ARTICLE IN BRIEF
Investigators report two advances in a technique called targeted muscle innervation, which have led to the development of an artificial leg that responds fluently to signals from the brain.
A surgical procedure to reposition nerves in the stump of an amputated leg, combined with a bionic prosthesis that reads the electromyographic signals from those nerves, results in an artificial leg that responds fluently to signals from the brain.
Zac Vawter, a software engineer from Yelm, WA, who was the first person to receive the bionic prosthesis, lost his right leg above the knee at age 31. During the amputation surgery two nerve transfers were performed — increasingly common procedures during amputation because they prevent the formation of painful neuromas. (See “More about Nerve Transfers.”)
“The surgery was the first application of this type of surgery to a lower extremity,” said Levi J. Hargrove, PhD, lead author of an article in the Sept. 26 issue of the New England Journal of Medicine (NEJM) describing the case. “We had used it previously for high-level upper limb amputees, and it has been very successful in providing unprecedented control of artificial arms for high-level amputees, so we extended the technique to the lower extremities, and made use of nerve signals to control the bionic leg prosthesis.”
The technique, known as targeted muscle reinnervation (TMR), was developed at Northwestern University and the Rehabilitation Institute of Chicago (RIC) in the early 2000s, and soon adapted to bionic arms. A team led by Todd A. Kuiken, MD, PhD, of the RIC and Northwestern University's Feinberg School of Medicine, performed TMR surgery on five arm amputees who were able to use the resulting electromyogram (EMG) signals to operate a sophisticated arm prosthesis. Their results were described in a 2009 article in the Journal of the American Medical Association.
INNOVATIONS TO THE PROCEDURE
Two innovations have allowed the same procedure to be adapted to a bionic leg, according to Dr. Hargrove, a research assistant professor in physical medicine at the RIC and at the Feinberg School of Medicine.
First, TMR was used for the first time in an amputated leg to provide signals that could be picked up by a bionic leg.
“The result is the reinnervated muscle amplifies the nerve signal by about 1,000 times, making it easy to measure the EMG signal and decode what the person is intending to do,” Dr. Hargrove said.
The surgery on Vawter involved the common peroneal nerve and the tibial nerve, which innervate muscles below the knee. Those nerves were redirected to a small portion of the hamstring muscle, enabling the spontaneous activation of the bionic leg when Vawter wants to walk, sit, or climb stairs. (Last November, Vawter climbed 103 flights of stairs in Chicago's Willis Tower, formerly known as the Sears Tower.)
The second innovation involves the leg itself, which incorporates an on-board computer designed to pick up nerve signals from the wearer's brain, and initiate appropriate leg movements. A gyroscope, for example, similar to the kind that enables cell phones to switch from landscape to portrait mode, monitors the orientation of the leg. An accelerometer monitors the leg's forward movement. A load cell measures how much weight is being placed on the leg. The sensors feed the signals into the computer, which deciphers them and instructs the leg to perform the movements intended by the person wearing it.
“The nerve in the hamstring muscle naturally fires, and the signal is decoded by the pattern recognition algorithms programmed into the leg,” Dr. Hargrove said. “We teach the computer to recognize how the patient moves. We tell patients, remember how you walked and climbed stairs and moved your leg before, and do the same thing. We teach the computer what to do based on signals from the patient's brain.”
The bionic leg also provides much greater power, according to Dr. Hargrove, thanks to input from collaborators at Vanderbilt University, and at Freedom Innovations, a prosthetics company based in Irvine, CA, which is working to commercialize the leg, perhaps within the next two years. Other advances in the bionic leg have come about as a result of innovations from related industries — computer chips used by the cell phone industry, for example, and compact batteries developed by the auto industry for electric cars. The development of the leg was accomplished in four years with the help of an $8 million grant to RIC from the US Army's Telemedicine and Advanced Technology Research Center (TATRC).
The leg is too loud and too heavy, in Dr. Hargrove's opinion, and efforts are under way to make it quieter, lighter, and more robust. Also, he hopes to see improvements in the steering and control systems.
“The device is an amazing piece of engineering,” he said, “but any errors in the system could cause the person using it to fall. The error rate has been reduced from about 13 percent to 2 percent, and we're working to make the error rate even lower. We believe improving electrode contact with the skin will help to reduce errors.”
The leg can be fooled. The most common confusion involves its failure to distinguish a small incline from level ground. “The person might notice that the ankle doesn't push off quite as vigorously going up a slope,” Dr. Hargrove said. “The leg might feel sluggish, but it won't cause a fall.”
THE RISK OF FALLS
Falls could be an obstacle to the widespread adoption of such prosthetics, said Karunesh Ganguly, MD, PhD, who develops brain-machine interfaces.“With prosthetics in the real world, you have almost zero tolerance for failure,” said Dr. Ganguly, a neurologist in the Neurorehabilitation Program at the University of California, San Francisco. “Imagine what just a 1–2 percent failure rate could mean for something as simple as crossing the street. It could be devastating. They have made important strides, but improving reliability is an obvious next step.”
