Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bayesian Network Modeling of Causal Influence within Cognitive Domains and Clinical Dementia Severity Ratings for Western and Indian Cohorts

Wupadrasta Santosh Kumar, Sayali Rajendra Bhutare, Neelam Sinha, Thomas Gregor Issac

TL;DR

The paper addresses how six cognitive-domain scores causally influence the global Clinical Dementia Rating ($CDR$) across Western (ADNI) and Indian (LASI) cohorts. It employs seven-node directed acyclic graphs learned with the Peter–Clark (PC) algorithm, quantifying edge strengths via a normalized chi-square metric to compare causal structures between populations. A key finding is that memory ($M$) exerts the strongest influence on $CDR$ in both datasets, but the overall DAG connectivity and edge strengths differ—ADNI is denser with high-strength edges, while LASI is sparser with variable weights—reflecting population-specific dementia pathways. The work demonstrates the value of causal DAGs for understanding domain-specific contributions to dementia severity in diverse populations and highlights the PC algorithm’s suitability for categorical clinical data, informing targeted interventions and field validation with clinicians.

Abstract

This study investigates the causal relationships between Clinical Dementia Ratings (CDR) and its six domain scores across two distinct aging datasets: the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Longitudinal Aging Study of India (LASI). Using Directed Acyclic Graphs (DAGs) derived from Bayesian network models, we analyze the dependencies among domain scores and their influence on the global CDR. Our approach leverages the PC algorithm to estimate the DAG structures for both datasets, revealing notable differences in causal relationships and edge strengths between the Western and Indian populations. The analysis highlights a stronger dependency of CDR scores on memory functions in both datasets, but with significant variations in edge strengths and node degrees. By contrasting these findings, we aim to elucidate population-specific differences and similarities in dementia progression, providing insights that could inform targeted interventions and improve understanding of dementia across diverse demographic contexts.

Bayesian Network Modeling of Causal Influence within Cognitive Domains and Clinical Dementia Severity Ratings for Western and Indian Cohorts

TL;DR

The paper addresses how six cognitive-domain scores causally influence the global Clinical Dementia Rating () across Western (ADNI) and Indian (LASI) cohorts. It employs seven-node directed acyclic graphs learned with the Peter–Clark (PC) algorithm, quantifying edge strengths via a normalized chi-square metric to compare causal structures between populations. A key finding is that memory () exerts the strongest influence on in both datasets, but the overall DAG connectivity and edge strengths differ—ADNI is denser with high-strength edges, while LASI is sparser with variable weights—reflecting population-specific dementia pathways. The work demonstrates the value of causal DAGs for understanding domain-specific contributions to dementia severity in diverse populations and highlights the PC algorithm’s suitability for categorical clinical data, informing targeted interventions and field validation with clinicians.

Abstract

This study investigates the causal relationships between Clinical Dementia Ratings (CDR) and its six domain scores across two distinct aging datasets: the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Longitudinal Aging Study of India (LASI). Using Directed Acyclic Graphs (DAGs) derived from Bayesian network models, we analyze the dependencies among domain scores and their influence on the global CDR. Our approach leverages the PC algorithm to estimate the DAG structures for both datasets, revealing notable differences in causal relationships and edge strengths between the Western and Indian populations. The analysis highlights a stronger dependency of CDR scores on memory functions in both datasets, but with significant variations in edge strengths and node degrees. By contrasting these findings, we aim to elucidate population-specific differences and similarities in dementia progression, providing insights that could inform targeted interventions and improve understanding of dementia across diverse demographic contexts.
Paper Structure (18 sections, 1 figure, 3 tables)

This paper contains 18 sections, 1 figure, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Studies on CDR modeling
  4. Studies on utilizing DAG for modeling causal influences
  5. Materials and Methods
  6. Clinical data
  7. Estimating a Directed Acyclic Graph
  8. Notation
  9. Results
  10. Similarities in the Bayesian network structure of ADNI and LASI
  11. Differences in the ADNI and LASI DAGs
  12. Discussions
  13. Differences in ADNI and LASI datasets
  14. Contrasting DAG with other models on CDR
  15. Advantages of PC algorithm, especially for the categorical data like CDR data
  16. ...and 3 more sections

Figures (1)

  • Figure 1: Reconstructed DAG structure for the two datasets, using PC algorithm. (a) ADNI dataset. (b) LASI dataset. The full form of the six cognitive domains are: M - Memory, O - Orientation, JPS - Judgement and problem-solving, CA - Community affairs, HH - Home and Hobbies, PC - Personal care