Human-machine links could operate robots

By CHARLES CHOI, United Press International  |  May 3, 2004 at 12:39 PM
share with facebook
share with twitter

NEW YORK, May 3 (UPI) -- U.S. scientists have developed a technique to record brain signals that could lead to controls for robotic limbs for patients suffering from paralysis or neurological diseases.

Scientists from Duke University Medical Center in Durham, N.C., recorded brain signals from 11 volunteers who suffer from either severe body tremors or refractory Parkinson's disease, a version of the illness that does not respond to medications.

Standard treatment for such patients involves surgery to implant permanent electrodes, which use small electric currents to stimulate areas deep in the brain, to help relieve symptoms. Neurosurgeons record the signals of a patient's brain during the operation to ensure the electrodes are placed in the most effective location.

Before implanting electrodes in the experiment, the investigators recorded each patient's brain electrical signals with arrays of 32 microelectrodes, each thinner than a human hair.

"We used volunteers in surgery where recordings were already being performed," Dr. Parag Patil, a neurosurgery resident at Duke, told United Press International.

Patients generally remain awake during surgical implantation of therapeutic electrodes. They experience no pain because there are no pain receptors inside the brain. The volunteers reclined in a semi-sitting position in front of a video screen.

"We had patients perform a task -- essentially play a video game with their hand," Patil explained. The volunteers held an electronic squeeze ball in one hand and were instructed to grip the ball to reach a target level of force as quickly as possible.

"Squeezing a ball is a very intuitively simple thing for a patient to do when they're having surgery. It's not distracting, they can learn it very quickly, perform it very easily," Patil said.

The volunteers played the video game for up to 10 minutes during the surgical procedure.

"We found we were able to get signals from the brain that predicted what the patients were doing with their hand," said Patil, who called the discovery a key step in developing a device that could allow patients to perform similar tasks even if they lacked muscular control of their hand.

"We only had five minutes of data on each patient, during which it took a minute or two to train them to the task," added Miguel Nicolelis, a Duke neurobiologist. "Despite the limitations on the experiments, we were surprised that our analytical model can predict the patients' motions quite well."

The Duke team will present its findings Tuesday in Orlando at the annual meeting of the American Association of Neurological Surgeons. They also will publish their results in the July issue of the journal Neurosurgery. The Defense Advanced Research Projects Agency and the National Institutes of Health supported the research.

"The goal of our project is to develop a neuroprosthetic brain-machine interface that will allow signals from the brain to be interpreted and allow people with handicaps to perform actions they wouldn't be otherwise able to perform," Patil said.

"Someone with quadriplegia who couldn't move their arms and legs could have a prosthesis that takes signals directly from their brain to operate a prosthetic arm or wheelchair or computer mouse," he said. "Our goal is to have a functioning neuroprosthetic prototype in the next three to five years -- if not sooner."

Although research by other groups has shown patients with individual implanted electrodes can direct a cursor on a computer screen, devices as complex as prosthetic limbs would require data from the kinds of large electrode arrays the researchers used in these new experiments.

"If you really want to extract motor commands processed in the brain, you have to use multiple electrodes. It's the only way to get the information with the requisite accuracy you need," John Chapin, a neuroscientist at the State University of New York Downstate School of Medicine in Brooklyn, told UPI.

Earlier studies from the Duke lab revealed monkeys could learn to control robot arms using only their brains. While those experiments recorded signals from the cortical surface of the brain, these new studies took signals from deeper in the brain in subcortical areas.

"This shows that one can extract information not only from cortical areas, but from subcortical ones too," Nicolelis said. "This suggests that in the future, there will be more options for sampling neuronal information to control a prosthetic device."

Another Duke team member, neurosurgeon Michael Turner, explained that subcortical electrodes also had the benefit of greater stability because they are buried deeper.

"The way the brain works, all the signals for motor control are filtered through these deep regions of the brain before they reach the final cortical output. So, they are theoretically easier to record from than cortical areas," Turner said.

He added the subcortical areas are denser, meaning there are more cells from which to record packed into a smaller area.

"This opens up a new scientific frontier -- to help those who need it most, people who are paralyzed," said biorobotic engineer Mandayam Srinivasan, director of the Touch Lab at the Massachusetts Institute of Technology in Cambridge.

"It gives us access to their thoughts," Srinivasan told UPI. "It's exciting to be able to do this at all. One of the holy grails of research is to understand the human brain."

Chapin added that the research "is something that could tremendously help improve the quality of life of people who are really suffering quite terribly."

The Duke researchers have partnered with Plexon, a neurotechnology company in Dallas, whose customers include 150 domestic and international military labs, pharmaceutical companies, academic institutes and research hospitals run by such institutions as the National Institutes of Health, GlaxoSmithKline in Britain and the U.S. Air Force. Competing research is being performed by a team includes Cyberkinetics of Foxborough, Mass.

"These two groups are going to be the ones to watch," Chapin told UPI.

The researchers emphasize, however, many years of further development and clinical testing will be required before any devices can be available.

"Obviously there are liability issues here," Chapin noted. "What if the insulation that is on the electrodes turns out to have bad interactions with tissues, for example?"

They already have applied for federal approval to implant experimental electrodes long-term in quadriplegic patients. Such tests, conducted over the next three to five years, would implant arrays in specific brain regions.

Experimenters would ask patients to perform actions more complicated than merely squeezing. They would play video games where targets are moved through two-dimensional or 3-D space, Patil said. Researchers then would study which tasks are optimally controlled by each region. They also would test for 30 days or so, to make sure the recordings are reliable over time.

"We're moving very deliberately here. Everyone in the field is aware of the need for safety at every step," Patil said.

Other devices on the horizon to help patients with nerve or brain damage include "sensory neuroprostheses," Chapin said, which would to help patients recover feeling in their body.


Charles Choi covers research for UPI Science News. E-mail

Related UPI Stories
Trending Stories