Correlational analysis of cerebral metabolic (rCMRglc) data obtained with positron emission tomography (PET) assesses group differences and has demonstrated reduced frontal-parietal rCMRglc interdependencies in Alzheimer's disease (AD). A multivariate analysis of rCMRglc data assesses individual differences. We recently identified discriminant functions, reflecting frontal-parietal rCMRglc interdependencies, that separated AD from control subjects. To test if the functions would identify an AD rCMRglc pattern in older Down syndrome (DS) adults with (DS DAT+) or without (DS DAT-) dementia, we applied the functions to longitudinal rCMRglc data in: young DS (n = 15), DS DAT- (n = 10), DS DAT+ (n = 4), and young controls (n = 15). All DS DAT+ and some of the later DS DAT- scans were classified as AD. The results provide additional validation of the functions and suggest their utility for the early detection of AD.