American Way

Robotic Prosthetics

For the last decade, Jan Scheuermann has been unable to move any of the muscles below her neck. Yet she has been working out like an athlete. (2013)


By Charlotte Huff

FOR THE LAST DECADE, Jan Scheuermann has been unable to move any of the muscles below her neck. Yet she’s been working out like an athlete since early 2012, exploring the outer boundaries of mind control and robotics.

Sitting in a nondescript building on the University of Pittsburgh campus, ­Scheuermann works with researchers and a robotic arm she calls Hector to perform feats that sound more like science fiction than fact. With the help of two tiny electrode arrays implanted in her brain and relentless work — eight to 12 hours weekly since February 2012 — Scheuermann can mentally direct where the arm moves next. As she focuses on a particular movement, her brain’s nerve impulses transmit to the computer that manipulates the robotic arm.

Because her work is part of an ongoing research project — and one of the most stunning advances in a federally funded effort to revolutionize prosthetic arms and hands — Scheuermann can’t take Hector home with her. But what might she achieve if she could?

“I could open the refrigerator and pull out a plate of food that someone had prepared and left for me,” says the 53-year-old Pittsburgh woman, who has a progressive neurological disease called spinocerebellar degeneration and steers her wheelchair with her chin. “I could feed myself. And I could bring a cup of liquid up to my mouth and drink through a straw.”

Such mind-bending capabilities are but a few of the numerous research avenues emerging from an ambitious effort by the Defense Advanced Research Projects Agency­ (DARPA) — sometimes dubbed the mad scientists of the U.S. military — and its Revolutionizing Prosthetics program to significantly improve the options for veterans with upper-limb amputations.

By late January of this year, there had been 1,585 amputations involving troops who’d served in Afghanistan or Iraq. Of those, nearly one in five involved a portion of the arm or hand, according to Department of Defense statistics. Unfortunately, existing arm and hand prostheses are typically cruder and less functional than those available for legs, says Dr. Geoffrey Ling, deputy director for DARPA’s Defense Sciences Office and program manager for Revolutionizing Prosthetics. That’s because of nature’s ­engineering; there are far more joint movements in the wrist and hand to replicate than in the leg. But Ling, a neuro-critical-care physician who retired as an Army colonel in 2012 after combat tours in Afghanistan and Iraq, is undaunted. “At DARPA, we don’t believe in anything that’s insurmountable,” he explains, with a touch of braggadocio.

Launched in 2006, Revolutionizing Prosthetics has been funded to the tune of $144 million to date and involves numerous medical centers and researchers. (DARPA estimated having more than 300 researchers in 2010, but officials didn’t venture a more recent figure.) And though it’s only been seven years, amputee patients, veterans and civilians alike are starting to benefit from the program’s advances in treatment.

For example, some amputees are taking advantage of a surgical technique called targeted muscle reinnervation that allows them to use remaining nerves to exert better control over a prosthesis. During the surgical procedure, the nerves are transferred into muscle tissue, enabling more natural movement. And by the time you read this, perhaps federal officials will have approved the commercial sale of another key Revolutionizing Prosthetics effort — the creation of a sophisticated motorized arm for mobile amputee patients. Called the DEKA Gen-3 Arm System, it’s considered a substantial improvement over more traditional, griplike prosthetic devices.

Dr. Justin Greisberg, associate professor of clinical orthopedic surgery at Columbia University Medical Center in New York, has been following these advances and tells patients that the evolving technology’s potential is enormous in terms of improving quality of life.

“It’s going to take some time and some work,” he says. “But it’s not if we’re getting a much more functional robotic arm — it’s just a matter of when.”


STRETCHING FROM THE SHOULDER to the fingertips, the human arm easily executes a daunting spectrum of bioengineering capabilities each day — lifting a gallon of milk, eating grapes, grasping and turning a doorknob. The arm is capable of 29 degrees of motion. (Actually, more: A simple joint movement, such as flexing the elbow open and closed, is considered one degree.)

While prosthetic arms and hands are improving, until recent years, they’ve been largely limited to griplike hands that only open and close, says Linda Resnik, a research health scientist at the Providence Veterans Affairs Medical Center in Rhode Island. Sometimes people will give up entirely, particularly if the prosthesis replaces a large portion of an arm. “They may be uncomfortable to wear,” she says, “and they don’t feel like they’re really that beneficial in helping them do more.”

Thus, DARPA’s program has been developing along two research tracks, one of which involves the more pressing goal of developing a significantly improved prosthetic — a “strap-and-go arm” that doesn’t require surgery. DARPA has achieved success with the DEKA Gen-3 Arm System, developed in collaboration with Manchester, N.H.-based DEKA Integrated Solutions. Last year, the company­ submitted an application to federal officials to sell the prosthetic commercially.

Resnik has been working with the robotic DEKA arm as the principal investigator for two related Veterans Affairs (VA) studies. Nicknamed “Luke” after Luke Skywalker, it comes in several different models, depending upon how much of a patient’s arm has been amputated. The full arm can achieve 10 degrees of freedom; users can operate it through various modes. One common approach is by using foot controls, in which a motion-sensor device (roughly the size of a matchbox) is inserted into the laces or otherwise is attached to the shoe.

The technology, which is tailored to the user’s preferences, then sends information about movement and speed wirelessly to move the Gen-3 arm. For example, it might be programmed so when the user lifts his toes on one foot, the arm will bend. And, when he lifts his heel, the same arm will straighten.

One woman in the VA study was able to return to her sewing passion, stitching together an apron. “A lot of our subjects want to cook and are very happy to be able to have two arms to use in cooking tasks,” Resnik says.

