DARPA’S RP Arms Progress, Nurture Related Wonders

Home > Articles > DARPA’S RP Arms Progress, Nurture Related Wonders
By Judith Philipps Otto

Photographs of the MPL courtesy of the Johns Hopkins University Applied Physics Lab.

Magic begets magic. Creation engenders further creation. As The O&P EDGE continues to follow the development of the two futuristic prosthetic arms conceived in 2006 by the Defense Advanced Research Projects Agency's (DARPA) Revolutionizing Prosthetics (RP) program, the news is not only about the progress of the two original designs, but also the possibilities for exploration and the related discoveries that have been made along the way-some of which already have or are on the verge of changing lives.

The "Luke Skywalker" arm, developed by DEKA Research and Development Corporation (DEKA), Manchester, New Hampshire, is on track to become commercially available soon, and the Modular Prosthetic Limb (MPL), which was created and unveiled at the Johns Hopkins University Applied Physics Lab (JHU APL), Laurel, Maryland, in December 2010, is moving along an equally exciting trajectory, albeit one without near-term commercial applications.

According to Geoffrey Ling, MD, PhD, DARPA program manager, "The focus of the Johns Hopkins University Applied Physics Lab's effort has shifted...from an emphasis on attaining commercialization to continued research on the brain-machine interface. The reason for this shift is that the Gen 3 Arm System being developed by DEKA will fill the immediate need for an advanced prosthesis, while the APL effort fills the need to further explore how to attain near-natural control of a prosthesis with sensory feedback provided directly to the brain.

"The challenge with commercializing the APL system is that the control system must go through approval at the same time as the arm system, but the enabling technology for the braincontrol system has to evolve before it can become commercially viable."

Mike McLoughlin, APL deputy business area executive and JHU Whiting School of Engineering instructor, cites the MPL team's experience with several trial patients, including spinal cord injury patient Tim Hemmes, whose success at controlling the prosthetic arm using an implanted electocortographical (ECoG) array made history in October 2011. Within 30 days of receiving the brain-computer interface (BCI) implant, Hemmes learned to control the robotic limb through thought alone and reached out to touch his girlfriend. (Author's note: View the video at www.youtube.com/watch?v=yff20TlHv34) The effort earned the University of Pittsburgh (Pitt), Pennsylvania, team that built the BCI and Hemmes a 2012 Popular Mechanics Breakthrough Award.

This success paved the way for a longer-term test with a patient using a different type of electrode array implant. Mike Boninger, MD, director of the University of Pittsburgh Medical Center (UPMC) Rehabilitation Institute and chair of the Pitt Department of Physical Medicine and Rehabilitation, announced in December 2012 that unprecedented levels of thought control of the prosthetic arm had been achieved by a subject named Jan Scheuermann.1

"The fact that Tim Hemmes was able to obtain three degrees of freedom control in less than a month was quite spectacular," Boninger says.

Scheuermann, who has quadriplegia, brings a chocolate bar to her mouth using a robot arm she is guiding with her thoughts. Researcher Elke Brown, MD, watches in the background. Photograph courtesy of UPMC.

But when the U.S. Food and Drug Administration (FDA) granted permission for a longer-term test with the second tetraplegic subject in February 2012, Boninger's team placed two tiny arrays that barely penetrate the motor cortex and found that Scheuermann was able to achieve three degrees of freedom in about three days. "She has now reached the point where she has seven degrees of freedom control," Boninger reveals.

Degrees of Freedom

Where one degree of freedom allows movement in a single plane-along a line on a computer screen, for example-two degrees allows two-dimensional movement on the screen, such as mouse movements up and down, as well as from side to side. Three degrees allows movement outside the plane and into three-dimensional space, as demonstrated when Hemmes moved the robotic arm through space to touch his girlfriend.

Six degrees of freedom adds three movements that control orientation of the wrist, Boninger explains: flexion and extension; ulnar and radial deviation; pronation and supination. The seventh degree of freedom is grasp.

