ABSTRACT

The observation that MCI is a heterogeneous classification, regardless of how it is defined (i.e., narrow to broad definitions), has led some researchers to question the use of MCI as a clinical construct because research, to date, does not support the often-held global etiological assumption that MCI necessarily reflects early-stage AD (Wolf & Gertz, this volume). Instead, some researchers have chosen to make no assumptions about the relation between behavioral presentation and underlying pathology; rather, they identify MCI only as a given level of cognitive impairment in the absence of dementia as evidenced by performance on measures of cognitive functioning (e.g., Palmer, Backman, Small, & Fratiglioni, this volume; Wilson et al., this volume). This approach also makes no assumptions about the relative importance of particular types of cognitive impairments and allows for empirical determination of risk factors and the most relevant features in relation to specific outcomes. Broad definitions, as discussed in Chapter 1, result in higher prevalence rates, with lower rates of conversion, but may yield some different types of information than that derived from studies using highly selective diagnostic MCI criteria. For example, in Tuokko, Frerichs, Graham, Rockwood, Kristjansson, Fisk, et al.'s (2003a) study of people identified with cognitive dementia-no dementia (CIND) from the Canadian Study of Health and Aging, all subgroups identified on the basis of presumed etiology declined at similar rates. Fabrigoule, Barberger-Gateau, and Dartigues (this volume) and Palmer et al. (this volume) demonstrate that it may be important to include people with impairments in activities of daily living within the construct of MCI, rather than excluding them.