Monkey's Brain Runs Robotic Arm


By Michael Schirber
LiveScience Staff Writer
posted: 18 February 2005                                                                  


WASHINGTON D.C. - Robotic arms used by amputees are typically controlled by moving some other part of the body, like the opposite arm. Researchers would like to make such prostheses respond to the whim of the brain.  

Now it turns out researchers have found a method so easy (well, relatively so) that a monkey can do it.  

In a new study, a monkey fed itself using a robotic arm electronically linked to its brain. The work was presented here Thursday at the annual meeting of the American Association of the Advancement of Science (AAAS).  

The robotic arm is about the size of a child’s, with a fully functional shoulder and elbow, as well as a simple gripper that can hold a piece of fruit or vegetable.

"It moves much like your own arm would move," said Andrew Schwartz from the University of Pittsburgh.  

The monkey’s real arms are restrained in plastic tubes. To control the robotic arm, 96 electrodes – each thinner than a human hair – are attached to the monkey’s motor cortex, a region of the brain responsible for voluntary movement. Although there is an area of the cortex generally associated with arm motion, the exact placement of the electrodes is not crucial, Schwartz explained.    "You don’t have to be exactly right because the brain is highly plastic," he said, referring to the fact that the brain will rearrange its structure to get things done. And food, it turned out, was a good motivator for the adaptable primate brain.  

About the size of a child’s arm, this robot version moves much like a natural arm, with a fully mobile shoulder, elbow and simple gripper to hold objects. Credit: A. Schwartz

Computing intention

the electrodes measure the firing rate of a single neuron. Each one of the billion or so neurons involved with arm motion is thought to have a preferred direction. There is, for instance, a set of neurons associated with moving the arm up, or down, or to the right.   With a special computer algorithm, the researchers are able to find an average direction from the small sample of neurons being measured. This average direction is used to move the robotic arm.

Rob Kass from Carnegie Mellon University, who was not involved in the new study, said this type of algorithm has been around since 1960. The computational shortcut has had a wide range of applications, including missile tracking and navigation.

"The benefit of the algorithm is that it allows for a more efficient use of data," Kass said. "It also provides a framework for learning."  

This learning was evident in the fact that, with practice, the monkeys became faster with the robotic control. The researchers also found that their subjects could adapt to different placements of the food.

A monkey successfully manipulates a robotic arm with signals from its brain to hold fruit. Credit: A. Schwartz

"Our algorithm is not exactly what is going on in the brain," Schwartz said. But the monkey’s brain actually adapts its neural signal to be closer to the algorithm. The reward for this rewiring is the snack.


Interestingly, in the beginning, the monkey’s restrained arms would twitch – as if they were trying to reach and grab the food. But after a day with the robotic arm, the monkey was completely relaxed.


"He no longer was trying to move his own arms," Schwartz said.


Next up: realism


Schwartz and his collaborators plan to move beyond the simple two-pronged gripper to a more realistic hand with fingers.


"That’s where we want to go next," he said. "We will need to connect electrodes to 50 or 100 more neurons – we think."


The research may one day lead to permanent artificial prostheses for those who have lost a limb, and it might also increase the mobility and dexterity of those suffering from spinal cord injuries or nervous system disorders, like ALS.


"We hope to move to human subjects in two to four years," Schwartz said.


A big hurdle, however, is the fact that biological material builds up around the electrodes, causing the signal to degrade over time. On average, the electrodes in the monkey brains only lasted six months. More bio-compatible materials, as well as devices that transmit their signal without wires, may be needed to make the jump to humans.


Monkey Brain Controls Walking 'Bot

By Noah Shachtman  January 15, 2008

THE PLAYERS Dr. Miguel A. L. Nicolelis, left, at Duke University, and Gordon Cheng in Kyoto, Japan, with the robot.

 When our robotic overlords finally do take over, there's a decent chance they'll do it with monkey brains. 

A few years back, Duke neuroscientists, funded by the Pentagon, figured out how to have monkeys control robotic arms with their little simian minds (above article).  Now, if that wasn't unnerving enough, the same Duke crew has discovered a way for one of the monkeys to make "a 200-pound, 5-foot humanoid robot walk on a treadmill using only her brain activity," the New York Times reports.  How far away are we from the ultimate sci-fi dystopia: Terminator and Planet of the Apes -- at the same time! 

And did we mention that these same researchers recently took out a patent on mind-controlled weapons

In preparing for the experiment, Idoya [the monkey] was trained to walk upright on a treadmill. She held onto a bar with her hands and got treats — raisins and Cheerios — as she walked at different speeds, forward and backward, for 15 minutes a day, 3 days a week, for 2 months.

Meanwhile, electrodes implanted in the so-called leg area of Idoya’s brain recorded the activity of 250 to 300 neurons that fired while she walked. Some neurons became active when her ankle, knee and hip joints moved. Others responded when her feet touched the ground. And some fired in anticipation of her movements.

To obtain a detailed model of Idoya’s leg movements, the researchers also painted her ankle, knee and hip joints with fluorescent stage makeup and, using a special high speed camera, captured her movements on video.

The video and brain cell activity were then combined and translated into a format that a computer could read. This format is able to predict with 90 percent accuracy all permutations of Idoya’s leg movements three to four seconds before the movement takes place.

On Thursday, an alert and ready-to-work Idoya stepped onto her treadmill and began walking at a steady pace with electrodes implanted in her brain. Her walking pattern and brain signals were collected, fed into the computer and transmitted over a high-speed Internet link to a robot in Kyoto, Japan.

Military scientists have tried all different combinations of animal and machine -- cyborg pigeons, radio-controlled rats, steerable shark spies. The Duke experiments are the "first steps toward a brain machine interface that might permit paralyzed people to walk by directing devices with their thoughts," according to the Times. Electrodes in the person’s brain would send signals to a device worn on the hip, like a cell phone or pager, that would relay those signals to a pair of braces, a kind of external skeleton, worn on the legs."

“When that person thinks about walking,” Duke's Dr. Miguel Nicolelis says, “walking happens.”

No lie.  And all kinds of researchers and military departments are working on similar efforts, to move towards thought-controlled exoskeletons and prosthetics.  Notice, too, the robot in this experiment was built by exoskeleton-maker Sarcos

But that still doesn't mean we shouldn't panic.