Multimodal signatures predictive of future psychosis transition in youths at clinical high risk - psychotic disorders are a leading contributor to the global disease burden, causing high levels of…

The National Institute of Mental Health awarded The Leland Stanford Junior University $721,645 on May 1, 2026, under the Mental Health Research Grants program (CFDA 93.242) to identify multimodal prognostic markers predictive of psychosis transition in youths at clinical high risk using artificial intelligence and large-scale multimodal data from clinical, cognitive, and neurobiological sources.

The recipient will develop and validate an AI-based computational framework combining explainable AI methods with integrative theoretical models of psychosis to improve prognostic precision for early identification and treatment of individuals who progress to psychotic disorders. Work focuses on identifying robust prognostic markers from integrated data across multiple consortia to enable earlier intervention before overt psychotic symptoms emerge.

Performance extends through January 31, 2031, with work performed in Palo Alto, California. The University of California, Los Angeles serves as a sub-recipient under sub-award R01MH140250-64036275-343022, dated July 28, 2026, delivering complementary research on multimodal signatures predictive of future psychosis transition in youths at clinical high risk.

Multimodal signatures predictive of future psychosis transition in youths at clinical high risk - psychotic disorders are a leading contributor to the global disease burden, causing high levels of disability and increased mortality. To improve outcomes, IT is essential to identify and treat patients in the early stages of psychotic disorders, especially before overt symptoms appear. Yet, despite decades of research, we are unable to accurately identify early on individuals who will progress to develop a psychotic disorder, even those who are clinically high risk for psychosis, due in part to small sample sizes and extant approaches that do not capture the multifactorial etiology of psychotic disorders. There is therefore an urgent need to substantially improve prognostic precision. Critically, accurate and robust prognostic markers are needed to understand the origins and progression of psychosis and to identify precise neurobiological targets for early treatment. Newly available large-scale multimodal data-clinical, cognitive, and neurobiological-as well as exciting recent advances in artificial intelligence models and methods that overcome limitations of extant approaches offer an unprecedented opportunity for developing accurate and robust prognostic markers for psychosis. The overarching goal of our proposal is to identify accurate and robust multimodal prognostic markers for psychosis using a novel data-driven AI-based computational framework. Building on our highly encouraging preliminary results, we will use an innovative approach combining our recent work on AI models and explainable AI methods as well as integrative theoretical models of psychosis with a wealth of newly available large-scale multimodal data from multiple consortia. The specific objectives of our proposed work are threefold. In aim 1, we will identify prognostic markers using clinical, cognitive, and neurobiological data to predict future psychosis transition, particularly in youths at clinical high risk. In aim 2, we will evaluate the generalizability and the temporal (longitudinal) stability of the identified prognostic markers. In aim 3, we will determine whether the identified prognostic markers predictive of future psychosis transition in youths at clinical high risk are a characteristic trait of psychosis. Through the successful completion of the work described here, our multidisciplinary team is uniquely positioned to transform our understanding of the mechanisms associated with the risk for and development of psychotic disorders, as well as identify neurobiological targets. Ultimately, these advances will lead to the development of individualized prognostic tools and early targeted treatments for psychosis and, more broadly, advance precision psychiatry.