April 14, 2023: Learning Opportunity Sparks New Insights

In image:  Jamie Graham, Masters of science student

Pictured above:  Jamie Graham, MSc student, Contributing Researcher

Participants in Dr. Lyle Muller (Mathematics) and Dr. Marieke Mur (Psychology, Computer Science)’s graduate course on “neural networks” explored the intersections between artificial intelligence and neuroscience. Last week, learners showcased their findings on neural activities—both biological and artificial.

Mauricio Cespedes Tenorio (MEsc student, Western) and Mohamed Yousif (MSc student, Western) discovered that humans and machines may be fooled by the same optical illusions. Their work with Convolutional Neural Networks—a type of artificial network that processes visual information—also helped them understand the neuroscience of sight in humans. As Tenorio and Yousif explained:

“The same way biologists use different models like cells and mice to test different theories and hypotheses, we use computational models to test our ideas of how the brain works. This course took the approach of showing us and breaking down different models of neurons and neural networks from earlier biophysical models to other more abstract models, which helps understanding better the principles behind these computational algorithms.”

This learning experience was organized as part of an ongoing partnership between the “Mathematics of Neural Networks” research theme at the Western Academy and the Fields Institute for Research in Mathematical Sciences. Participants included graduate and senior undergraduate students from Western and other Ontario universities.

The course is part of the research theme’s broader effort to foster a research community that is committed to the advancement of computational neuroscience and to the discovery of possible clinical applications.
Course participants also included contributing members of the “Mathematics of Neural Networks” research theme. Jamie Graham is a machine learning specialist and MSc student, who is currently studying under the supervision of Western Fellows Muller and Dr. Ján Mináč.

For Graham, the course played an important role in shaping his contribution to the “Mathematics of Neural Networks” theme. As he explained, “The course gave me more of an appreciation for machine learning in the context of computational neuroscience.”

As a participant in Muller and Mur’s course, Graham discovered new insights into a set of mathematical models known as Mackey-Glass equations. Biologists use Mackey-Glass equations to describe various physiological processes, including neural activities.

“From a neuroscience perspective, we are working towards understanding the mechanisms of visual processing and memory,” Graham said.

“One particular application of my research is that of understanding the mechanism of memory consolidation, which may be mediated by a particular rhythm of brain activity called sleep-spindles. This oscillatory activity in the brain during sleep has been found to be associated with memory performance.”

Graham’s enthusiasm stems from the idea that mathematical innovation may spark important medical discoveries.  As he explained:

“If we understand how memory consolidation occurs in the brain at a fundamental level, this may pave the road for design of therapies and interventions for memory-related neurodegenerative disorders such as dementia and Alzheimer's."
 
Learn more about the “Mathematics of Neural Networks”:  https://uwo.ca/academy/themes/neural_networks/index.html