Investigating causal pathways between chronic exposure to air pollution and Alzheimer's disease in the UK Biobank using a data-driven causal graph approach
Approved Research ID: 95432
Approval date: March 20th 2023
Alzheimer's disease (AD) is the most common form of dementia, affecting more than 55 million people globally, and is becoming a significant public health threat. Existing evidence has shown that ambient air pollution exposure over the long term is associated with abnormal brain structure, cognitive decline, and dementia risk. Air pollution may affect AD dementia via multiple pathways involving oxidative stress, neuroinflammation, and blood-brain barrier damage. Previous research has also suggested that chronic inflammatory disease may mediate the relationship between air pollution exposure and brain health. However, the underlying causal mechanisms between air pollution exposure and AD dementia and how chronic comorbidities modify such causal pathways remain to be fully understood. This research project aims to fill this gap by developing artificial intelligence (AI) models to examine (1) the causal effect of air pollution exposure on AD dementia and (2) the role of chronic inflammatory disease (asthma and COPD) in the causal pathways between air pollution exposure and AD dementia. The AI models capitalize on graph neural networks and the whole cohort data, identifying the causal graph that delineates the statistical relationships between air pollution exposure, AD outcomes, and various control variables ranging from genetic and sociodemographic factors to clinical and behavioral measurements, thus better disentangling the causal effect of air pollution exposure on AD dementia. The proposed project is estimated to complete in 12 months. The proposed project is expected to help better understand the role of air pollution exposure in AD dementia and inform public health policymaking about air pollution control and AD prevention.