Machine Learning to Infer and Control Brain State

Apr 17

This event has passed.

Wednesday, April 17, 2024 – 2:00PM to 3:00PM

Computational Medicine Seminar Series featuring David Carlson


David Carlson, Assistant Professor of Civil & Environmental Engineering

We are excited to announce the next session of the Spring 2024 Computational Medicine Seminar Series featuring David Carlson. This seminar will explore "Machine Learning to Infer and Control Brain State." It is increasingly possible to treat psychiatric disorders by making targeted interventions on the brain. However, finding an appropriate protocol requires many choices. We propose a method that identifies electrical dynamics across brain regions related to illness states or behaviors and employs these patterns to design intervention protocols. Specifically, we design machine learning methods that can infer electrical networks from brain data, and the expression of these networks define a brain state that predicts disease state, behavior, or outcomes. These networks are interpretable in their spectral power and directional relationships between brain regions, facilitating the design of testable protocols on key relationships. We will provide a case study on how these techniques can be used on social aggression. Using an inferred brain network related to social aggression, we develop a closed-loop protocol that intervenes only when necessary to reduce aggressive behavior. This machine-learning controlled protocol suppresses aggression, but not pro-social behavior, whereas an open-loop (constant) stimulation procedure suppresses both. I will end with a discussion on our current efforts in causal discovery and mediation analysis to further understand and improve this system.

Join us Wednesday, April 17th at 2:00 pm for this VIRTUAL event at the following Zoom meeting link:

Join us for an IN-PERSON post-event coffee hour and bites, where great ideas will be brewed over a cup, and you can meet the seminar speaker!
Location/Time: Wilkinson 1st Floor, 3:00 - 4:00 PM