The ADNI dataset stimulated work in imaging genetics, as evidenced by the large number of published papers using the data (all ADNI genetics studies from the period 2009–2012,are reviewed by Shen et al., 2013; this volume). Even so, data from other elderly cohorts is needed to determine how well ADNI’s genetic findings generalize to other non-selected cohorts. ADNI deliberately selected participants who are predominantly Caucasian, and relatively free from drug or alcohol abuse and vascular disease. Genetic associations detected in ADNI may be stronger or weaker, or even not supported at all, in cohorts with different ethnic composition, or with other co-morbid health conditions. Because of this, it is important to recognize that some genetic associations may depend on the cohort studied, diagnostic or demographic criteria, and we should not expect all true genetic associations to be detectable in all cohorts.
