TORONTO - In a high-tech variation of "monkey see, monkey do,'' U.S. researchers have taught two macaques to feed themselves with a human-like robotic arm using only signals from their brains.
The scientists at the University of Pittsburgh School of Medicine implanted tiny probes in the area of the animals' brains where voluntary movement originates as electrical impulses. Specially designed computer software then transmits these impulses to the robotic arm, which carries out the actions the monkey intended to perform with its own limb.
Using this "brain-machine interface,'' the monkeys are able to direct the robotic arm and open and close a two-finger gripper to feed themselves marshmallows and chunks of fruit while their own arms are gently restrained in tube-like devices.
The technological advance lays the groundwork for development of prosthetics for people with spinal cord injuries and physically "locked-in'' diseases like Lou Gehrig's disease (amyotrophic lateral sclerosis), say the researchers, whose work was published online Wednesday in the journal Nature.
"Our immediate goal is to make a prosthetic device for people with total paralysis,'' said senior author Andrew Schwartz, a professor of neurobiology. "Ultimately, our goal is to better understand brain complexity.''
"Now we are beginning to understand how the brain works using brain-machine interface technology,'' Schwartz said in a statement. "The more we understand about the brain, the better we'll be able to treat a wide range of brain disorders, everything from Parkinson's disease and paralysis to, eventually, Alzheimer's disease and perhaps even mental illness.''
Schwartz's lab had previously focused on brain-machine interfaces for controlling cursor movements on a computer screen, a task the monkeys successfully mastered.
The macaques first learned to manoeuvre the robotic arm to deliver tasty treats into their mouths using a joystick, then moved on to hands-free control using brain signals alone.
"The monkey learns by first observing the movement, which activates his brain cells as if he were doing it,'' Schwartz said. "It's a lot like sports training, where trainers have athletes first imagine that they are performing the movements they desire.''
Co-author Chance Spalding, a graduate student in the department of bioengineering, said the two monkeys -- named Arthur and Pearce -- took only about a week to train.
"It was very quick,'' Spalding said Wednesday from Pittsburgh, noting that macaques and other primates use sticks as tools in the wild and the robotic surrogate arm is just another tool for them to manipulate.
The time it takes for the monkey to see the food and activate the arm is about 150 milliseconds, about the same time it takes a human to decide to move and for a limb to respond, he said.
While the robotic arm looks somewhat industrial and unwieldy, the interface with the monkey's brain is anything but. The probes, or electrodes, inserted into the neuronal pathways in the animal's motor cortex are as fine as a human hair.
The Pittsburgh scientists' latest endeavour isn't the first to employ brain-machine interface technology, but Spalding said it's the first time the technology has been used in this way.
The goal is to help people who have lost physical function through injury or disease.
"When either amputations happen or spinal cord injuries or something occurs, neural capacity in your brain still exists,'' he said. "We believe those signals that normally control your arm are functioning just fine.''
"If there's no longer function or you no longer have an arm, then we can just tap directly into that neural system in the brain and control an artificial device for you to regain function.''
In an accompanying commentary, physiology professor John Kalaska at the Universite de Montreal, said the research takes the notion of thought-provoked movement to a whole new level and "provides a heartening example of what, in due course, may be possible.''
"One encouraging finding was how readily the monkeys learned to control the robot ... Learning could be even quicker in human subjects, facilitated by verbal instructions from a trainer,'' writes Kalaska.
"This also suggests that neuroprosthetic devices could minimize the frustration often encountered by patients in current rehabilitation programs when their diminished motor capacity results in only small performance gains despite prolonged, intense effort.''
However, Kalaska said there are still hurdles to overcome before "neuroprosthetic robots'' are developed for humans and make their debut at rehabilitation clinics.
The long-term reliability of the implantable electrodes must be improved, he said. With current technology, the quality of recorded neural activity often deteriorates within weeks or months.
The devices also must be made more portable: experimental models involve a large array of recording, computer and robotic-control hardware that requires constant attention from a skilled technician, he said.
Spalding agreed but said his lab hopes to begin trials of the brain-controlled robotic arm in people with severe immobility disorders within two years.
"Obviously we have further advances we have to do before this is a device that people will just be able to pick up off the shelf and use in a rehab environment,'' he said. "But it's definitely a step towards that.''
