scanning session as the perfusion acquisitions, generating gray matter and white matter posterior probability maps for each participant in native space. The posterior probability maps were then thresholded using a minimum probability of 0.70, minimizing partial volume effects for each tissue type, yielding a binary gray matter mask and a binary white matter mask. The T1-weighted anatomical acquisition was processed using FreeSurfer reconstruction Inhibitors,research,lifescience,medical (Dale et Inhibitors,research,lifescience,medical al. 1999; Fischl et al. 1999), which generated separated left and right cerebral hemisphere cortical ribbon masks and cortical parcellation using the Desikan–Killiany atlas for each participant. Left and right

masks were combined to form the cortical ribbon mask. The whole brain geometry for each participant’s mean perfusion data was established by concatenating the inferior 9 axial slice and superior 9 axial slice relative CBF maps generated by the scanner, along the slice (z) direction using Analysis of Functional Neuroimages (AFNI) (Cox 1996). The AFNI MATLAB library, freely available from http://afni.nimh.nih.gov/afni/matlab, was used to convert Inhibitors,research,lifescience,medical each whole brain perfusion array into an AFNI-compatible 3D format, having the same geometry as the whole brain relCBF dataset. The FreeSurfer cortical ribbon, anatomically based cortical parcellation (Desikan–Killiany atlas) and binary masks, were then aligned with and resampled to the same geometry Inhibitors,research,lifescience,medical as the perfusion data using AFNI/SUMA (Surface Mapping with AFNI). A whole brain perfusion map was then created using the following formula: Whole brain perfusion = (binary gray

matter mask + binary white matter Inhibitors,research,lifescience,medical mask) × (perfusion data). Alignment verification of the cortical ribbon mask, cortical parcellation, whole brain mask, and whole brain perfusion map in 3 mm × 3 mm × 6 mm space was done others for each participant using the AFNI viewer. Following alignment and resampling, the mean and standard deviation (SD) of all perfusion values between 1 and 100 was calculated for each region of interest (as outlined in the Desikan–Killiany atlas), the cortical ribbon, and the whole brain. Left and right hemisphere frontal, temporal, parietal, and occipital lobe perfusion composites were also computed that consisted of the sum of the perfusion values of the respective regions of interest for each lobe. Morphometric analyses For morphometric analyses, T1 volumes were segmented into gray, white, cerebrospinal fluid, and nonbrain tissues using the FreeSurfer software package.