"This new subject was able to reach in space, orient her wrist, and grasp objects," he notes. "We did a clinical test called the ARAT [Action Research Arm Test], which looks at how well you can function with one arm, and she came very close to being normal in certain tasks.... This is someone who was previously unable to move at all and can now control a robotic arm to stack cones, to move objects, and shake your hand.

The subject was even able to feed herself chocolate for the first time in a long, long time, which was one of the highlights of the experiment for her."

Since this subject improved with practice, Boninger wonders how Hemmes and the ECoG array might have improved in terms of degrees of freedom if he had been given longer than a month to practice.

"We get very good function with seven degrees of freedom. We think if you get to ten or 12 degrees of freedom, you'd really have a pretty fully functional arm. I don't know if you could play the piano with it-we're not sure-but you could do most of the tasks that need to be done.

"We're continuing to do experiments to get more control, and we're getting this amazing insight into the brain," he marvels.

Cells in the brain fire just by seeing an arm move, he explains. The researchers record which nerve cells fire and what patterns are created when the subject watches the arm move in different ways or directions. Using the identified patterns, they create and train a neural decoder, which enables the subject to control the prosthesis simply by thinking about moving the arm.

Such decoder training has been done in virtual reality and using the MPL. Both systems-and both types of implant arrays-have advantages and disadvantages, which will be explored more thoroughly in future experiments, Boninger says.

Continuing experiments will focus on several goals. "Right now, the subject has to be looking at the arm for it to work," he explains. "We want people to be able to touch something and feel it, so we'd like to be able to incorporate sensation at some point. We would like to be able to have someone control two arms at once, and we would like to be able to leave the implant in forever. And at this point, there are obstacles to that, so we'd also like to have a wireless device."

How soon might a neurally controlled prosthesis be ready for commercial development and use?

Depending on further pursuit of the ECoG part of the experiment, and the time to train to acheive more degrees of freedom, it could be sooner than one might think, Boninger hints.

"With volunteers and with funding, and with some great researchers like the team we have at Pitt, I think that could be something that could be available in five years," he speculates. "I think it's going to be a little bit longer for the more complex technology that we used on Jan, but I think clearly it's the near future-not some very distant phenomenon...."

Test subjects, he stresses, are vital to the program, which is always recruiting. Research volunteers are encouraged to contact the study coordinator at 412.383.1355, or visit www.upmc.com/BCI

Revolutionizing Amputation Surgery

The RP program can count another significant discovery to its credit-one that can make an immediate difference in amputees' potential to succeed even with conventional prostheses, believes Albert Chi, MD. A JHU trauma surgeon and biomedical engineer, Chi has been working with McLoughlin's team to develop innovative surgeries for people with upperlimb amputations to enhance their ability to utilize as many of the capabilities of the MPL arm as possible-as naturally as possible.

Patients with transhumeral amputations typically control their prostheses using muscle activation from the two muscle groups available to them, the biceps and triceps, Chi points out. Controlling the hand using those contractions, i.e., with non-intuitive control, requires conscious flexion and extension of the elbow.

"Most people get very frustrated with the traditional control in state-of-the-art myoelectrics because they call it a huge cognitive burden to perform these types of unnatural movements," he says. "There's a high abandonment rate."

Photograph courtesy of the Johns Hopkins University Applied Physics Lab.

Chi says the solution lies in his unique adaptation of targeted muscle reinnervation (TMR) surgery, pioneered by Todd Kuiken, MD, PhD.

"Imagine you're missing your limb above the elbow. All the nerves from your brain to your spinal cord down to that missing limb are still intact, but they have nowhere to go. Essentially that information is sent down and it's just lost. It goes off into space."

Chi locates the nerve endings that are going nowhere and dissects out the nerve innervations to three separate nerves- the bicep, the brachioradialus, and one of the heads of the triceps. In each case, he preserves one of the heads and disconnects another, then moves up one of the nerves that used to serve the hand. The process creates three extra nerve inputs and a natural amplifier that contracts these muscles when the amputee thinks about moving his or her hand and wrist.

Working with the MPL researchers and using pattern recognition algorithms to identify individual muscles that are contracting, how well they communicate with each other, and their amplitude and frequency, Chi is able to classify up to 12 separate motions for a transhumeral amputee.

