A1117
Title: Causal network analysis identified mental disorders and phenotypic age acceleration as causes of dementia
Authors: Hui Guo - University of Manchester (United Kingdom) [presenting]
Abstract: A number of biological and lifestyle factors have been associated with dementia. However, causal risk factors of the disease, which are imperative for interventions, remain elusive. Natural language processing models are utilized to select candidate risk factors of dementia from 5,505 measured variables in the UK Biobank. A holistic machine learning causal network approach is taken, fast causal inference in combination with mixed graphical models, to explore the complex causal mechanisms underlying dementia from imputed data of 120 selected variables. Of these, it is shown that eight risk factors may directly or indirectly cause dementia. The work systematically investigated causal risk factors of dementia, which paves the way for a fuller insight into its causal mechanisms. It is also shown that natural language processing models have the potential for selecting variables from high-dimensional data.
