Scientists have been working with prosthetic robot arms as a way to help people who have lost limbs to regain their freedom of movement. However, these robot arms require special brain implants for them to work. Now, a group of researchers developed a new kind of prosthetic robot arm that no longer needs these implants, and the surgeries needed to implant them.

For the first time ever, researchers from the Carnegie Mellon University, in collaboration with the University of Minnesota, have demonstrated the use of a non-invasive brain-computer interface (BCI) to control a robot arm. Using this non-invasive BCI, the researchers have demonstrated the possibility of making this novel technology more accessible to a wider range of patients, by no longer requiring surgeries to implant BCIs.

Moving robot arms with the mind

While robot prosthetics have been around for some time now, these all required BCIs implanted into a patient’s brain. However, inserting implants into a person’s brain not only requires utmost surgical skill and precision, it also costs a lot of money. This is in addition to the health risks that such a surgery exposes a patient to, such as infections and even brain damage. As such, developing a non-invasive BCI to control these prosthetics has been one of the biggest challenges of the technology.

“There have been major advances in mind controlled robotic devices using brain implants. It’s excellent science,” said team lead Bin He, professor at Carnegie Mellon. “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 and his team have been looking at ways to develop a high-fidelity, non-invasive method of connecting the brain to robot prosthetics – but doing so hasn’t been easy. Previous brain-computer interfaces have not been able to decode neural signals reliably, preventing them from controlling robot limbs smoothly in real time. To overcome this challenge, the team turned to another new technology that’s being used to improve robotics in other fields — machine learning. (Related: Engineers develop “Vegebot” that can harvest lettuce.)

Using machine learning to connect man and machine

To create a BCI that could read brain waves quickly and control a robot limb in real time, He and his team used machine learning techniques to build a reliable “connection” between the brain and the robot arm. Using this, the team were able to allow a person, for the first time, to control a robot arm in real time, demonstrated by having the arm follow a moving cursor on a screen.

While this approach required a higher amount of user training, the researchers demonstrated that it improved BCI learning by approximately 60 percent. The new approach performed 500 percent better than previous non-invasive BCI systems in the same cursor-following test.

So far, the team has tested the new BCI technology on 68 able-bodied participants who took part on up to 10 sessions each. With the success of the preliminary trials, the team is now hoping to conduct clinical trials with actual patients in the future. Their hope is to eventually be able to make the technology pervasive and economical for the people 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,” stated 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.”

Sources include:

MedicalNewsToday.com

Engineering.CMU.edu