Mind boggling: Spaun opens a window into how the brain ticks

Monday, February 16, 2015

Meet Spaun, the world’s largest model of a functional brain. It can’t beat chess champions or win big at Jeopardy. And its handwriting would make any school teacher sigh in dismay.

But what it offers — a window into the workings of human grey matter — has researchers around the globe salivating with excitement.

Developed by Chris Eliasmith and his multidisciplinary team at Waterloo’s Centre for Theoretical Neuroscience (CTN), Spaun combines 2.5 million simulated neurons with a visual recognition system and a simulated mechanical arm.

Just like its human counterpart, it can read, answer questions, play simple games and memorize lists. It can decode numbers written in unfamiliar handwriting and tackle basic logic problems. Like people, it even makes mistakes, faltering at complex questions or getting tripped up when lists get too long. And Spaun learns, adapting its behaviour based on feedback from the world around it.

The brainchild of Eliasmith, a Waterloo systems design engineering and philiosophy professor, promises to unleash a slew of discoveries and innovations. On the medical side, Spaun could shed light on the impact of Alzheimer’s disease or help pharmaceutical researchers predict the effects of potential new treatments.

“I think it’s something that really might revolutionize how we do drug testing, and come to understand the effects that chemicals have on behaviour,” Eliasmith predicts.

He’s collaborating with researchers at Stanford to develop computer chips that work more like the human brain, using noisy, low power processing units that act like neurons. That’s a huge advantage when it comes to energy usage. While the average desktop computer consumes 90 watts or more, the human brain uses just 25. If you’ve ever reached for your smartphone only to discover it’s out of juice, you’ll appreciate the benefits of squeezing as much computation power as possible out of limited battery life.

Eliasmith’s internationally-acclaimed work was recognized in February 2015 with the 2015 Natural Sciences and Engineering Research Council of Canada John C. Polanyi Award. The prestigious honour acknowledges an individual or team whose Canadian-based research has led to a recent outstanding advance in the natural sciences or engineering, and is given out only once a year.

His technology has been featured on the BBC, the Discovery Channel and the CBC, and in publications such as Wired and the New York Times, just to name a few. It’s also included in his new book How to Build a Brain, which explains how to build computer models of the human brain.

Eliasmith says his systems design engineering undergraduate education gave him a solid technical foundation along with the flexibility to take courses in philosophy and psychology.  It also equipped him well for conducting highly-interdisciplinary work.

Today, he appreciates that same atmosphere of cross-pollination in both systems and the CTN. He points out that the faculty members in Waterloo’s systems design engineering department come from a variety of different backgrounds including mathematics, kinesiology, and computer science, and engineering disciplines such as mechanical, biological, biomechanical and electrical.

“We’re interested in each other’s work and willing to talk about it and provide feedback,” Eliasmith says. “It’s a great environment for a researcher.”