Still, the NEJM paper represents “an important proof of principle that targeted muscle reinnervation allows intuitive control over lower-limb prosthetic devices,” Dr. Ganguly said. “Through TMR, they can direct nerves in the stump after amputation and allow discrete control.”
The bionic leg and the surgery needed to enable it represent an enormous advance in lower extremity prosthetics, said Gerald E. Loeb, MD, professor of biomedical engineering, and director of the Medical Device Development Facility at the University of Southern California.
“Until about 25 years ago, such prostheses were pretty crude affairs — not much better than the peg leg of pirates,” said Dr. Loeb, who wrote an editorial about the bionic leg in the same issue of the NEJM. “Then people started using sensors and microcontrollers to actively control the locking and unlocking of joints, particularly the knee, at different points in their gait. When you're standing, for example, you want the knee to be locked in a weight-supporting position so the leg doesn't buckle, but when you're walking you need the knee to be fairly mobile.”
Sensors could barely pick up signals from the brain through the skin of an amputated leg, but repositioning nerves in a way that amplified those signals made them usable.
“By transferring nerves into muscles that remain in the proximal part of the limb, you can record signals the patient would have sent to the muscles that no longer exist, and you can replace those muscles with a sophisticated limb containing sensors, motors, and variable impedance joints,” Dr. Loeb said.
Finding a way to pick up and interpret signals from the brain for use by a bionic leg represents the major innovation in this work, said Dr. Loeb. “That's really the obstacle they've overcome in this paper,” he said. “They've included not just the EMG signals, but a sophisticated pattern recognition algorithm that lets them figure out how to combine all the command signals into intelligent decision making about how the prosthesis should behave,” he said.
Dr. Loeb suspects the biggest obstacle to adoption of this system may be the high quality of existing bionic legs.
“What we have here is a very sophisticated system that produces some level of improved performance in a field where the basic level of performance of prosthetic limbs is already quite high,” he said. “Will the added benefit persuade clinicians and insurance companies to go with this much more complex and sophisticated surgical procedure, rehabilitation process, and prosthetic technology?”
One additional innovation, however, may soon make the bionic leg clearly superior — using implanted transmitters similar to the chips implanted in pets that will send nerve signals from the patient's residual leg to the bionic leg, with no need for sensors that make contact with the skin.
“Myoelectric systems are now compromised by the quality of the signal you can pick up on surface of skin,” Dr. Loeb said. “Telemetry for EMG would be a huge improvement because EMG signals recorded from within the muscles and sent out wirelessly wouldn't be so encumbering.”
MORE ABOUT NERVE TRANSFERS
The widespread adoption of the bionic leg depends heavily on surgeons learning to perform targeted muscle reinnervation (TMR), which ideally should be done at the time of amputation. Learning such a specialized procedure just to prepare a bionic prosthesis may not be a priority for many surgeons, but the procedure also eliminates neuroma pain — a common side-effect of amputation — which could encourage surgeons to add TMR to their repertoire of skills.
“What we feared when we started doing this surgery was that we might create painful neuromas due to the mismatch between the large major mixed nerve of the amputation and the small motor nerve it was going to,” said Gregory A. Dumanian, MD, professor and chief of the Division of Plastic Surgery at Northwestern University's Feinberg School of Medicine.
“What we found unexpectedly was the opposite — that people's pain was getting better,” said Dr. Dumanian, who developed the technique with his colleague, Todd Kuiken, MD, PhD, professor and associate dean of academic affairs at the Feinberg School. “We were not creating neuromas in situ; we were helping chronic localized pain from end-neuromas.”
Dr. Dumanian and his colleagues tested this observation on a rabbit amputee model and reported in a 2012 paper in the Journal of Hand Surgery that targeted innervation surgery did indeed calm neuromas.
“It's been the supposition since primate studies conducted in 1985 that muscle is a more favorable place for a nerve to live,” he said. “Our paper is the first evidence of a cured neuroma, which means a nerve brought back toward its previous anatomy.”
TMR cures the neuroma by bringing the number of fascicles back down to normal, and the size of the fascicles back up towards normal, Dr. Dumanian explained. “Giving the nerve somewhere to go and something to do calms down the nerve. Every other neuroma treatment — and there are hundreds — tries to hide the neuroma, or hinder the neuroma from forming.”
As a result of this discovery, the number of people who are candidates for TMR has gone up 15-fold, Dr. Dumanian said. “You don't need the goal of prosthetic control to undergo this type of nerve transfer surgery.”
Also, he and his colleagues have received a grant to do a multicenter randomized trial involving 200 amputees who will be evaluated for chronic pain in their residual limb, and then receive either surgery to bury the nerve ending in muscle, or a nerve transfer.
If the study results suggest that TMR should be performed routinely to prevent neuromas, it could make virtually all above-the-knee amputees candidates for the type of bionic prosthetic developed at the Rehabilitation Institute of Chicago.
“Currently, targeted innervation is reserved for prosthetic control,” Dr. Dumanian said, “but if you're going to do it for chronic localized pain, this could be a surgery that could be beneficial for millions of amputees, or anyone with a painful end-neuroma.”