Using Artificial Intelligence to understand how the brain interacts with the external environment

Faculty of Science
Cellular and Molecular Medicine
Students on campus near O-Train station.
Among the most common illnesses are mental or neurological disorders, which currently affect approximately 450 million people worldwide, and are increasingly known to impact physical health as well.

Fortunately, as mental health awareness grows, more energy and money is invested into research to enable more personalized treatment of psychiatric disorders.

A mathematician and computer scientist from the Department of Mathematics and Statistics, Professor Maia Fraser is part of a multidisciplinary research team that uses Artificial Intelligence to study the brain and its interactions with the external environment. She helped assemble this group together with Dr. Georg Northoff of The Royal’s Institute of Mental Health Research, a leading neuroscientist and clinician whose research highlights the role of brain-world alignment in mental health. The aim of their joint work is to eventually create not only novel diagnostic and therapeutic tools for mental health professionals, but progress in AI as well. Together, Profs Fraser (co-PI) and Northoff (PI) represent the Canadian segment of this collaborative endeavour led jointly between Canada and the United Kingdom. This international team was awarded more than $1.5 M over three years through the Canada-United Kingdom Artificial Intelligence Initiative.

Maia Fraser

Professor Fraser is a key force on the mathematical and computational side of the project. Her work in hierarchical learning focuses on ways biological and machine systems make use of hierarchy to enable learning. In particular, the methodology of the current project was precipitated by bringing the Free Energy principle of Lead PI Karl Friston (University College, London) to bear on the brain-world alignment paradigm of Northoff and the ongoing questions studied by Fraser and Prakash Panangaden at McGill/MILA in hierarchical reinforcement learning. Together with John Griffiths (UToronto), who specializes in advanced math tools in computational neuroscience, this team is developing novel Math-AI methods to capture patterns and structure within the complex activity of the brain.

Significant impact is expected to be created through this project. First, economic impacts will be generated through links to various AI companies, offering a direct pathway for commercialization of the novel mathematical-computational model being created. Clinical impacts are expected in the area of mental health, through improved diagnostic and therapeutic tools for use by mental health professionals. An example would be better diagnostic classification of psychiatric disorders such as schizophrenia and depression, as well as AI-guided personalized therapy to shift brain activity away from problematic patterns. The Canadian team will implement the new model in clinics in Ottawa (Royal Ottawa Mental Health Centre) and Toronto (UToronto/CAMH). Finally, the ethics portion of the project will lead to social impacts, such as guidelines and criteria for the human self in its encounter with the artificial self; i.e., what it means to be humans.

Read more: