Prediction of mild cognitive impairment progression using time-sensitive multimodal biomarkers.

Abstract

Alzheimer's disease (AD) develops silently for years before symptoms emerge, making early identification of at-risk individuals essential for prevention trials and early intervention. We tested whether combining neurophysiological, imaging, and blood biomarkers improves prediction of progression to mild cognitive impairment (MCI) in cognitively unimpaired older adults with a family history of AD (<i>n</i> = 102; 31 progressors; mean follow-up of 5.9 years). Magnetoencephalography, magnetic resonance imaging, plasma biomarkers, and amyloid and tau positron emission tomography each captured complementary aspects of disease biology. Multimodal models predicted progression more accurately than demographic and genetic factors alone. Higher MEG alpha power was associated with increased near-term risk, whereas higher gamma activity predicted lower near-term risk; both effects weakened over time. Higher neocortical amyloid burden predicted increasing risk over follow-up, whereas plasma biomarkers and entorhinal tau predicted higher risk without significant time-varying effects. These findings support a time-sensitive multimodal framework for identifying cognitively unimpaired individuals at risk of MCI due to AD.