Hypergraph clustering for analyzing chronic disease patterns in mild cognitive impairment reversion and progression.

Identifying the sequential patterns of chronic conditions that precede the onset of mild cognitive impairment (MCI) is essential for understanding both the progression and the potential reversal of MCI. This study identifies common sequences of chronic conditions preceding MCI and introduces a novel, network-based clustering framework for characterizing patients with similar progression patterns linked to cognitive trajectories. We used the Mayo Clinic Study of Aging (MCSA) cohort and categorized participants of MCSA into two groups (i) stay at MCI or progressed to dementia, or (ii) reversion to normal within 5 years after the first onset of MCI. We curate the state transition network (patient level) for identifying and introduced hypergraph clustering (patient group level) to characterize participants with similar sequences. We identified generic key indicators (e.g., chronic kidney disease) and highlighted sex-specific potential indicators (e.g., arthritis, hypertension) associated with MCI reversal, opening new research directions to explore potential differences between males and females. There are certain ssequences of chronic conditions (e.g., originating from arthritis) that are more commonly observed in females. However, these observations warrant further validation. The proposed framework - hypergraph clustering - offers a promising method for uncovering similarities in patients through unique trajectories of chronic conditions that precede MCI.