The first experiments by scientists, using a noninvasive, high-fidelity interface to control a robotic arm, have been successful. In the future, the researchers aim to perfect the technology to make it more widely available.
Robotic arms and other robotic instruments may sound like a futuristic development, but they have been around for years,
Less common, though, are prosthetic, robotic arms that allow people who have lost a limb to regain freedom of movement.
The man can control his robotic arm thanks to a “rerouting” of certain nerve endings, yet so far this prosthetic — developed by scientists from Johns Hopkins University in Baltimore, MD— is not available to other people who may also need it.
Another project — from the University of Chicago in Illinois — has been testing
Now, researchers from Carnegie Mellon University in Pittsburgh, PA, and the University of Minnesota in Minneapolis have managed, for the first time, to use a noninvasive brain-computer interface to control a robotic arm. The scientists report their success in a study paper that appears in the journal Science Robotics.
Prof. Bin He, from Carnegie Mellon, leads the research team that used an interface that does not require a brain implant — which is an invasive procedure — to coordinate the movements of a robotic arm.
Prof. He and colleagues want to develop a high-fidelity, noninvasive method of connecting the brain and flexible prosthetics because inserting brain implants needs not just high surgical skill and precision, but also a lot of money, as implants are costly. Moreover, brain implants come with a number of health risks, including infection.
All of these aspects have contributed to the low number of people receiving robotic prosthetics, so the scientists at Carnegie Mellon and the University of Minnesota have been seeking to turn the tables by developing a noninvasive technology.
Yet there are many challenges in doing this, particularly the fact that previous brain-computer interfaces are unable to decode neural signals from the brain reliably, and so cannot control robotic limbs smoothly, in real time.
“There have been major advances in mind controlled robotic devices using brain implants. It’s excellent science,” notes Prof. He, commenting on previous steps towards finding a “dependable,” technology.
“But noninvasive is the ultimate goal. Advances in neural decoding and the practical utility of noninvasive robotic arm control will have major implications on the eventual development of noninvasive neurorobotics,” he adds.
In their current project, Prof. He and team used specialized sensing and machine learning techniques to “build up” a reliable “connection” between the brain and a robotic arm.
The team’s noninvasive brain-computer interface successfully decoded neural signals, allowing a person, for the first time, to control a robotic arm in real time, instructing it to continuously and smoothly follow the movements of a cursor on a screen.
Prof. He and colleagues showed that their approach — which included a higher amount of user training, as well as an improved neural signal “translation” method — improved brain-computer interface learning by approximately 60%. It also improved the robotic arm’s continuous tracking of the cursor by over 500%.
So far, the researchers have tested their innovative technology with the collaboration of 68 able-bodied participants who took part in up to 10 sessions each. The success of these preliminary trials has rendered the scientists hopeful that they will eventually be able to bring this technology to the individuals who need it.
“Despite technical challenges using noninvasive signals, we are fully committed to bringing this safe and economic technology to people who can benefit from it,” says Prof. He.
“This work represents an important step in noninvasive brain-computer interfaces, a technology, which someday, may become a pervasive assistive technology aiding everyone, like smartphones.”
Prof. Bin He