Novel Multimodal Precision Medicine Approaches and the Relevance of Developmental Trajectories in Bipolar Disorder

Abstract

There is a pressing need to establish objective measures to improve diagnosis, prediction, prevention, and treatment of bipolar disorder (BD). Multimodal artificial intelligence (AI) tools could provide these means by incorporating various layers of data orthogonally related to BD, including genomics and other omics, environmental exposures, imaging measures, electronic health records, cognition, sensing devices, and clinical variables. These rapidly evolving AI models hold promise to capture the multidimensional complexity of BD and delineate clinically relevant developmental trajectories that could guide clinical care and therapeutic strategies. In this review, we describe the potential of mapping developmental trajectories that underlie BD, outline how novel multimodal models could improve the prediction of BD and related outcomes, and discuss specific clinical use cases and key ethical and practical challenges regarding the development and potential implementation of these multimodal AI solutions to advance precision medicine approaches in BD.