The proposed project is designed to assess early diagnosis of a major form of mental impairment associated with aging: Alzheimer disease. The project will measure cerebral glucose metabolic patterns using positron emission tomography (PET) in people considered to be at high risk for developing familial Alzheimer disease (AD). This high risk group will consist of persons with Age-Associated Memory Impairment (AAMI) who have at least one first-degree relative with AD (familial AAMI or FAAMI group). PET measures, using the glucose analog 18F-fluorodeoxyglucose (FDG), will be obtained on FAAMI subjects at their initial assessment. These measures will be compared to those from age-matched controls and clinical profiles will be repeated at three-year follow-up evaluations to identify subjects who progress to probable AD. The study builds upon prior PET work at UCLA, which identified abnormal metabolic patterns in AAMI subjects who later developed nonfamilial AD. The proposed project aims to replicate these investigate the applicability of such metabolic patterns early in the course of familial AD. It is hypothesized that FAAMI subjects will show abnormal PET scan results compared to controls; and that FAAMI subjects will be more likely to develop AD in two years than will controls. Whether PET scan results predict which FAAMI patients develop AD in two years, as well as differences among AD subgroups, also will be explored. Early identification of patients with AAMI that progresses to AD is critical not only for guiding optimal clinical management and planning but also for identifying potential subjects who are most likely to benefit from experimental treatments. Because AD probably represents a multi-causal disorder, longitudinal refining diagnostic accuracy of homogeneous disease forms. Further study of such homogeneous subgroups would increase the likelihood of finding etiologies, which could eventually lead to specific treatments.