Investigating the causal network of dementia using exposome data by employing a causal discovery approach combined with natural language processing models

Xinzhu Yu,Artitaya Lophatananon,Vivien Holmes,Kenneth R Muir,Hui Guo

DOI: https://doi.org/10.1101/2024.11.05.24316664

2024-11-06

Abstract:INTRODUCTION Comprehensively studying modifiable risk factors altogether to explore how they contribute to dementia mechanism is imperative for effective interventions. METHODS This study utilized natural language processing (NLP) models to select candidate risk factors of dementia from 5,505 variables in the UK Biobank. We took a holistic machine learning approach, fast causal inference in combination with mixed graphical models, to explore the complex causal mechanisms underlying dementia from 120 imputed variables. RESULTS In the identified causal network around dementia, eight risk factors may directly or indirectly contribute to dementia. In particular, mental disorders due to brain damage and dysfunction and to physical disease were identified as direct causes as well as mediators on the causal pathways to dementia. Evidence for a direct causal impact of phenotypic age on dementia was less pronounced. DISCUSSION The identified causal network offering valuable insights into the diseases mechanisms. Beyond direct connections to nerve or brain disorders, the potential direct link with biological age highlights its possible value in dementia management. Moreover, the use of NLP models for variable selection introduced an innovative application in medical research, highlighting a promising future for advanced tools in large-scale data analyses.