Background Cognitive function may serve as an early indicator of prodromal Alzheimer’s disease (AD). We examined how cognitive measures over the life course are associated with the AD-related biomarkers of amyloid (Aβ) status and whole brain volume in a population-based sample free of dementia and other major neurological disorders.

Methods Data were from 458 (49% female) dementia-free participants from Insight 46, a sub-study of the National Survey of Health and Development (1946 British birth cohort). At age 70–72, participants underwent 18F-Florbetapir amyloid-PET and multi-modal MRI imaging. Regression analyses and multilevel modelling examined patterns between cognitive measures (spanning age 8–69) and the neuroimaging outcomes of Aβ status (±) and whole brain volume.

Results Of a range of cognitive tests, word-list learning (WLT) at age 69 was specifically associated with Aβ status at the 5% level (OR 0.75 (0.57, 0.96)). Aβ+ individuals additionally showed faster WLT decline in the 26 years preceding imaging (M1: β=-0.06 (-0.09,-0.01)). Measures of reaction time (M1: b=-0.17 (95% CI -0.27,-0.08) and search speed (M1: b=0.12 (95% CI 0.03, 0.22)) at age 60, and decline in search speed over 26 years (M1: β=0.01 (0.00,0.01)), were associated with smaller whole brain volume. These patterns remained similar after adjusting for childhood cognition, sex, education, child and adult SEP, affective problems and concurrent pathology. APOE-e4 status attenuated the association between verbal memory and Aβ.

Conclusion Associations between cognitive function and neuropathology at age 70–72 appear to manifest between ages 60–69 in a population-based sample without dementia or other major neurological problems. In particular, decline in WLT is associated with Aβ+ and is partially attributed to the effects of APOE-ε4; whereas level of reaction time and rate of decline of search speed are associated with smaller brain volume and are APOE-ε4-independent. Our findings are consistent with evidence of cognitive changes as part of an AD prodromal syndrome in early older-age; and provide evidence that cognitive domains can differentiate underlying pathophysiology associated with AD.