Large Banner Ad
Small Banner Ad

January 10, 2013

Spaun: A human brain model that can think

The Canadian Charger

More by this author...

University of Waterloo professor of systems design engineering Chris Eliasmith and his team of researchers has not only created the world's largest functional brain model, they've created the only one that can perform the specific tasks researchers ask it to do, according to Terry Stewart, post doctoral research associate and project manager for the brain model called Spaun.

Spaun, the model can see, remember, think and write using a mechanical arm. Mr. Stewart said that while other researchers randomly connect brain neurons and see what the computer model of the brain does, his team developed software that can program Spaun to respond to specific requests.

He said the team build the software in the lab using the concepts in the book Neural Engineering by Chris Eliasmith and Charles Anderson.

“We give it (Spaun) hundreds of components and the software takes these instructions and connects thousands of neurons for each component.”

For example, if Mr. Stewart asks Spaun to perform some addition or memorization, the software program will tell Spaun how to connect the neurons and the rates they should fire at. He showed the computer model in action, with some neuron connections firing at a noticeably faster rate than others, when memorizing a sequence of numbers.

“There could be thousands of neurons all connected to each other. That means millions of different connections. The software program tells the brain model how to connect the neurons and how they should function.”

When memorizing a sequence of numbers, the computer screen not only shows neurons in Spaun interacting with pulsating lines, it shows what Mr. Stewart called a “thought bubble” which tells the researchers what Spaun is “thinking,” while these neurons are firing. In the example Mr. Stewart demonstrated, the thought bubble had a sequence of number, with the number 8 fading.

“The number eight is fading so it's forgetting that number,” Mr. Stewart said. “It's exciting because the model is making the same mistakes as people make. If it works, it should make the same kind of mistakes as people. This is part of what makes us think we're on the right track, because we didn't program in these mistakes.”

To simulate the ageing process, Mr. Stewart said he and his team randomly kill off neurons in the computer model and witness it gradually increasing its number of errors. Because neuro-scientists know how individual neurons are damaged by ageing and diseases such as Parkinson's disease, Mr. Stewart said he can damage the neurons that way in the computer brain model and see how the model behaves when performing a variety of tasks.

“Now we can look at treatment. Once we understand how a big chunk of the brain reacts, this should lead to treatment.”

To illustrate the potential this research has, he used the example of deep brain stimulation – a technique used now for some Parkinson's patients – whereby doctors send electrical impulses into the section of the brain affected by Parkinson's disease, by inserting a wire. The result is that some patients who were barely able to move are able to get up and move freely.

While he said it's possible that at some point, doctors may be able to replace parts of the brain, that's a long way off. But he thinks the current research could lead to the brain being able to control artificial limbs.

Meanwhile, the educational aspect of learning how the brain functions is an ongoing process which Spaun is making contributions to. Mr. Stewart said his team is learning from the computer model how people memorize things and may thus be able to arrange material in a way to make it easier for the brain to memorize it.  And he foresees a time when researchers will be able to see how the brain is functioning when a person becomes trapped in incorrect ideas that he or she has difficulty letting go of.

However, in part, because Spaun has only 2.5 million neurons and the human brain has 10 to 100 billion, Mr. Stewart stressed that – although they've made great strides - researchers are at the very, very beginning of the process of understanding the brain.

He said this helps to explain why researchers all over the world are dependent upon each other to advance our understanding of the brain.

“All of our research is available publicly. That's the currency of academia – so other people can build on it. Many other people make use of that information. We encourage other people to do research because it must be a community effort.”

  • Think green before you print
  • Respond to the editor
  • Email
  • Delicious
  • Twitter
  • Facebook
  • MySpace
  • StumbleUpon
Subscribe to the E-bulletin

M. Elmasry

Subscribe to our YouTube Channel