Alzheimer's disease poses a looming crisis for the health care system as well as society in general. The low efficacy of current treatments for those already affected with this disease has prompted the suggestion that interventions might be more successful if they were applied before the development of significant pathology, that is, when individuals are clinically asymptomatic. Currently, the field requires a sensitive and specific diagnostic tool for identifying those individuals destined to develop this disease. As a first step, we present here an analysis of cross-sectional data for 95 asymptomatic offspring (50–75 years of age) of autopsy-confirmed late-onset familial Alzheimer's disease cases and 90 age-matched controls, studied with functional magnetic resonance imaging (fMRI) to investigate brain activation patterns. Analysis of activation in response to a paired-associates memory paradigm found significantly different patterns in these groups. At-risk individuals showed more intense and extensive activation in the frontal and temporal lobes including the hippocampus during memory encoding, an increase unrelated to the APOE ε4 allele. They also showed decreased activation particularly in the cingulum and thalamus during both the encoding and recall phases of the task. These results demonstrate that asymptomatic individuals, at genetic risk for development of late-onset Alzheimer's disease by virtue of familial clustering, show functional activation patterns distinct from those without such risk more than a decade before their parent's onset age. While longitudinal study is needed to determine whether these patterns, or a subset of them, are predictive of disease onset, these findings suggest that functional neuroimaging holds promise as a method of identifying pre-clinical Alzheimer's disease.