NEUROIMMUNOLOGY AND PSYCHIATRIC DISORDERS: BRIDGING BRAIN AND IMMUNE SYSTEMS

Abstract

To take precision psychiatry further, precision psychiatry requires a knowledge of how the relationship between the immunological and nervous systems functions.  The following study was a multi-modal, multi-method experimental study that examined the synergy between the immune system and the brain in individuals with schizophrenia, bipolar, and major difference disorder.  We quantified the peripheral cytokines (IL-6, TNF-a, IFN-6, and CRP) using ELISA. We also had access to structural and functional brain data in the form of MRI scans and were able to turn them into voxel-based morphometry and graph-theoretical analysis.  We found that the machine learning models (Random Forest and SVR) are able to predict the extent of bad psychiatric conditions based on immunological data and neuroimaging data with good accuracy (R 2 more than 0.82).  The feature interpretation using SHAP showed that the best predictors were the hippocampal connection and IFN- \gamma.  The relationship between the cytokines and the artificial substrates was demonstrated through nine quantitative tables and 12 figures. Qualitative expert interviews theme analysis proved the biological and clinical importance of the results.  We have demonstrated that the neuroimmune biomarkers reflect more than just mental symptoms, but may become a source of prediction and applied to other fields.  The paper demonstrates an effective AI-powered system to retrieve brain and immunological data. This creates new opportunities to diagnose, personalize the treatment and discover more about the systematic causes of mental illnesses.