Multivariate data analysis statistical procedures were applied to identify regional cerebral blood flow (rCBF) patterns in patients with multi-infarct dementia (M.I.D.) and Alzheimer’s disease (A.D.) when compared to a control group (C). A fronto-temporal-parietal pattern was identified for the A.D. group. The M.I.D. group demonstrated a predominant temporal-parietal pattern. A discriminant function analysis classified 96% of the A.D. and C. patients correctly. One hundred percent hit rate was obtained in discriminating between M.I.D. and C. Discrimination of A.D. and M.I.D. obtained 82% diagnostic accuracy. When the three groups were compared the hit rate was 93%. Mean rCBF and mean relative weight of the gray matter (Wg) differences were found when each dementia group was compared with C.M.I.D. and A.D. did not differ in mean rCBF but significant differences were found for Wg, with M.I.D. having greater reduction. The performance on standardized psychological tests of two independent samples of A.D. and M.I.D. were compared. The Memory Quotient from the Wechsler Memory Scale was consistently more impaired for the A.D. group. No correlation between degree of psychological impairment and rCBF or W g reduction was found.