Bayesian network modeling of statistical dependency within cognitive domains and clinical dementia severity ratings for Western and Indian cohorts.

Abstract

Background: Dementia is studied rigorously in the current times, to assess the potential factors that contribute towards its progression, such as demographic factors, geographical background as well as to understand its impact on essential life skills. Alzheimer's disease, the most common form of dementia, is a major focus due to its prevalence and lack of curative treatments.

Objective: We aim to study dementia progression across two geographically distinct populations, by analyzing Clinical Dementia Rating (CDR) scores and its domain-specific components, which reflect the decline in essential life skills. This analysis may support clinicians in designing targeted interventions.

Methods: This study investigates the statistical dependencies between the CDR and its six defining domain scores across two distinct aging datasets: the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Longitudinal Aging Study of India (LASI). We analyzed the dependencies among domain scores and their influence on the global CDR in both datasets using data-driven directed acyclic graphs, within Bayesian network models.

Results: Distinct notable similarities and differences in the associations and the edge strengths between the Western and Indian populations were identified. Memory showed similar prominence in contributing towards global CDR, while the incoming edges on memory domain reduced in LASI over ADNI, specifically from the domains, Home and hobbies, and Community affairs.

Conclusions: The analysis highlights a stronger dependency of CDR scores on memory functions in both datasets. The results overall elucidate population-specific differences and similarities in dementia progression across diverse demographic contexts.