Scientists have invented a device that allows paralyzed patients to control a robot arm directly from their brain, bypassing their damaged central nervous system.
This brainwave-connected device would then be able to grab objects that the user wants. The system uses tiny electrodes implanted directly into the primary motor cortex, the part of the brain that controls movement. Signals are then routed through a tiny box in the scalp, which is then connected by wire to a small refrigerator-size computer. The computer then translates the brain movement patterns into an algorithm that can be transmitted directly to the robot arm.
PHOTOS: Real Life Robotics Better Than ‘Transformers’
“The ultimate goal is to develop neural technologies to restore mobility specifically for people with no control of their arms or hands,” said Leigh Hochberg, a neurologist at the Department of Veterans Affairs in Providence, R.I., who also has appointments at Massachusetts General Hospital, Brown University and Harvard University. “We’re hoping to provide technology directly from brain signals back to commands that control assisted devices or limbs.”
WATCH VIDEO: Cutting-edge robots, recently unveiled by NASA and General Motors
Hochberg and John Donoghue, director of the Brain Institute at Brown, had previously collaborated on the “BrainGate” project that produced the 2006 study showing how a patient could control a computer cursor using a brain-to-computer neural interface. Their latest study, which appears in today’s issue of the British science journal Nature, goes a step further. It is, in effect, developing a separate neural pathway to deliver messages from the brain to an arm, in this case an artificial one.
Hochberg said the experiment has worked with two patients, a man and a woman who both lost the use of their limbs and their voice as a result of a stroke.
SEE VIDEO HERE:
Both patients were able to move the robot arm to grab foam balls. And the woman was able to pick up a metal coffee container and drink through a straw for the first time since she was injured 15 years earlier.
“The smile on her face was something I and our research team will never forget,” Hochberg said.
The researchers cautioned that the efforts with the two patients only worked successfully about two-thirds of the time, and are not as fast or accurate as a human arm. The experiment does give hope to millions of patients who have become paralyzed as the result of stroke or other physical trauma.
Researchers implanted an array of electrodes the size of a baby aspirin near the top of the motor cortex. From their, 96 hair-thin electrodes pick up signals and send them to a penny-sized device in the top of the scalp. A matchbox-sized transmitter on top of the head then relays the brain signals to the computer which controls the robot arm.
Each neuron is like a radio broadcast tower putting out signals that generates a pattern, Donohue said.
“It’s like a QR code coming out at all times. The computer takes that pattern and translates it into a command that moves it to the left or right."
NEWS: Robotic Arm Inspired by Elephant Trunk
The next step is to rig the BrainGate system to a prosthetic limbs that a patient could wear. And after that, perhaps to the muscles in the paralyzed limbs themselves.The researchers compared the development of their project to the years of engineering and neuroscience used to develop cardiac pacemakers and deep brain stimulators which are now used to help Parkinson’s disease patients. At first these devices were experimental and expensive, but are now common and affordable.
“There’s no doubt that for this device to be successful, it has to reach people who would benefit and it will have to be affordable,” Hochberg said.
“Affordable means that it’s in a range that can be acquired privately or reimbursable by insurance.”
The BrainGate project is also leading to new understanding about the capabilities of robots, according to Patrick van der Smagt, a director of biomimetic robotics and machine learning labs at the German Aerospace Center and the Technical University of Munich, and a co-author on the Nature paper.
“We start to understand the properties of a muscle,” van der Smagt said, “and we want to use that understanding to build future robots more efficiently and more safely.”