Artificial Intelligence Unravels Brain Structure’s Role in Predicting Antidepressant Treatment Response: Insights from a Collaborative Study” Researchers from the Institute of Psychiatry, Psychology & Neuroscience (IoPPN) at King’s College London, the University of East London (UEL), and the University of Pennsylvania have employed artificial intelligence to analyze brain images in individuals diagnosed with major depressive disorder (MDD). The study, titled “Neuroanatomical dimensions in medication-free individuals with major depressive disorder and treatment response to SSRI antidepressant medications or placebo,” and published in Nature Mental Health, reveals that the quantity of gray and white brain matter in MDD can predict the response to traditional antidepressants (SSRIs) and placebo medication.
Gray matter, responsible for various functions including sensation, perception, movement, learning, speech, and cognition, and white matter, facilitating communication between different areas of gray matter, were found to play a crucial role in predicting treatment responses. Despite the global impact of MDD on over 320 million people, the lack of biomarkers for predicting treatment response has been a challenge. The research aimed to uncover distinctive brain mechanisms underlying the presentation of the illness.
Analyzing brain scans from 685 participants with MDD and 699 healthy controls, the study identified two distinct dimensions—Dimension 1 (D1) with preserved gray and white matter, similar to healthy controls, and Dimension 2 (D2) with widespread decreases. The findings suggest potential biomarkers for defining depression.
Professor Cynthia Fu, the study’s joint first author, emphasized the significance of the research in identifying biomarkers for depression using machine learning. The study also investigated how these dimensions related to the clinical response to antidepressants. Participants in D1 exhibited a significantly greater response to SSRI medication compared to placebo, while those in D2 showed no significant differences in the effectiveness of SSRIs or placebos, indicating a potential biomarker for early identification of treatment resistance.
Dr. Mathilde Antoniades, another joint first author, highlighted the collaboration with researchers worldwide sharing data from medication-free MDD participants. Professor Christos Davatzikos from the University of Pennsylvania emphasized the use of state-of-the-art AI methods in this unique dataset. Moving forward, researchers aim to define disease-specific dimensions in depression and those shared with other mental health disorders.
