Mount Sinai Health System and IBM Research have unveiled an innovative research initiative aimed at improving mental health treatment for young people through artificial intelligence and behavioral data. The Phenotypes Reimagined to Define Clinical Treatment and Outcome Research (PREDiCTOR) study seeks to address the challenge of lacking objective measures in psychiatry by integrating advanced AI with data collected from clinical interviews, smartphone sensors, and cognitive tests.
Funded by a substantial $20 million grant from the National Institute of Mental Health (NIMH), the PREDiCTOR study involves a collaboration among leading institutions, including Harvard, Johns Hopkins, Columbia, Carnegie Mellon, and Deliberate AI. The study will focus on developing objective, scalable, and cost-effective methods to identify novel clinical signatures that can enhance individual-level prediction and clinical decision-making in mental health care.
Dr. Rene Kahn, Chair of Psychiatry at the Icahn School of Medicine at Mount Sinai and Mount Sinai Health System, will co-lead the project alongside Dr. Cheryl Corcoran, Program Leader in Psychosis Risk at Icahn Mount Sinai, and Dr. Guillermo Cecchi, Director of Computational Psychiatry and Neuroimaging at IBM Research.
Dr. Corcoran emphasized the potential of leveraging detailed behavioral data: “Every clinical visit provides a wealth of untapped data, including spoken language, eye contact, and facial expressions from both patients and clinicians. By analyzing audiovisual data from clinical interviews and integrating smartphone data—such as physical activity, social interactions, sleep patterns, and diary audio—we can develop clinical signatures that indicate key outcomes.”
The study will focus on patients aged 15 to 30 who seek treatment for the first time at six outpatient mental health clinics within the Mount Sinai Health System. This age range is critical as it encompasses a developmental period where many mental health conditions emerge, and accurate diagnoses and prognoses can be challenging.
Dr. Kahn highlighted the significance of the study: “Individualized prognosis and clinical decision-making during this critical developmental window can profoundly impact the long-term trajectory of these young patients. Our team’s expertise in electronic health record analysis, behavioral data collection, and computational analysis uniquely positions us to undertake this important work.”
Participants will have their clinical visits recorded both audibly and visually over one year. The study will assess cognitive function at the outset and track patients over the year to develop clinical signatures for significant outcomes, including treatment disengagement, emergency room visits, and hospitalizations.
Dr. Cecchi noted the transformative potential of the study: “Our goal is to understand what factors predict whether young people remain in treatment or drop out, and what factors lead to worsening symptoms requiring acute care. We believe advancements in AI are now robust enough to be applied effectively in real-world clinical settings.”
This research is part of the NIMH’s Individually Measured Phenotypes to Advance Computational Translation in Mental Health program, an initiative aimed at utilizing behavioral measures and computational methods to create novel clinical signatures for mental health treatment. The program’s importance was recently highlighted by the White House Office of Science and Technology.
Related topics:
