SEMANTIC CLUSTERING ON THE C.V.L.T.-II IS REDUCED IN COGNITIVELY INTACT OLDER ADULTS WITH AUTOPSY-CONFIRMED ALZHEIMER’S DISEASE NEUROPATHOLOGIC CHANGE

Abstract

Semantic clustering (SC) is a learning strategy of grouping information with shared attributes and is measured on memory such tests as the California Verbal Learning Test-II (CVLT). Recent research suggests that reductions in SC parallel clinical progression in Alzheimer's disease (AD) such that SC for intact individuals is greater than mild cognitive impairment which is greater than dementia. The objective of the present study was to examine SC in an autopsy cohort, longitudinally. It was hypothesized that participants with greater AD pathology who show reduced SC over time. Participants were older adults, cognitively intact at baseline, never demented, and came to autopsy after ≥3 years of annual CVLT administrations. Brain autopsies followed standardized protocols. Braak staging was used to classify participants for analysis: AD- (Braak 1-3) and AD+ (4-6). Linear mixed models (LMMs) controlling for age, sex, and education were used to examine total learning trial SC (z-scores) longitudinally with a group × administration interaction. 39 participants were classified as AD- and 10 as AD+. The groups (AD-, AD+) were statistically similar in age at baseline (79.1y, 80.4y), education (16y, 17.5y), sex (20F, 4F), age at death (86.4y, 87y), post mortem interval (4.8hrs, 10.5hrs), and administrations (4.9y, 5.4y), ps > 0.19. A LMM indicated that the AD+ group exhibited reduced SC at each CVLT administration (β = −0.20, p = 0.05). Excluding one AD+ participant with a precipitous decline in SC nullified the group × administration interaction but the AD+ group exhibited significantly reduced SC compared to the AD- group (β = −1.02, p = 0.021) at all administrations. The results demonstrate that SC is reduced in non-demented older adults with autopsy-confirmed AD. Although we hypothesized a reduction in SC over time, it appears that SC was reduced prior to clinical indication of cognitive decline. This is in line with evidence that AD pathology accumulates gradually over the span of years. A small, well-educated cohort limits the generalizability of this study; but, it provides evidence that SC may offer an indicator of preclinical AD using a widely-available neuropsychological test.