Bionic limbs learn to open beer

Andrew Rubin is sitting with the Surface tablet, watching a skeletal hand on the screen squeeze and unclench the fingers. Ruby had his right hand amputated a year ago, but he repeats these movements with the help of a special device attached to his shoulder.

The electrodes on his arm are connected to a box that records sequences of nerve signals, which allows Rubin to train the prosthesis to act like a real hand. “When I think of squeezing the fingers, it causes a contraction of certain muscles in the forearm,” he says. “The program recognizes the sequences that occur when I bend or pull a hand that I don’t have.”

A 49-year-old college professor from Washington, DC, several times a month goes to a startup Infinite Biomedical Technologies, using deep learning algorithms to recognize signals in his shoulder that correspond to different hand movements.

Each year, more than 150,000 people undergo hand amputations as a result of accidents or for various medical reasons. Then, most of them get a prosthesis that can recognize a limited number of signals for hand or foot control.

However, Infinite and another company decided to take advantage of improved signal processing, sequence recognition software and other advanced engineering technologies to create new prosthetic controllers that might be able to make life easier for Rubin and other people. The key is to increase the amount of data a denture can accept and help it interpret it. “The goal of most patients is to use more than a couple of functions, such as squeezing and unclamping fingers or turning the hand. Sequence recognition gives us this opportunity, ”says Raul Kaliki, director of Infinite. “Now we are able to intercept more active signals in the limb.”

The Kalika team, consisting of 14 employees, creates this electronics, which then goes into prostheses manufactured by other companies. Their electronic controller Sense records data from several, up to eight, electrodes on the Rubin's shoulder. Thanks to many hours of training with a tablet app, this device is able to recognize intentions encoded in Rubin's nerve signals when it moves its shoulder in a certain way. Then Sense sends instructions to the prosthesis that forms a certain grip.

Last Friday, Kaliniki from Infinite received a notification from representatives of the Food and Drug Administration (FDA) of the United States approving the sale of Sense to the United States. Kaliki says he expects to start installing the system in prostheses by the end of November. In 2017, FDA officials approved a similar system from the Chicago-based company Coapt. Today, more than 400 people use it at home, according to the company's director, Blair Locke.


Coapt

Locke began working as an engineer 13 years ago at the Chicago Rehabilitation Institute, associated with Northwestern University. He worked with surgeons who repaired nerve damage in patients after amputation. One day he realized that it would be easier to create an improved prosthesis if he could figure out how to improve the reception of signals from the body, he says. “The innovation is that we provide a more natural, more intuitive way to control with bioelectronic signals,” says Locke.

In earlier versions of prostheses, electrodes recorded the strength of the signals, “but it was as if you would know only how loud the instruments played while listening to the orchestra,” Lock said. “It was quite difficult to understand the content and accuracy of the signals.” The Coapt system works inside the prosthesis, and costs from $ 10,000 to $ 15,000, depending on the number of required personal adjustments. Artificial limbs, Lok says, could cost a person between $ 10,000 and $ 150,000.

Nicole Kelly received a new prosthesis along with a control system from Coapt a year ago. Now a 28-year-old Chicago resident is able to grind fresh peppers and hold playing cards, as well as open a beer.

“For many actions, it turned out that I wasn’t that I couldn’t do them at all before - they just suddenly started giving me a lot easier,” says Kelly, born without a left forearm. Her prosthesis is “not my body, and not 100% natural,” she said. - My body communicating with this technology has a learning curve. Even in such things as keeping the salt shaker and pepper shakers, I’ll, in fact, start for the first time. ”

In the Coapt system, there is also a reset button, which allows Kelly to restart the sequence recognition system if the hooks do not work as it should. “If at some point it seems to me that she is doing something strange, I can click on the reset,” says Kelly, a former participant in the Miss America contest, and now she’s a disabled rights activist. She says that now it takes about two minutes to re-train her arm.

And this is not the only innovation. Infinite Biomedical engineers distribute RFID tags ( RFID)), so that people without limbs can put them on door handles, kitchen appliances and other household objects - useful devices that require certain grips. The idea is that the controller in the prosthesis recognizes the RFID signal, and automatically changes the grip, with, say, one that is required to rotate the door handle, to the one needed to take the newspaper. According to Kaliki, the project is being developed with financial support from the National Institutes of Health.

These technologies are still new and unavailable for everyone. It takes a lot of training to learn how to use them, and, of course, not all insurance companies will pay for the most complex prostheses or new control systems. However, patients such as Andrew Rubin hope that many of these breakthroughs will appear soon enough. So far, if he needs to take a cup, and then open the door, he has to use the application for the smartphone every time he needs to change the grip on the prosthesis.

“This is a slow process, and I think that we will eventually come up with something that will allow me not to rely on the phone to change the grip,” he says. Rubin says he likes to train at the Infinite office in Baltimore every week, as well as at the Johns Hopkins University bioengineering laboratory, which develops a glove that can feel pain like a real hand. But Rubin - suffering from an infection with sepsis, and having survived a leg amputation a few years ago - I would like to go to the point where he could use his right hand to release the bolt on his mirror, balance the plate or even write with a pen. And he, as the first person to experience at home a new sequence recognition system from Infinite, is not too far from this point.