Estimates of the prevalence rates for dementia vary significantly among countries. Such variation may be explained, at least in part, by methodologic differences in studies. The disparities in prevalence rates of dementia subtypes, particularly Alzheimer's disease and multi-infarct dementia, are especially apparent in studies conducted in Eastern and Western countries. In Japan and China, the prevalence of multi-infarct dementia exceeds that of Alzheimer's disease, whereas in the West, Alzheimer's disease predominates in the vast majority of studies. Clearly, cross-cultural studies of incidence using standard methods are needed to investigate whether a true difference in risk exists, and which risk factors differentially contribute to this variation. Migrant studies of genetically homogeneous populations offer a unique opportunity to answer these questions. This article explores the value of migrant studies, their application to etiologic questions of dementia and its subtypes, and recommendations concerning how to conduct such studies.