"Our patient, with natural thought, can flex his wrist, point, pinch, [and] open and close his hand; he is also able to deviate his wrist toward the thumb and the pinky.... So where before you only had two motions you could control with intuitive thought, now we've gone to 12. It's great!"

About 50 patients worldwide have had the TMR surgery to date, and Chi has access to three for daily training.

Even for users of commercial prostheses who don't choose pattern-recognition training, the surgery offers benefits. "Instead of doing a complex elbow-flexion, elbow-extension effort, now they're doing natural movement with the commercially available arm-including simultaneous multiple motions-like an elbow flex while closing the hand."

As of this writing, one of Chi's patients was scheduled for a landmark fitting in January with a commercially available pattern-recognition arm-one of the first available.

Driven by another exciting discovery, Chi is also urging trauma surgeons to change the way they perform amputations.

"Currently, the standard of care for any kind of amputation is to cut the nerve back as far as you can and tie it off with a suture, and that's so barbaric. Instead of tying off the nerve," he suggests, "tuck it into a little muscle pocket. This local muscle reinnervation procedure is actually easier...and you can actually complete the amputation faster!"

If, after rehabilitation and six months of mental imagery exercises, the patient doesn't succeed, he or she can still get TMR. The patient has lost nothing, Chi notes.

His enthusiasm for the new procedure is fueled by his discovery of a patient with a very high transhumeral amputation who was fitted 14 years post-surgery without any surgical revision-but who was nonetheless able, within 30 minutes, to demonstrate six separate motions that Chi classified using pattern recognition.

"Theoretically, this is impossible to do without local reinnervation," Chi marvels. "My gut feeling is that I'll be able to classify and fit others with advanced prosthetics if they have had local reinnervation done.

"This concept is so novel, beneficial, and it seems so simple, yet no one is using it. So right now, I'm really on my soapbox, paper pleading with all trauma surgeons to change how they perform these amputations, and I'm using my patient as a case report/proof of concept." (Author's note: Chi's paper will appear in an upcoming issue of the New England Journal of Medicine or the Journal of Trauma.)

Pointing to the interesting new directions Boninger and Chi are exploring, McLoughlin underscores that the MPL team's focus is canted toward the research side.

"We have an arm that has 26 degrees of freedom, it's got over a hundred sensors on it, and we currently only know how to utilize a fraction of its capabilities," he says. "Although that fraction is a big jump over what has been done previously, it's still only the beginning of what the MPL could ultimately do. So our interest has been in facilitating the development of much more highly capable systems in order to accelerate advancements on the technology front.

"The other pathway, which is the DEKA Arm, is pushing down the commercialization route."

According to Ling, "DEKA is currently answering FDA review questions and has to complete some more trials and testing as part of the process. Once all directed testing satisfies the FDA, the DEKA Arm System will become commercially available."

Stewart Coulter, PhD, project manager for DEKA's highcapability "Luke" prosthesis, reports that the latest model of the arm, currently called the Gen 3, has undergone significant clinical studies, engineering testing, and additional take-home studies. "Based on those results, we're in the review process with the FDA now," he explains. "Once we get through this review, we'll be able to address next steps more clearly."

Photograph of the DEKA GEN 3 arm courtesy of Stewart Coulter.

"The Gen 3 offers a lot of improvements versus the Gen 2," Coulter says, "and those improvements are based on the feedback we got from the thousands of hours of clinical use of the arm system. We really want to thank the clinical study participants for their time and feedback on the arm system. Their feedback has been critical in the process. The VA [U.S. Department of Veterans Affairs] has been a great partner in this process as well."

Coulter adds that the study participants, prosthetists, and researchers have also provided valuable feedback on the design. "The Gen 3 Arm has been updated...to give it much cleaner lines," he says. "We've done a lot to make it look nice and be ready for the next step. Now we're pushing through the final pieces here. It's been a tremendous effort from that perspective."

Coulter is enthusiastic about the efforts of his team's prosthetist partners at Next Step Orthotics & Prosthetics, Manchester, New Hampshire, and biodesigns, Westlake Village, California. "They've been great to work with on this, and they are continuing to push some of the ideas on the socket side that they have been using. People are already benefiting from these developments, and I'm very happy to see that happening...in parallel with our progress on the arm itself."

