The concept of mild cognitive impairment (MCI) has been proposed by Petersen et al1,2 as a nosologic entity referring to elderly persons with mild cognitive deficit without dementia. MCI is a seductive concept for clinicians and researchers because, unlike older concepts such as Age-Associated Memory Impairment (AAMI)3 or Age-Associated Cognitive Decline (AACD),4
it is assumed to be pathology-based and hence amenable to treatment or preventive actions. MCI is being widely used in epidemiologic and clinical studies as an intermediate stage between cognitive normality and dementia. A high number of persons have been found to suffer from MCI in the population at large, and many are now being referred to memory clinics. There is an expectation of even larger numbers of individuals with memory complaints coming forward for assessment as the results of the ongoing randomized clinical trials become available. MCI appears to be a heterogeneous clinical entity.5 Multiple sources of heterogeneity have been described: heterogeneity in etiologic factors (various types of degenerative lesions, vascular lesions, psychiatric features, and association with nonneurologic pathologic conditions), heterogeneity in clinical symptoms,6 and heterogeneity in clinical course.7 The heterogeneity of MCI has been recognized,8 and the first consensus report has recommended that the term ‘MCI’ be qualified by an appropriate modifier. We thus propose a subclassification of MCI for use in research and in clinical practice.