When the brain processes sound, neurons – signal-firing cells – spike, and in different ways, depending on factors such as whether the sound has a high or low pitch.

Vision, smell, hearing – this kind of "dynamic sensory information" – relies on neurons firing in various patterns.

University of Illinois Professor Todd Coleman wants to understand those patterns better, with the idea of being able to model them mathematically, test the models statistically and, eventually, replicate the action in the brain in "engineered systems" – which could be computers or hearing aids – designed to operate in similar fashion.

"Not only is it interesting for pure science, but it has practical applications," the UI electrical and computer engineering professor said recently of his research in "computational neuroscience."

It could even help lead to brain-controlled video games where you "think" the action on the screen, or contribute to systems that allow people with physical disabilities to control prosthetic devices with their thoughts.

Coleman's interest in computational neuroscience – at one level, creating quantitative ways to characterize how the brain functions – goes back to graduate school at the Massachusetts Institute of Technology. Friends with a biological and medical bent urged him to apply the intricate math he was using in solving communications problems to frontiers in bioscience.

His doctoral work focused on compressing data distributed over computer and communication networks and interference problems in wireless communications. He still works on developing methods to make com-municating information simpler and more reliable.

He and colleagues already have a patent on a technique for shoehorning additional pieces of information between the informational "packets" sent over a network, such as the Internet, making better use of available bandwidth.

Coleman thought about concentrating on bioinformatics in his postdoctoral work – that is, methods for juggling and making sense of the gigantic piles of biological data scientists are generating in studying genes, proteins and the like.

Then, the idea of investigating how the brain represents information presented itself.

"That seemed really compelling to me," he said.

He ended up studying with Emery Brown, an MIT and Harvard brain researcher who is both a medical doctor and a statistician, in part because he figured Brown's doctorate in statistics gave them a common professional language.

A Dallas native who did his undergraduate work at Michigan, Coleman landed at the UI a year ago, attracted by its highly rated computer and electrical engineering program, collegial atmosphere and interdisciplinary research focus, and at the urging of his doctoral adviser at MIT, Muriel Medard, who started her career here.

Medard had never pushed him in a particular direction during their time working together, so it stood out when she did.

"She told me I'd be a complete idiot not to go to Illinois," Coleman said.

UI Professor Richard Blahut, head of the Electrical and Computer Engineering Department, hired Coleman even before his postdoctoral stint.

"We hired him a year before he came," Blahut said. "I wanted him. He's very intelligent and has a very strong research reputation. He's got a very strong, likable personality. He's a team builder. He'll be a good role model for young faculty and students."

Coleman's lab is preparing to begin looking at how the brain functions in set situations by using, among other tools, electroencephalography, or EEG, to capture electrical signals from volunteers as they perform designed tasks at a computer wearing something like a swimming cap festooned with sensors.

Coleman said discerning the brain's modus operandi, even on known tasks, presents challenges because the action in the brain is dynamic and even brain structure is "plastic," meaning changeable.

The brain also tends to work in a probabilistic fashion, reacting in a way that appears most likely to be the proper response.

Still, Coleman said, researchers can probe systems of the sort by, say, taking a reductionist approach, changing individual variables to see what happens and to get a piece of the puzzle, with enough pieces assembled over time to make a larger picture.