Abstract
Objective: Increased intraindividual variability (IIV) in function has been linked to various age-related outcomes including cognitive decline and dementia. Most studies have operationalized IIV as fluctuations across trials (e.g., response latencies) for a single task, with comparatively few studies examining variability across multiple tasks for a given individual. In the present study, we derive a multivariable operationalization of dispersion across a broad profile of neuropsychological measures and use this index along with degree of engaged lifestyle to predict risk of cognitive impairment. Participants and Methods: Participants (n = 60) were community-dwelling older adults aged 65+ years (M = 74.1, SD = 6.5) participating in a cross-sectional investigation of risk factors for amnestic mild cognitive impairment (a-MCI) and probable Alzheimer's Disease (AD). Participants were classified into three subgroups based on test performance and clinical judgement. Healthy controls (n = 30) scored better than -1 SD relative to existing norms on all classification measures, in the absence of memory complaints or functional impairments. The a-MCI group (n = 23) had self- or informant-reported memory complaints and scored 1 SD or more below the mean for at least one memory task while scoring better than 1 SD below the mean for all other cognitive domains, in the absence of functional impairments. The AD group (n = 7) scored at least 2 SD below the mean for two cognitive domains (including memory) with impairments in functioning. Measures spanned a range of cognitive domains (episodic memory, executive function, language), with the derived dispersion estimates reflecting variability across an individual's neuropsychological profile relative to the group average. Further, an Activities Lifestyle Questionnaire, indexing social, cognitive, and physical behaviors, was administered to assess the protective benefits of engaged lifestyle. Results: Multinomial logistic regression models examined the risk of being classified as a-MCI or AD as a function of increased dispersion, (dis)engaged lifestyle, and their interaction. Greater dispersion was associated with an increased likelihood of being classified with AD, with protective engaged-lifestyle benefits apparent for a-MCI individuals only. Conclusion: As a measure of IIV, dispersion across neuropsychological profiles holds promise for the detection of cognitive impairment.