May 14, 2005
Monkey See, Robotics Do
Primates' brains easily learn to manipulate hi-tech arm
HealthDay News -- Executing some serious monkey business, neurobiologists say the nimble brains of monkeys quickly learn to absorb an external robotic arm into their sense of self -- controlling a wired mechanical limb as if it was the real thing.
"We are quite optimistic," said study author Mikhail Lebedev, a senior research scientist in the department of neurobiology at Duke University. "The monkeys adapted quite well. They were moving the robotic arm as if they were moving a part of their body."
Scientists hope that such sophisticated artificial arms, hands and legs may one day offer a broad range of improved mobility and function to handicapped and paralyzed individuals.
By taking a more in-depth look at data they collected in an earlier study, the researchers found that in the process of figuring out how to manipulate the artificial arm, some of the monkeys' brain cells appeared to reorient themselves.
In essence, brain cells linked to arm movement incorporated this new demand, directing their attention away from the monkeys' flesh-and-blood arms and towards controlling the "fake arms."
That didn't lessen the primates' overall dexterity, however: The researchers say the monkeys were able to operate the robotic appendage while simultaneously retaining full use of their real arms.
Comparing the monkey's initial interaction with the robotic arm to the way a human might interact with an unfamiliar car, Lebedev said there were initial control issues that took some trial and error for the monkey's brain to sort out. And he said it was during this fine-tuning process that certain brain cells -- or neurons -- ultimately came "to represent the movement of the robotic arm more than the movement of the monkey's arm."
In initial studies, Lebedev and his colleagues focused on two female rhesus macaque monkeys, implanting a series of tiny electrodes in targeted areas of their brains. Thin wires smaller than a human hair connected the electrodes in their brains to a control box, which in turn operated a large robotic arm -- about 3 feet tall -- located in a separate room.
As the researchers recorded the monkeys' brain signals, they used fruit juice rewards to motivate the monkeys to engage in 60- to 90-minute training sessions over the course of several days. One monkey participated in nine such sessions, while the other completed 20 sessions.
As first, the primates were shown how to manipulate a joystick that moved a cursor onto a target on a video screen. The cursor was linked to a video representation of the movement of the robotic arm, allowing the monkeys to see the effect the joystick had on the arm's manner and speed of movement.
Eventually the joystick was de-linked from the robotic arm, leaving the monkeys in full brain-only control of the robotic arm -- despite the fact that they continued to jostle a nonfunctioning handle. After a few days, later sessions involved the ultimate removal of the joystick itself, and the monkeys were allowed to see the actual robotic arm.
Lebedev and his team found that after a shaky adjustment period, the monkeys were able to deftly control the robotic arm with only their brains -- without access to the joystick, and without mimicking the movement that their arms had made when moving the joystick about.
In this latest update on the findings, the researchers pored over brain cell activity data from these experiments. These signaling patterns show that as the monkeys improved their ability to control the robotic arm, a larger portion of their neurons started to "fire" specifically while manipulating the robotic arm.
Finding no reduction in the monkeys' ability to also move their own arms at will, the authors concluded that the monkeys seemed to have actually incorporated the robotic arm into the neural structure of their brains.
The extremely flexible nature of this mind-machine interaction points to the ability of primates to continually expand their sense of identity, leaving open the possibility that human brains can do the same with high-tech tools.
Lebedev said that finding is a critically important step in the drive to make prosthetic devices that the mind can accept as a natural, dexterous extension of the body.
"I would say that, or course, maybe right now this looks like something that may not be immediately available to humans," said Lebedev. "But if maybe five years ago one would have asked about this kind of work, research scientists would have been very skeptical, that this is too difficult. So, I wouldn't be surprised if in maybe another five years we would have real implications for humans."
"This basically confirms what we've been saying," said Andrew Schwartz, professor of neurobiology at the University of Pittsburgh School of Medicine. "That when the monkey uses a prosthetic device, he is learning. And the learning takes place as a change in the relationship between the monkey and the device."
Schwartz added, "The same thing would go on with humans as well. And humans would probably learn better than monkeys can."