NSF gives nearly $5 million to Brown faculty working on IT

Five cutting-edge computer science projects at Brown were among the winners of the highly competitive National Science Foundation awards for information technology research. These extraordinary projects — which, if fully realized, will allow medical students to participate in a surgery that may have taken place months ago, or quadriplegics to move a robotic arm simply by thinking about it — won the scientists nearly $5 million of the $156 million distributed. See additional articles:

• Charniak and others push into new areas of speech recognition
  
• Van Dam project hopes to marry 3-D graphics with interactive electronic books that one day may train surgeons using virtual reality
  
• Upfal project explores dynamic behavior of networks
  
• Van Hentenryck in the hunt for algorithm that takes uncertainty into account
  

by Cynthia Ferguson

It’s a project that almost borders on science fiction, but the work of an interdisciplinary team at Brown could move the medical world closer to giving those who are severely disabled some real mobility. If the goal of this team is ultimately reached, victims of spinal cord injuries, such as actor Christopher Reeve, will be able to operate a robotic arm or other device simply by thinking about it.

Michael Black of computer science (top left), Elie Bienenstock of applied mathematics (middle left) and John Donoghue of neuroscience (bottom left) were awarded $446,970 from the National Science Foundation to pursue this ambitious project. It involves modeling the behavior of brain cells so that a computer chip implanted in the brain can interpret the signals and relay them to a robotic appliance.

The task, of course, is daunting because the brain is so enormously complex. "There are more cells in the brain than there are people on the planet," notes Black. "Being able to see what all of them are doing would be like trying to listen to every person in the world."

What they are doing is studying 25 to 100 brain cells — considered by today’s standards a large sample — to make optimal inferences about the cells’ individual behavior and intentions. They are trying to determine, for example, which cells tell us to raise an arm and which ones move it to the right.

When a person has lost all mobility, the path between brain and limbs has been cut off or impaired. "What we’re trying to do is restore that pathway," says Black. "More important, we want to restore that person’s pathway to the world."

Even a crude robotic arm — one that can just lift a drink, turn up the heat or turn off the television — gives a quadriplegic some control over his or her environment, Black notes. If successful, the coupling of the appliance and the person would become so tight that it would function as an extension of the body.

"When you’re first learning a musical instrument, you’re aware of everything your fingers are doing," Black says. "Eventually it becomes automatic. The instrument becomes part of you."

The project requires the expertise of all on the team. Donoghue is gathering the neural data and exploring basic neuroscience questions to determine how the brain "codes" information. Bienenstock, who is trying to describe how the brain works through the language of mathematics, crunches the numbers.

"The mathematical challenge is enormous," Bienenstock concedes, but he is encouraged by the work already done. With Donoghue’s data, Bienenstock’s numbers and Black’s computer models, they have begun to replicate some crude arm movements made by monkeys.

"Each cell has a task," says Bienenstock. "Each cell has a preference. But it’s just numbers."

The robotic arm envisioned by Black would look something like a human arm, featuring joints and a crude "hand" that can open and close. It might be attached to a wheelchair someday, allowing a disabled person to move around in the world and interact with objects. The same technology, Black notes, could extend to a service robot on wheels that, like the arm, would be operated from the brain.

One difficulty Black faces in developing the necessary software is that the brain is always trying to adapt to new influences and is constantly changing. This means the software must adapt as well.

Daunting as the challenges may seem, Black is optimistic that the project will yield tangible results. "We’ve assembled a terrific team of people with all of the necessary expertise," he says. "Brown is really good at this kind of collaborative effort."

Photos by Glenn Turner