From Our 2008 Archives

Monkey Controls Robotic Arm With Brain

By Alan Mozes
HealthDay Reporter

WEDNESDAY, May 28 (HealthDay News) — Relying solely on brain signal manipulation, monkeys have learned to operate human-like robotic arms to feed themselves, U.S. researchers reported Wednesday.

This cutting-edge development in the field of neuro-prosthetics was made possible by linking up neural pathways in the voluntary movement region of the monkey's brain to a specially designed computer software program. In turn, the monkey's mental "firings" were turned into fluid and natural prosthetic movements — enabling arm-restrained primates to grip and eat marshmallows and fruit with a claw-like robotic hand.

"This a step on the way to fully mobile arms and dexterous prosthetic hand that will be controlled with brain activity recorded directly from the individual brain cells," said the study's lead author, Andrew Schwartz, a professor of neurobiology at the University of Pittsburgh School of Medicine.

The research, published Wednesday in the online edition of Nature, is being touted as a significant advance towards designing functional prosthetic devices for fully paralyzed individuals.

As such, it is a real-world leap beyond Schwartz' own previous endeavors, which explored similar brain-machine connections that would enable monkeys to use their thoughts to manipulate cursor movements on a computer screen.

As with the earlier effort, the current work, done by a research team at the University of Pittsburgh and Carnegie Mellon University, focused on the primary motor cortex — the part of the brain where thousands upon thousands of nerve cells fire in unison to issue voluntary movement instructions.

Unable to chart the massive number of neural firings typical of movement control, the researchers inserted probes the width of human hairs into approximately 100 neurons located in this brain region in two rhesus monkeys. The probes were then linked directly to a computer outfitted with a complex piece of software capable of reading and expanding upon the selected electrical impulses, to fill in the blanks and paint a full picture of movement intent.

The researchers then restrained the monkeys' own arms, after which they used a computer to instruct the robotic arms to grab food and bring it to the monkeys' mouth.

After repeatedly observing the computer-controlled process, the monkeys were given several hours a day to practice launching and mimicking the robotic movement simply by means of neural signaling. After several days of trial and error, the monkeys eventually came to take full control of the arms — displaying a consistent adeptness at initiating and completing brain-signal-controlled prosthetic movements.

The researchers concluded, in fact, that the monkeys ultimately came to view the arms as extensions of their own bodies.

Schwartz said that he and his colleagues are already working on improving their current achievements.

"Right now we can control the arm very well in three dimensions," he explained. "We can reach and grasp with a simple gripper. But there's no dexterity, no fingers. So we're trying to go way beyond what we've already accomplished by adding wrist and finger control in the next year or two. And we're making good progress."

"And in the next couple of years I'm quite sure there'll be humans walking around with these kind of implants," Schwartz added. "I don't think it's very far away at all. In fact, other teams have already implanted a few people with versions of these kind of devices — although they haven't yet matched the kind of performance we achieved with our work."

John F. Kalaska, a professor of neuroscience at the Universite de Montreal in Quebec, said that while several different labs are busy exploring the promise of neuro-prosthetic science, Schwartz's is the first to successfully put the various pieces of the puzzle together.

"This study team is by no means the only one looking into this, and much of this technology has been used elsewhere," noted Kalaska. "But this team has clearly shown proof of principle that the concept is workable."

"This is important," he added, "because there are hundreds of thousands of neurological patients that have severe motor difficulties caused by disease and spinal cord accidents. Severely disabled patients. And they have no quality of life, they have no independence. So if you can somehow develop a technology that allows these people to simply think about what they want to do and have a device simply do it for them as a result, you can give them a quality of life that they didn't have before. You can take their intentions, and turn them into reality."

SOURCES: Andrew Schwartz, Ph.D, professor of neurobiology, University of Pittsburgh School of Medicine, Pa.; John F. Kalaska , professor of neuroscience, Universite de Montreal, Quebec; May 28, 2008, online edition, Nature

Copyright © 2008 ScoutNews, LLC. All rights reserved.