 Deep biomarker, a deep learning model developed by researchers at Stanford University, was used to accurately predict the onset of psychosis in Alzheimer's disease, AD, patients. The model was found to be significantly better than traditional risk assessment methods, providing valuable insight into the underlying causes of AD plus P. Metabolic syndrome, inflammatory markers, and liver function were identified as key contributors to the development of AD plus P. Additionally, the model provided mechanistic insights into the pathophysiology of AD plus P, suggesting potential targets for future therapeutics. This article was authored by Peihao Fan, Ashen Miranda, Shi Guangqi, and others.