What Lies Ahead?

With a staggering number of possible avenues to pursue, the challenges of managing a diverse group of collaborators appear equally challenging, as McLoughlin enumerates the efforts at Pitt and the California Institute of Technology (Caltech), Pasadena, and an amputee at Walter Reed National Military Medical Center, Bethesda, Maryland, who plans to incorporate autonomous or reflexive behaviors in a prostheses using conventional myoelectric control paired with pattern recognition, just to name a few.

"When you're trying to understand how to interface with the human brain, an extremely complex organ, you're going to need a network of research institutions looking at this problem," he points out. "It's our idea that we kind of pull that network together. You as a researcher can take your expertise and combine it with other technology and ultimately make progress orders of magnitude faster than would have been possible if you were off working by yourself."

Every DARPA researcher comments proudly-almost reverently -on a universal attitude that values partnership and teamwork in achieving goals, as well as a commitment to continually place the pursuit of knowledge and the needs of those who will benefit from it above personal or profitability concerns.

"Part of what DARPA is exploring are ways that we can make our findings available to the research community so that anybody who wants to do research can come in and work with the technology that's come out of this program. So...if you want to do neural research to look at ways of controlling an arm, you don't have to go and reinvent an arm of your own; you can utilize what has already been developed," McLoughlin says.

"It is clear that DARPA will not be able to do this alone," Ling reflects. "Achieving a wireless brain-control system for an advanced prosthetic limb and reliable implants that will last for decades is a challenge for the scientific and engineering communities. Federal strategies can inform the way ahead for these efforts, but public, private, foundation, and other creative arrangements will be required to achieve the necessary technological breakthroughs."

Ethical Questions

As research and development in neural engineering and neurally controlled prosthetic devices continues to progress, a number of ethical questions have been posed: How far do we, can we, should we go with research that taps into the human brain?

"The primary ethical challenge faced by any claimed 'revolutionary' device or system," says Ling, "is that it will replace everything that is currently used. This carries the potential to create false expectations. We must remember the nuances of each individual with loss of limb or loss of the ability to control their natural limbs; some have comorbidities that prohibit the use of certain types of technology.

Photograph courtesy of the Johns Hopkins University Applied Physics Lab.

"The widely used split-hook body-powered technology that is offered to every military upper-extremity amputee was invented 100 years ago and has served several generations of amputees. It is unlikely that the DARPA-developed arm systems will completely replace the split hook. Testimonials from users who have tried the DARPA arms during clinical and take-home trials indicate that amputees are excited about the restoration of many functions that the split-hook simply cannot perform, but these are different for each user. Our ongoing human trials using implanted devices to achieve brain control of a prosthesis should reveal in the near future how excited a quadriplegic is with the prospect of being able to apply the technology to accomplish activities otherwise impossible due to her condition.

"The goal of this program has been and always will remain near-natural restoration of capability to the injured. Distribution of resources to competing priorities will always exist, but the role of a program manager at DARPA is to continue to show the merit of an investment with measurable results. The Revolutionizing Prosthetics program continues to produce such results, so the effort still continues well after its originally planned termination date."

Boninger agrees. "I just spent a day seeing patients in a wheelchair clinic, where I saw half a dozen people whose lives could be changed immeasurably if this technology were available. So for me, it's really easy to remember what this is all about. Everybody understands-and it kind of makes you row a little harder."

Judith Philipps Otto is a freelance writer who has assisted with marketing and public relations for various clients in the O&P profession. She has been a newspaper writer and editor and has won national and international awards as a broadcast writer-producer.


  1. Collinger, J. L., B. Wodlinger, J. E. Downey, W. Wang, E. C. Tyler- Kabara, D. J. Weber, A. J. C. McMorland, M. Velliste, M. L. Boninger, and A. B. Schwartz. 2012. High-performance neuroprosthetic control by an individual with tetraplegia. The Lancet 10.1016/ S0140-6736(12)61816-9.