It has been demonstrated that the biological data (e.g., genetic and neuroimaging data) from the UK Biobank can provide valuable information in diagnosis prediction of mental disorders. However, it is less clear whether non-biological data (e.g., behavioral, cognitive and clinical assessments and health records) is as predictive for mental disorders. If so, will the biological data further enhance the accuracy of mental disorder prediction when they are combined with non-biological data? In this study, we aim to address the two questions by combining the various forms of data from the UK Biobank and applying advanced machine learning algorithms to this combined data. This project is expected to take 5 years to complete. Through this project, we will develop a computational tool for individualized mental disorder identification, which may improve the diagnosis and facilitate better understanding of mental disorders.
