On the basis of neurological evidence that autistic patients have fewer Purkinje and granule cells in the cerebellum as well as vermal cerebellar hypoplasia, the authors tested the hypothesis that autistic patients have cerebellar hypofunctioning. They used positron emission tomography of the cerebellum with 18F-labeled 2-deoxyglucose to study seven autistic patients and eight age-matched control subjects. The results showed no significant difference in mean cerebellar glucose metabolism between the two groups, but all mean glucose rates of the autistic patients were either equal to or greater than those of the control subjects. The implications of these findings are discussed.