Linking dynamic connectivity states to cognitive decline and anatomical changes in Alzheimer's disease

Alterations in brain connectivity provide early indications of neurodegenerative diseases like Alzheimer's disease (AD). Here, we present a novel framework that integrates a Hidden Markov Model (HMM) within the architecture of a convolutional neural network (CNN) to analyze dynamic functional connectivity (dFC) in resting-state functional magnetic resonance imaging (rs-fMRI). Our unsupervised approach captures recurring connectivity states in a large cohort of subjects spanning the Alzheimer's disease continuum, including healthy controls, individuals with mild cognitive impairment (MCI), and patients with clinically diagnosed AD. The framework successfully identified distinct brain states associated with different clinical stages of AD, demonstrating a progressive reduction in functional flexibility as disease severity increased. Specifically, we observed that patients with AD spend more time in brain states dominated by unimodal sensory networks, while healthy controls exhibited more transitions to polymodal, cognitively demanding states. Importantly, the fraction of time spent in each state correlated with cognitive performance and anatomical atrophy in key regions, providing new insights into the disease's progression. Our findings suggest that the disruption of dynamic connectivity patterns in AD follows a two-stage model: early compensatory hyperconnectivity is followed by a decline in connectivity organization. This framework offers a powerful tool for early diagnosis and monitoring of AD progression and may have broader applications in studying other neurodegenerative conditions.