A model of brain function to predict tau across Alzheimer's disease and related disorders.

A machine-learning framework using a computational brain-function model and FDG-PET predicts AD-type tau measured by flortaucipir-PET across AD and ADRD, with participant-level R² up to ~0.71.

• Why it matters: AD-type tau varies across Alzheimer’s disease and related disorders, and measuring it with FTP-PET is limited; a data-driven way to estimate tau from FDG-PET is needed.
• What they did: The model was trained on 1088 research participants with FDG-PET and FTP-PET, then validated in a longitudinal subset (n = 215), ADNI (n = 295), and a heterogeneous clinical cohort (n = 264; n = 28 with FTP-PET).
• The result: It predicted FTP-PET at the participant level (R² ~ 0.34–0.71), classified low versus high tau stages with 79–95% accuracy, and estimated tau in the heterogeneous cohort without FTP-PET.

AI summary of the abstract · may contain errors · not medical advice · read the paper.