The normative predictome for precision psychiatry

• Applicant: Dr. Thomas Wolfers, Ph.D.
• Subject Area: Biological Psychiatry ; Biological Psychology and Cognitive Neuroscience ; Human Cognitive and Systems Neuroscience ; Medical Informatics and Medical Bioinformatics
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 513851350

Mental disorders are affected by many different biological and environmental factors acting throughout development, yet their diagnosis and treatment is solely based on the assessment of symptoms. The reliance on symptom-based classifications and predominant case-control analyses have hampered the identification of clinically predictive markers. To improve this, I codeveloped the normative modelling framework which allows me to chart brain development and aging. Using this approach, I was able to capture a high degree of brain heterogeneity among patients with the same severe mental disorder. While mental disorders are considered to affect the brain as a network this outstanding characteristic has not been incorporated into this framework in a principled way. Therefore, I propose to chart the normative predictome for which I will predict the architecture or function of one brain region from the collection of other brain regions using machine learning methods. This approach yields predictome derived probability maps which indicate to what degree individuals deviate from the estimated predictive pattern across brain regions. My models will be estimated on one of the largest reference samples assembled to date, improve the performance of earlier brain charting approaches, can be used across imaging modalities and in rapidly acquired longitudinal designs. Combined these properties enable unprecedented possibilities towards more precision in psychiatry.

• DFG Programme: Independent Junior Research Groups
• International Connection: Netherlands, Norway
• Cooperation Partners: Professor Dr. Andre Marquand, Ph.D.; Saige Rutherford; Professor Dr. Lars T. Westlye, Ph.D.