Undergraduate research contributed to BrainGate

Paninski '99 developed key algorithm used in device

When Matt Nagle plays Pong on his computer, people watch with amazement.

He doesn’t win all the games, and most of the time he plays against himself, but he doesn’t use his hands or any other part of his body to control the paddle on the screen. Nagle is paralyzed and plays using his mind, thanks in part to Liam Paninski ’99.

Brown researchers’ creation of BrainGate – a device that allows quadriplegics to operate computer cursors, and even robotic arms, just by thinking about it – is now well-known and has been featured on television and in print media. But few of those familiar with BrainGate have heard of Paninski, who spent two years as an undergraduate attempting to decode brain activity – research that allowed for the creation of BrainGate years later.

Paninski played the role of translator, allowing technicians and researchers to give meaning to the mess of the electrical impulses sent out by the brain. Nearly a decade later, four people with quadriplegia – total paralysis from the neck down – can perform tasks once thought impossible using the power of their minds.

Professor of Neuroscience John Donoghue, lead researcher for the BrainGate program and head of the lab where Paninski worked, recounted his first encounters with the ponytail-wearing kid who “fit the mold of a hippie.”

“He was one of those people that was on fire with excitement,” Donoghue said. “He wanted to do an independent study project right away.”

In his independent project, Paninski researched brain activity and multiple-neuron recordings. The math-intensive task of analyzing brain activity came easily to Paninski, who Donoghue said was “extraordinarily skilled in mathematics.”

“It changed the way my life went,” Paninski wrote of his undergraduate research in an e-mail to The Herald. “I never really pictured myself pursuing research as a career until I worked in a lab at Brown.”

Donoghue recalled a highly advanced book called “Spikes” that discussed brain signals and how they can be combined to understand what the brain is attempting to do – look left or right, move an arm or raise an eyebrow.

“(Paninski) went off and read the book, and probably knew more than anybody I knew at the end of a weekend of reading – including me,” Donoghue said.

Using his strong mathematic ability, specifically in statistics, and rapidly growing knowledge of brain signals, Paninski modified an existing algorithm and adapted it for the interpretation of brain activity.

“He took existing mathematics and applied it to that problem in a way that was novel and creative,” Donoghue said.

Paninski’s algorithm was applied to a device Donoghue would later develop as a key component of BrainGate. The specialized device picks up brain signals, which Paninski’s algorithm translates into something meaningful. This way, the crackles and pops emitted by the brain can be seen as motor functions, such as the movement of limbs.

The BrainGate project began after Paninski graduated, but it is in part thanks to the technology he helped develop as an undergraduate that Nagle is now able to play Pong.

“If you damage your spinal cord, then the messages from the brain can’t get to the muscles, so you’re paralyzed,” Donoghue said. “But the part of the brain that puts out those movement signals is still functional.”

Using Donoghue’s device as the hardware and Paninski’s algorithm as part of the software to interpret brain activity, BrainGate was implanted in Nagle in 2004. A videotape of the trials shows Nagle staring at a prosthetic forearm and hand, commanding the hand to open and close using his mind.

“(Paninski) left a lasting impression,” Donoghue said. “In the laboratory, he was a driving force as far as solving the problems of understanding how the brain was coding information.”

Paninski also started a Brown tradition known as the Pop Code meeting, a weekly brainstorming session for faculty and students from the fields of mathematics, engineering, physics, computer science, neuroscience and cognitive science.

“That was 10 years ago, and it’s still going,” Donoghue said. “People still talk about (Paninski’s) amazing abilities.”

Now 28 years old, Paninski is already an assistant professor of statistics at Columbia University after a position at University College London.

“People just don’t have career paths like that without being extraordinary,” Donoghue said.

Paninski praised the University’s focus on undergraduates.

“Brown is pretty special in how undergraduate students are encouraged to do research, welcomed in labs and given exciting, important projects to work on,” he wrote.

Though Paninski’s contribution was unusually significant, he wrote that he believes professors can learn quite a bit from their students.

“It’s part of research that students will discover new facts and ideas that (professors) hadn’t thought of before,” he wrote. “It happens a lot to me now that I’m a professor.”

Paninski wouldn’t comment on how Brown measures up to Columbia, but he did offer his view on his experiences at both schools.

“They’re both great from a research perspective,” he wrote. “Although I think Brown is a better place to be an undergraduate – more freedom.”