Associations between ventricular boundary shift integral and composite cognitive domains across the alzheimer’s disease continuum

Cognitive impairment is hallmark of Alzheimer’s disease (AD). Although structural MRI has consistently demonstrated widespread brain atrophy in AD, the comparative utility of different imaging sequences for predicting cognitive decline remains uncertain. The Boundary Shift Integral (BSI), a longitudinal measure quantifying brain volume change over time, provides a dynamic alternative to conventional cross-sectional volumetric measures and could more sensitively reflect cognition-related neurodegeneration. This study investigated the associations between BSI-derived atrophy and domain-specific cognitive performance across the AD continuum. Participants (Cognitively Normal CN = 155, Mild Cognitive Impairment MCI = 283, AD = 100) underwent comprehensive cognitive testing, and composite scores were computed for memory, executive function, and language domains. BSI metrics for the whole brain, ventricles, and bilateral hippocampi were derived from serial MRI scans acquired at baseline and 12-month follow-up and were directly compared with conventional T1-weighted volumetric measures. Linear regression models were used to examine associations between imaging markers and domain-specific cognitive scores, adjusting for relevant covariates. In the MCI group, BSI measures demonstrated significant associations with all cognitive domains, consistently outperforming traditional volumetric measures. Whole-brain and ventricular BSI were significantly associated with memory, executive function, and language, while hippocampal BSI showed particularly robust associations with memory. In the AD group, BSI again exhibited more consistent associations with cognitive performance than volumetrics, especially for whole-brain and right hippocampal measures. No significant associations were observed in the CN group. Direct comparisons confirmed that BSI provides enhanced sensitivity to cognition-relevant structural changes, particularly in early and prodromal stages of AD. BSI measures offer superior sensitivity in detecting early structural brain changes associated with cognitive decline. These findings support the utility of BSI as a dynamic and clinically relevant tool for early detection and longitudinal monitoring of the disease.