May
Tuesday, May 27, 2025
8:00-9:00am PT
Watch Recording Here
Killian Pohl, PhD
Professor of Psychiatry & Behavioral Sciences and, by Courtesy, Electrical Engineering
Stanford University
Title: Crafting Machine Learning Models for Neuroscience Discovery
Abstract: Machine learning has had limited impact on improving the diagnosis and prevention of psychiatric diseases as their findings often fail to generalize beyond the neuroscience data they were trained on. In this talk, I will review the most critical challenges in using machine learning to advance discovery in neuroscience. For example, the presence of confounding effects often results in data-driven inference identifying spurious and biased associations. I will show that traditional approaches are often unsuitable for minimizing their effect on 3D brain MRI studies and propose alternative strategies, such as augmenting training data via synthetic 3D MRI generated by conditional diffusion models. I will review findings of the proposed deep learning approaches on large publicly available data sets (such as ABCD study, > 10K samples) and smaller in-house studies (< 100 samples).
Dr. Pohl is a Professor of Psychiatry and Behavioral Sciences and, by courtesy, of Electrical Engineering, and the Director of the Computational Neuroscience Laboratory (CNSLab) at Stanford University. The focus of his laboratory is to advance computational neuroscience in identifying biomedical phenotypes that enhance personalized medicine toward the diagnosis and prevention of psychiatric disorders from childhood to old age. The CNSLab identifies phenotypes by coupling findings from unbiased, machine learning-based searches across highly dimensional biological, cognitive, neuroimaging, and behavioral data with insightful interpretations by Dr. Pohl’s clinical collaborators. Dr. Pohl is the principal investigator on awards from Stanford’s Institute for Human-Centered Artificial Intelligence and the National Institute of Health (NIH). Before joining Stanford, Dr. Pohl received his Ph.D. in computer science from the Massachusetts Institute of Technology and was faculty at Harvard, IBM Research, the University of Pennsylvania, and SRI International.