High resolution research tomograph scanner

Dynamic PET scans were acquired on the High Resolution Research Tomograph scanner (Siemens/CTI, Knoxville, TN, USA) for 60 min starting with [¹⁸F]-FE-PE2I bolus injection over 1 min. After reconstruction and motion correction⁵² , individual images were registered to MNI space by registering the PET image to the high-resolution MPRAGE image of the same subject and visit using SPM12⁵³ . The MPRAGE images were spatially normalized to MNI space^{54 (link)} and the estimated warping parameters were subsequently applied to the corresponding PET images. The registration and normalization quality of the MPRAGE and PET scans was checked visually. Three bilateral regions of interest were identified in each PET image; the putamen and caudate from the AAL template in the MNI space^{55 (link)}, and a hand-drawn SN mask which was used in our previous studies^{56 (link)–58} . For every region, a time-activity curve was extracted and modeled using the simplified reference tissue model (SRTM) with the cerebellum as reference region to estimate the regional binding potential (BP_ND) and relative input parameter (R1). A voxel-level approach, still using the cerebellum curve as reference, was applied to normalized images to create parametric BP_ND maps^{59 (link)}.

One-hundred-twenty min ¹¹C-PBR28 PET scans were performed on the Siemens High-Resolution Research Tomograph (HRRT) scanner and 3D PET brain images were reconstructed into a 256 × 256 × 170 matrix. Demographic and clinical characteristics of subjects under study are summarized in Table 1.