Strong Armed

THE PLACE DIDN'T LOOK like a cutting-edge research laboratory. The work benches in the little room at Johns Hopkins University's Applied Physics Laboratory (APL) in Laurel, Maryland, were cluttered with oscilloscopes and other gear for testing electronics, circuit boards, computer workstations, drills, files, miscellaneous bits of wire, and other gear reminiscent of a hobbyist's well-equipped workshop.

A mannequin at the door, wearing an articulated metal and plastic arm in place of its usual plastic, and a cabinet shelf stocked with realistic rubber hands betrayed the lab's true purpose: the creation of a prosthetic arm so close in form, appearance, and function as to be virtually indistinguishable from a natural human limb.

Thirty-five-year-old mechanical engineer and former Marine reservist Jon Kuniholm, solidly built and still maintaining a short military-style haircut, had more at stake in the proceedings than his colleagues. In place of his right forearm, he wore a carbon fiber prosthesis that terminated in a split hook. Its design had changed little since 1912, when inventor D. W. Dorrance, who had lost his own arm in an industrial accident, patented a split hook prosthesis. Standing in the APL lab in the summer of 2007, speaking with quiet intensity, he turned his body to show the straps that linked the simple yet elegantly effective hook mechanism to his shoulder. As Kuniholm reached his prosthesis forward, the hook opened, then closed when he retracted it.

Even though the hook was unashamedly a hook, without cosmetic adornment, its simplicity and utility had made it more popular with amputees than any other design. Kuniholm, like his colleagues at APL, hoped that their latest high- tech prosthetic project would be different— but he wasn't counting on it.

Most of the challenges to the hook's supremacy to date had failed at least in part because of an imperfect marriage of form and function. Engineers had worked hard to create a prosthesis that closely resembled a flesh-and-blood arm and, in so doing, had sacrificed utility. The tension between aesthetic appeal and usefulness still asserts itself today, as researchers seek to answer essential questions about prosthetic design, and raises larger questions about the source and inspiration for such technological innovation. Is the best path to useful innovation through well-funded, but historically short-lived, government programs? Or can private enterprise best meet the needs of users through the demands of the market? What about engineers and designers freely sharing their expertise in the service of a common goal via the open-source model of development?

Humans have sought artificial replacements for their limbs for as long as they have managed to endure their loss. The earliest example of a prosthesis to survive to modern times, found in an excavation at Capua, Italy, dates back to about 300 B.C.E. Reinforced with bronze fittings, the wooden core attached to the wearer's body (found with his artificial leg) by means of a belt of sheet bronze and leather. Reports of artificial limbs go back as far as to Herodotus in fifth century B.C.E. Greece, and even earlier in the Hindu Rigveda , dating around 800 to 1500 B.C.E. From the beginning, prosthetic legs have presented an easier prospect than artificial arms for restoring their wearers to something approaching their normal function. Legs, after all, "only" have to enable their users to walk, run, jump, and occasionally bend enough to sit. Prosthetists are further helped by a leg's position, which is firmly anchored between body and ground. Hands, with their 27 separate possible movements or "degrees of freedom," as they are known to prosthetists, prove a much greater challenge. Swinging freely from the body, arms and their attached hands can drive a sledgehammer or grasp a small paintbrush. "My hand is to me as a god," wrote Virgil. Few poets have reserved such praise for the lowly foot.

Such difficulties did not stop the German knight Götz von Berlichingen, who lost his right arm in 1504 while fighting in the Landshut War of Succession between rival Bavarian duchies. After recovering from his wounds, the 24-year-old instructed an armorer to craft an iron arm and hand. The resolute knight wore his prosthesis in various conflicts over the next 40 years, his iron hand either locked around the reins of his horse or the pommel of his sword.

