For the Tau tracer AV1451 scans, does the procedure for signal intensity normalization use the whole brain mask or is it another mask (e.g., cerebellum)?
I'm not sure if you are talking about the image data or the CSV (numerical) spreadsheet data. There is no "fixed" intensity normalization procedure; this is intended to be carried out by individual researchers. The AV1451 CSV file contains a number of possible reference region SUVRs, including the cerebellar grey matter, which is a frequently used reference region for AV1451 analysis.
The intensity normalization for the "pre-processed" uploads to LONI for the Tau scans use the cerebellar gray mask.