Amp Schizophrenia: Clinically Actionable Lessons From the Largest (and Ongoing) Study of Clinical High Risk for Psychosis-View Presentation

Video Transcription

Video Summary

The session introduced the AMP Schizophrenia project, a large, ongoing NIMH-funded public-private partnership studying people at clinical high risk for psychosis. Dr. Carrie Bearden explained that the project aims to improve prediction of who will convert to psychosis or experience other significant outcomes, using longitudinal data from over 2,100 high-risk participants and about 650 healthy controls across 42 sites. The study collects multimodal data including clinical interviews, cognition, MRI, EEG, blood and saliva biomarkers, smartphone data, and speech samples. She also highlighted that the data are publicly available through NIH archives and that the project has already produced a new harmonized symptom measure, the PSYCS.<br /><br />Brandon Staglin emphasized the importance of integrating lived experience into mental health research. He argued that including people with psychosis improves science, ethics, feasibility, and trust, and he described how lived experience leaders have shaped AMP Schizophrenia’s design, outreach, data sharing, and participant burden reduction. He shared best practices for meaningful inclusion: build an inclusive culture, create sustainable systems, and provide adequate resources.<br /><br />Dr. Phil Wolfe presented early results on language as a biomarker. Using AI and natural language processing on structured and open-ended interviews, his team identified semantic and discourse features that distinguished clinical high-risk participants from controls and showed promising predictive power for risk classification.<br /><br />Dr. John Torres discussed digital biomarkers from smartphones, including surveys, activity, sleep, mobility, GPS-derived features, screen time, and audio diaries. He showed that participants were generally willing to share data and that digital patterns may help stratify risk, though careful validation is still needed.<br /><br />The talk ended with a discussion of exclusion criteria, terminology, data integration, and practical challenges in using these biomarkers clinically.

Keywords

AMP Schizophrenia

clinical high risk

psychosis prediction

longitudinal study

multimodal biomarkers

MRI EEG

speech analysis

lived experience

natural language processing

smartphone data

digital biomarkers

public NIH data