Last year, VA researchers launched a home-based study to track how the DEKA arm functions outside of a research setting. If the arm is commercially approved, Veterans Affairs and Department of Defense officials “are ready to go and acquire these for the veterans who qualify for them,” Ling says. As of this spring, no purchasing decision had been made — but roughly 300 veterans have expressed a strong interest in becoming the first users.


THE MOON SHOT, THOUGH — and the second research avenue that involves Scheuermann — focuses on integrating an artificial arm with the body’s own neuro-circuitry. “Clearly, we had to tie this directly into the brain so that the patient could just think about it and make it work,” Ling says. Led by Johns Hopkins University’s Applied Physics Laboratory (APL), the idea is to enable more natural prosthetic movement by accessing the brain’s own nerve pathways. The concept, which Ling describes as “pretty Space Age–thinking,” is not as farfetched as it might initially sound. When you close your eyes and move your arm up and down, you still sense where it’s located, says Bob Armiger, a biomedical engineer at the Johns Hopkins lab. “When you lose an arm, you don’t lose that sense of the limb itself,” he says, explaining that the related brain region likely remains intact.

Scheuermann knows this well. A Jeopardy! fan, she keeps track of her own score while watching the quiz show by mentally pushing down various fingers — immobile in reality — to keep count. “In my mind, I’ve always had control of my fingers,” she says.

In one of the most intriguing nerve-related breakthroughs, one that’s already assisted some mobile amputees, surgeons have moved nerve endings to help power the prosthetic arm. Armiger describes how a procedure at Johns Hopkins helped a West Virginia man who had lost his lower arm and part of his upper arm to cancer. In that man’s case, key nerves were transferred just above the amputation to the muscles that controlled the biceps and triceps, Armiger says.

Then, by using electrodes situated over those muscles, the nerve signals were amplified and sent to a small, computerized device at the socket of the amputated arm. “It’s the computer that turns muscle signals into movement,” Armiger explains. The surgical technique, first pioneered by a Chicago physician, even can assist veterans using simpler prostheses, such as the gripper devices. “It means you can just naturally open and close that gripper in the same way that you think about opening and closing your hand,” he says.

According to Columbia University’s Greisberg, as more advances are achieved, it’s possible that those nerve signals could be more directly captured. In one scenario, he says, a cufflike device might be located near the amputation site to read the nerve signals and transmit that information to and from the robotic arm. “That’s not where the research is now, but that’s certainly not crazy to think about,” he says.


THE EVENTUAL GOAL is to pair these sorts of nerve connections with the more sophisticated arm the researchers at the Johns Hopkins lab have built. That arm has potential for 26 degrees of freedom, including notable finger dexterity. Scheuermann’s early 2012 surgery, during which the two electrode arrays were implanted, opens the door for this robotic-limb technology to be used one day by individuals with quadriplegia related to traumatic injury or diseases such as muscular dystrophy.

During the four-hour procedure, the arrays were situated near the brain regions ­responsible for hand and shoulder movement, says Dr. Elizabeth Tyler-Kabara, Scheuermann’s neurosurgeon at the University of Pittsburgh School of Medicine. (Prior MRI scans had pinpointed both the anatomy of Scheuermann’s brain as well as which areas were active when she imagined specific movements.) The two square arrays, each no larger than a pencil eraser, contain 96 functional minielectrodes, which penetrate slightly into the motor cortex, according to Tyler-Kabara.

“We are recording the responses from single neurons,” she explains. “And so in order to pick up the responses from single neurons, you actually need the electrodes to be next to a neuron.”

Following the surgery, Scheuermann recalls being connected to the computer for the first time and a researcher asking her to imagine moving a finger. “And I did,” she says. “And the neurons started popping.”

Those first neuron firings were the start of a long process in which Scheuermann and the team have been training the robotic arm to move, creating a feedback loop between her nerve signals and the attached computer­ that moves the robotic arm. These days, Scheuermann can pinch the robotic fingers together and scoop up items as small as a one-inch cube.

Even now, her achievements give Ling tingles. “I’ve had eminent, eminent scientists who will remain unnamed who told me that it was not possible within 25 years,” he says.

The long-term hope is that endurance training by Scheuermann, and eventually other patients, will help design a more automated system that doesn’t have to be tailored to each user, Tyler-Kabara says. “They need to be able to wake up in the morning and it’s working,” she says.

Other hurdles also will have to be cleared before the Johns Hopkins arm will be available outside of a research-lab setting, Ling says. Another goal: to make the brain technology wireless. At this point, the computer that controls the robotic arm is wired to the two port terminals sitting roughly along where Scheuermann’s hair parts. “You look like a Frankenstein creation,” Scheuermann quips.

The APL arm also has sensory capabilities, with 74 sensors in the hand itself, including for vibration and temperature, according to Ling. The next step: to see if two-way sensory feedback can be achieved between the robotic arm and the brain.

As of this spring, researchers were in the process of recruiting another individual with quadriplegia to build upon what Scheuermann has achieved, Tyler-Kabara says. For that surgery, electrode arrays will be implanted in areas of the brain that process sensation, as well as movement.

But all of this innovation doesn’t come cheap. “The strap-and-go arm right now is pushing about $100,000,” Ling says, adding that as more are manufactured, that cost likely will decline due to economy of scale. “This is as high as [the price] is going to be.”


SCHEUERMANN, WHO HAS TWO grown children and relies on a caregiver while her husband works, also foresees numerous practical uses. With a mind-controlled arm, she could turn pages rather than rely on audio books. She could push a button to activate a phone and make a call, including to 911, a capability currently out of her reach.

The Pittsburgh woman is prone to naming items — she calls her wheelchair Sven because it has a Swedish manufacturer. She selected the name Hector for the APL arm because it has a hefty masculine look to it, she explains. It was only later that she learned one of the name’s meanings in Greek: “to grasp.”