Optimizing Treatment Strategies in Bipolar Disorder Spectrum with Artificial Intelligence: A Systematic Review of Performance, Bias, and Clinical Applicability (Preprint)

BACKGROUND Bipolar disorder (BD) is a complex and heterogeneous psychiatric condition, characterized by fluctuating clinical courses that affect approximately 1-2% of the global population. Despite pharmacological advances, treatment response varies significantly among patients, making the identification of individualized treatment strategies a major challenge. Recently, artificial intelligence (AI) has emerged as a powerful approach in precision psychiatry to identify subtle patterns in complex data and inform personalized clinical decisions. OBJECTIVE Provide a structured synthesis of current evidence on AI-supported treatment optimization in BD spectrum. METHODS This systematic review was conducted in accordance with the PRISMA 2020 guidelines. Four databases (PubMed, Web of Science, Scopus, and EMBASE) were searched for original studies published after 2015 on the application of AI in the treatment of BD in adult patients. The methodological quality, risk of bias, and clinical applicability of the predictive models were assessed using the PROBAST+AI tool. RESULTS A total of 35 studies were included, divided into three main categories. Treatment response prediction, focused primarily on lithium response, with accuracies up to 100% in multimodal models. Relapse risk prediction, where models demonstrated feasibility in predicting relapses and rehospitalizations with AUCs between 65% and 85%. Patient stratification, used to identify clinical subgroups and pharmacological profiles, with excellent predictive capabilities (AUCs up to 99%). However, the PROBAST+AI assessment revealed a high risk of bias in most studies, primarily due to data analysis limitations, small sample sizes, and lack of external validation. CONCLUSIONS The adoption of AI tools in BD serves as a driver for therapeutic optimization, although current AI tools in BD should still be considered exploratory rather than ready for clinical use. Effective implementation in real-world clinical scenarios requires more robust, transparent, and externally validated models to ensure reliability and generalizability.