Impact at the intersection of neuroscience and AI

By Alison Palmer, Evaluation and Special Projects Lead

As part of our mission to fund bold brain research, Brain Canada has, together with its donors and partners, supported dozens of large-scale research projects and platforms to develop and apply cutting-edge AI approaches to advance our understanding of the brain and identify solutions for brain diseases and disorders.

In fact, the most highly cited Brain Canada-funded publication, co-funded with CIFAR, is from AI pioneers Drs. Yann Le Cun, Yoshua Bengio and Nobel Prize Laureate Geoffrey Hinton. Dr. Hinton was awarded the 2024 Nobel Prize in Physics with John Hopfield for foundational discoveries and inventions that enable machine learning with artificial neural networks.

This accomplishment – and the impact it enabled, within and beyond academia – is a testament to the value of supporting high risk, high reward research. Right now, through the analysis of vast datasets from neuroimaging, genetic studies, and clinical records, AI tools are enhancing researchers’ ability to uncover the mechanisms underlying brain conditions and their progression, promising earlier detection and better, more personalized treatments.

Here are just a few examples of how Brain Canada-funded projects at the interface of neuroscience and AI are improving outcomes.

Improving the diagnostic accuracy of routine EEG testing for epilepsy

Dr. Elie Bou Assi at l’Université de Montréal used his 2022 Future Leader in Canadian Brain Research grant to test a bold idea – could AI help analyze routine electroencephalograms (EEGs), tests of the brain’s electrical activity, to improve the accuracy of epilepsy diagnosis?

Currently, a routine EEG will identify the tell-tale signals of epilepsy, known as spikes, in less than half of patients with the condition. Misdiagnosis remains a challenge in clinical practice, highlighting the need for complementary tools to support EEG interpretation. Dr. Bou Assi and his team sought to address this challenge with DeepEpilepsy, a machine learning model they developed to automate the processing of EEGs. They trained and tested the model on over 800 EEGs. DeepEpilepsy was able to detect subtle signal patterns linked to seizure risk, even when no visible spikes were present. The model achieved an area-under-the-curve (AUC) of 0.76, compared to 0.69 for standard EEG review based on visible spikes. When combined with review based on spikes, Deep Epilepsy’s performance rose to 0.83.

This proof of concept shows that AI can detect subtle brain-wave patterns specialists may miss, offering a promising way to enable earlier diagnosis, earlier treatment, and improved outcomes. With these results in hand, Dr. Bou Assi and his team successfully obtained additional funding that will allow them to conduct a validation study at the Centre Hospitalier de l’Université de Montréal (CHUM).

Beyond the scientific outcomes, the grant enabled close collaboration between engineers, neurologists, and data scientists, accelerating the translation of AI methods into tools that could ultimately improve diagnostic accuracy and accessibility in epilepsy care.

By revealing patterns that aren’t visible to the naked eye, AI gives us a new way to read EEGs and to better understand seizure risk. The Future Leader grant from Brain Canada made it possible to take this step from concept to real-world testing.
Dr. Elie Bou Assi

Read more here.

Better detecting metastatic brain cancer without surgery

Dr. Reza Forghani and colleagues at McGill University and international partners received a Spark Grant in 2021 from Brain Canada, the Canadian Cancer Society and CIHR to create an AI model that detects the spread of metastatic brain cancer using routine MRI scans—without aggressive surgery.

The proof-of-concept study tested the AI model using MRI scans from over 130 patients who had surgery to remove brain metastases at The Montreal Neurological Institute and Hospital. Comparing the results to what doctors observed by examining brain tumour samples under the microscope, the team found that the AI model identified cancer cells in surrounding brain tissue with 85 per cent accuracy.

Brain metastases are the most common type of brain cancer, occurring when cancer cells from other parts of the body spread to the brain. Surgery to confirm this spread is often risky or impossible. By spotting subtle MRI patterns often too faint for the human eye, the AI model Dr. Forghani and colleagues developed offers a non-invasive way to guide treatment decisions earlier and more accurately. With further testing and development, this technology could help doctors identify patients at higher risk of tumour regrowth and tailor therapies, potentially improving survival and quality of life for people living with metastatic brain cancer.

Read more here.

Supporting shared decision-making between patients with major depression and clinicians

Dr. Manuela Ferrari at the Douglas Research Centre, an expert in bridging knowledge and practice, is supported through the Bell Let’s Talk–Brain Canada Mental Health Research Program for her project testing an AI-based smartphone app designed to better manage the treatment of depression.

The tool, which was developed by a startup in Quebec, boosts communication between clinicians and patients, enabling clinicians to monitor their patients’ symptoms and responses to medication between appointments. Combining the best clinical evidence with machine learning, the tool uses the data gathered to propose a personalized treatment plan and support shared decision-making between patient and clinician. Dr. Ferrari and colleagues tested the efficacy and safety of the tool via a multicenter, randomized controlled trial across nine sites.

Almost 30 percent of patients using the tool with their clinicians achieved remission compared with none in the control group, with symptoms improving significantly faster. No tool-related adverse events were reported. These findings provide strong preliminary evidence that AI can uncover subtle patterns clinicians may miss, offering a promising way to speed and improve depression care.

We need to extend care outside the walls of traditional health-care settings, and outside these one-time interactions. We need to be present where the person is. And this is where technology can help us.
– Dr. Manuela Ferrari

Brain Canada