The birth of modern prosthetics had to wait, however, until the industrial age and the onset of the American Civil War. A Union canon ball tore off James Hanger's left leg in 1861, giving him the dubious distinction of being the first to lose a limb in the war. Dissatisfied with the prostheses available to him, he invented his own. The company he started later, now known as Hanger Prosthetics and Orthotics, produces prosthetic limbs to this day. As the number of war-related amputations grew, the federal government and many states rushed to provide limbs at no cost to maimed veterans, yet fewer than half of those took advantage of the benefit.

"It is probably more necessary that he should have an eye than it is for most soldiers to have arms," wrote one veteran's brother to a state board charged with providing prosthetics to veterans. "There are very few who will be most benefitted except in looks by artificial arms."

Such arms proved little better than decorative appendages, bearing hands generally fixed in one position. No wonder so many vets ignored the cumbersome prostheses, pinned up their sleeves, and wore their wounds proudly.

D. W. Dorrance's 1912 breakthrough came with his realization that a prosthetic arm need not be created to resemble a natural arm. What Dorrance wanted—and needed—was a means of restoring as much of his natural abilities as possible. He didn't care what the final system looked like as long as it worked. The result was the Dorrance hook, split to maintain a fixed surface opposed by a movable opposite member. Rubber bands held the two hooks together under tension. To open the hook, the user extended his or her arm; to close it, he bent his elbow. Closing their devices around objects such as umbrellas or knives, Dorrance and the thousands of users who followed him could resume some semblance of normal life. Variations of this so-called body-powered hook have remained the prosthetic arm of choice for amputees ever since, and they continue to be manufactured by the Hosmer Dorrance Corporation.

After returning from Iraq and undergoing multiple surgeries to remove what was left of his forearm after an improvised explosive device (IED) exploded near him, Kuniholm was issued three separate types of prosthesis: a split hook; a specialized arm best suited for holding a writing instrument; and a state-of-the-art, electrically powered and realistically styled prosthesis called a myoelectric, or myo, arm, which took its cues for movement from electrodes pressed to the skin of the remainder of his arm.

Kuniholm found that he far preferred the hook for everyday tasks. Its light weight enabled him to wear it for hours with minimal discomfort. It required little maintenance and responded more reliably to his commands than the high-tech myo arm. He could more easily grasp objects and help dress himself with the hook. Kuniholm disparaged the heavier, clumsier, and at times unpredictable myo arm as "a rigid, hand-shaped electric clamp."

Kuniholm knew there must be a better way to build a prosthetic arm. He and other members of his North Carolina high-tech firm, which had worked on such projects as designing an IED-disabling robot, looked into the matter. What they discovered surprised them. Not only had the design of the arm he now favored changed little since 1912, but attempts made over the ensuing decades had met with limited results.

As in wars past, World War I sparked a surge in demand for prosthetic limbs. European countries looked to the United States for the expertise and manufacturing capability that had grown following the Civil War. "American-made artificial limbs are the best in the world," boasted one prosthetics manufacturer to a New York Times reporter after visiting Paris in 1915. "The French are asking that American artificial limb factories be established over there so that the demand may be met on the spot." Incremental improvements followed in the 1920s, resulting in the development of the Trautman hook, which received a patent in 1925. The Trautman was a refinement on Dorrance's design, with flat pads on the two curving "fingers" in place of hooks, and only three metal parts fastened together with a couple of screws.

World War II brought renewed interest in prosthetics research, leading to the formation of the U.S. Committee on Prosthetic Devices (later the Prosthetic Research Board) in 1945. With access to annual funding of $1 million by 1948 (approximately $8 million in today's dollars) and the participation of dedicated laboratories around the country, the project pushed the envelope in prosthetics. The Army Prosthetics Research Laboratory (APRL) at Walter Reed Hospital produced the APRL hand, which substituted a spring for the more usual rubber bands holding earlier systems closed. It attempted to combine form and function, essentially blending a hook-type design (with a body-powered closing function, as contrasted with the Dorrance hook's opening function) with a realistic vinyl cosmesis. Neither that device nor the APRL hook, which were developed in parallel, however, succeeded in unseating the old split hook design. (One World War II veteran complained that the hook was "a slingshot in an atomic age.") Still, such efforts opened the research campaign that produced today's myoelectric arms.

