Choline acetyltransferase activity was measured postmortem in five brain regions to determine if such activity provided biochemical support for clinical and pathological subgrouping of Alzheimer's disease. Seven patients with Alzheimer's disease were divided into groups based on age at onset, severity of neuropathological changes, history of myoclonus, family history of dementia, cerebellar amyloid plaques, and congophilic angiopathy. Thirty-two age-matched normal control subjects and 17 neurological control patients with Huntington's disease were also studied. Patients with early-onset and late-onset Alzheimer's disease did not differ in the clinical duration of their disease. Choline acetyltransferase activity was significantly lower in patients with early-onset Alzheimer's disease than in age-matched control subjects in frontal cortex, temporal cortex, hippocampus, and cerebellum. In contrast, choline acetyltransferase activity in patients with late-onset Alzheimer's disease was significantly lower than in age-matched control subjects only in hippocampus. There was a tendency for choline acetyltransferase activity to be lower in cortex from patients with early-onset Alzheimer's disease compared with cortex from the late-onset group, and this difference was significant in temporal cortex. Choline acetyltransferase activity was also measured in the substantia innominata from 9 patients with Alzheimer's disease and 5 age-matched control subjects. Subjects with early-onset Alzheimer's disease had significantly lower choline acetyltransferase activity in substantia innominata than did control subjects. Patients with Alzheimer's disease and a history of myoclonus had significantly lower choline acetyltransferase activity than did affected patients without myoclonus. Multivariate regression analysis showed myoclonus to be the single best predictor of low brain choline acetyltransferase activity. These results provide further evidence for clinical, pathological, and biochemical heterogeneity in Alzheimer's disease.