Intraindividual Variability as a Predictor of Cognitive Decline in Elderly

Mild cognitive impairment (MCI) is conceptualized as a transitional state between normal aging and fully developed clinical features of dementia. The literature on MCI is notable for varied measurement approaches and lack of stability in the diagnostic entity, with many individuals remaining stable or reverting to normal cognitive status. Researchers agree that multiple neuropsychological domains should be assessed to enhance the assessment and prediction of cognitive decline. In addition, within-person assessments capture trajectories of decline, which are better suited for understanding individual change than simple comparison to group norms. The current study investigated the ability of within-person change on novel experimental cognitive tasks, referred to as intraindividual variability (IIV), to enhance characterization and prediction of MCI compared to traditional neuropsychological measures. Participants were 426 older adults from the Einstein Aging Study, a longitudinal community-based study, who were classified as healthy controls (HC) with no cognitive or functional impairment at a baseline assessment (mean age=79.89, SD=5.05). Participants completed a two-day testing battery that included standard (Free and Cued Selective Reminding Task, Trail Making Test Part A, Trail Making Test Part B) and experimental tests (Cued-Recall Retrieval Speed Task, Number Match Task, N-Back 2 task) of episodic memory, processing speed and executive function. For standard tasks we recorded the number of items correct and/or time to task completion. For experimental tasks we recorded reaction time IIV and accuracy IIV. The outcome of interest was MCI, which was defined by two classification approaches to prevent diagnostic circularity. Participant cognitive status was assessed annually for up to 10 years following initial baseline testing. Within-domain analyses aimed at characterizing future MCI status revealed that participants who transitioned to MCI performed worse at baseline on the episodic memory standard task and all experimental tasks; they also exhibited greater IIV on the episodic memory experimental task. Within-domain analyses aimed at predicting future MCI status revealed that baseline performance on the episodic memory standard task predicted future incident MCI and baseline performance for the episodic memory experimental task predicted future incident MCI when utilizing the alternate MCI criteria. Across-domain analyses aimed at predicting future MCI status revealed that episodic memory and executive function standard tasks predicted incident MCI better than experimental tasks. Across-domain time-dependent analyses revealed that the episodic memory standard and experimental tasks and the executive function standard task distinguished between individuals who transitioned to MCI within three years and HC. However, this analysis was unable to distinguish between participants who transitioned to MCI four or more years after initial testing. Overall, results revealed that baseline differences on cognitive tasks present up to 10 years before MCI diagnosis. Episodic memory IIV classified transition of MCI for both diagnostic MCI definitions. In addition, incident MCI can be predicted from baseline performance on episodic memory standard and experimental tasks and on an executive function standard task. These measures could be considered an early detection approach and should be further examined to facilitate early diagnosis of MCI and related preclinical conditions