Machine Learning software for adaptive medical treatmentsIn some diseases, such as chronic depression or epilepsy, medical treatments may need to vary or be adjusted over time. Professor Joelle Pineau is studying machine learning (computer programs that can automatically learn) techniques for optimal treatment sequences. Here is Professor Pineau's desciption her projects.
Automated learning of adaptive treatment strategies for chronic depression: Most chronic disorders, such as depression, epilepsy, substance abuse, HIV/AIDS, require clinical treatment over a long-term period. Consequently the best clinical care requires adaptive changes in duration of treatment, treatment type and dose over time. Adaptive treatment strategies use decision rules to individually tailor a course:w of treatment for an individual patient. The goal for computer scientists is to automatically learn optimal treatment sequences from data collected during multi-stage randomized studies. The STAR*D trial, funded by the US National Institute of Health, was the largest depression study ever conducted. We are currently using computational techniques (e.g. reinforcement learning) to automatically extract individualized optimal treatement strategies from data collected during this study.
Computational modelling and adaptive treatment of epilepsy: The goal of this project is to investigate the use of reinforcement learning to optimize deep-brain stimulation strategies for the treatment of epilepsy. We begin by investigating the problem of automatic seizure detection from electrophysiological recordings. We are currently implementing a mathematical model of the brain neural network, which exhibits the synchronous patterns that are characteristic of epilepsy. We will then investigate the use of reinforcement learning to optimize treatment strategies using this computational model, as well as in-vitro data from rat hippocampal slices.
Some courses that are relevant for this area are: