Journal of Clinical and Experimental Neuropsychology · Nov 16, 2026
Comparing multivariate cognitive change in unimpaired older adults to those with questionable impairment or mild dementia: a NACC study
By Grant JG, Hunter JM, McMillan TM et al.
Introduction We explored how frequently unimpaired older adults show declines across multiple cognitive measures and examined the extent to which their rates of multivariate cognitive change compared to older adults with questionable impairment or mild dementia. Method Two-year change scores were calculated for 1208 participants in the National Alzheimer’s Coordinating Center (NACC) database. The number of declines in performance falling below the 5th percentile in the natural distribution of change scores (the total change count) among the unimpaired participants (n = 742) was compared to those with questionable impairment (n = 389) and those with mild dementia (n = 77). Results Only 9.9% of the cognitively unimpaired participants obtained two or more significant declines in performance, compared to 27.5% of the questionably impaired participants and 49.4% of those with mild dementia. The total change count demonstrated significant discriminative ability for distinguishing between cognitively unimpaired participants and those with questionable impairment (AUC = .621) or mild dementia (AUC = .759). The threshold of two or more declines maintained high specificity (90.1%) for identifying atypical cognitive decline. In a multinomial logistic regression, a change count of two or more was a robust predictor of both questionable impairment (OR = 3.47) and mild dementia (OR = 8.55), beyond the effects of demographic covariates (age, gender, education, race/ethnicity). Conclusions Less than one tenth of cognitively unimpaired older adults show two or more significant declines in their performance on 13 measures from the NACC UDS 3.0 neuropsychological test battery. Individuals who exhibit two or more significant declines in their raw scores across this battery are likely experiencing atypical cognitive decline. Establishing longitudinal base rates of cognitive change across a test battery offers a viable, empirically grounded approach to identifying progression along the dementia continuum.
doi.org/10.1080/13803395.2026.2756260Open accessPubMed 43037353Biomarkers