The demographics of aging identify an immediate need for the early diagnosis and development of dementia prevention strategies. Recent neuropathological studies have pointed to the early involvement of the hippocampus and entorhinal cortex in the progression of Alzheimer's disease in the brain. In particular, these studies have implicated tau-related pathology as an important cause of neuronal death. In addition, there is a large body of evidence showing that beta-amyloid, which has a predilection for the neocortex, is also involved early in the course of the disease and may also have toxic effects on cells. In vivo cerebrospinal fluid studies have shown that markers for these brain changes have a diagnostic value for Alzheimer's disease and that some measures also provide diagnostic specificity for Alzheimer's disease. Structural and metabolic imaging studies demonstrate brain changes in impaired and at-risk individuals. While currently available magnetic resonance and positron emission tomography techniques are not by themselves specific for the pathologic features of Alzheimer's disease, there are patterns of change that have been useful for the early diagnosis. As such, both prediction and longitudinal imaging studies demonstrate a capacity to recognize abnormalities that relate to future Alzheimer's disease and most recently to future mild cognitive impairment. This review highlights cross-sectional, prediction, and longitudinal magnetic resonance and positron emission tomography imaging studies and attempts to put into perspective their utility for the early diagnosis of Alzheimer's disease, and for their utility to provide diagnostic specificity. It is concluded that there is considerable promise for an early and specific diagnosis for Alzheimer's disease by combining information from imaging and biomarker modalities.