Early on, government researchers had used an electromyograph and surface electrodes to measure the electrical signals naturally produced by moving human leg muscles in order to develop their own designs. Thence it was a natural step to use the signals generated by a person's remaining muscles to control an electrically powered prosthetic. In 1964 work in the Soviet Union culminated in the first myoelectric prosthetic arm. The user could open and close the system's pincer grip (shaped like a human hand and covered with a rubber cosmesis) by contracting his remaining forearm muscles. Electrodes on the skin's surface picked up the electrical signals generated by underlying muscles and translated them into movement.

Canadian researchers who studied the arm that same year reported that while "the Russian bioelectric prosthesis at present has only two motions: opening and closing of the hand ... there is no doubt that [it] represents a major contribution in the field of prosthetics." Western manufacturers improved on the design with better quality materials in time for veterans of the Vietnam War. Myoelectric arms gained further traction in the 1980s and into the 1990s. But many users, especially veterans still in the prime of their lives, came to the same conclusion that Kuniholm would: the bulky, underpowered, and temperamental myoelectric arms, such as the APRL hand before them, just couldn't effectively compete with the simple body- powered hook for everyday use.

The wars in Iraq and Afghanistan have provided the impetus for the latest burst of innovation in prosthetics. Col. Geoffrey Ling, both an active duty Army trauma doctor and a program manager with the Defense Advanced Research Projects Agency (DARPA), the Pentagon research and development branch behind such transformative technologies as the Internet and GPS, saw, during separate tours of Afghanistan and Iraq, the toll those conflicts were taking on human limbs and vowed to do something about it.

"DARPA had the money, a minimal bureaucracy, the right attitude, and a stated mission of pushing back the frontiers of science," he said. "I came back really invigorated to do this." The result was DARPA's Revolutionizing Prosthetics Program, launched under Ling's direction in 2005. Its aimed at nothing less than the development of a prosthetic arm and accompanying hand that would restore all—or almost all—of a user's natural abilities, including a range of activities that would enable him or her to return to active military service if they so chose. This $100 million project, the largest and best-funded of its kind, could forever change the notion of what it means to be disabled.

Kuniholm jumped at the chance to participate in the program as a Ph.D. student working for the project's lead contractor and project integrator, the Applied Physics Laboratory. As the only member of the engineering team who actually needed a prosthetic arm, he had much to contribute.

In the lab Kuniholm strapped on a carbon fiber sleeve that sprouted a nest of data cables running to a control box near his feet and thence into a card in the back of the personal computer at which he was working. On the LCD panel in front of him, a virtual silvery prosthetic hand moved through a series of grasps in response to Kuniholm's muscle flexions. Pattern matching software running on the computer sought to interpret his flexes, transmitted by the electrodes pressed to his forearm, into preprogrammed movements. This refinement on conventional myoelectric controls served as the control method for the so-called Proto 2 arm, which the team was feverishly attempting to finish before a semipublic demonstration at the DARPA Technology Symposium the following week. They hoped to replace the bundle of cables and the control box that for now tethered Kuniholm to the PC with a microchip.

A main part of APL's mission is to integrate the work of some 20 other organizations around the world also working on the Revolutionizing Prosthetics Program. Along with APL, the other members of the team are working on the three major problems of prosthetics: articulation, control, and power. At a workbench across the room from Kuniholm, a couple of engineers seemed to have licked the problem of articulation. Their intrinsic Proto 2 hand, so named because it contained all of the motors necessary for flexing the fingers within itself, was capable of 24 separate degrees of freedom— dextrous enough in theory to allow its user to play the piano. In practice, the myoletric controls driven by surface electrodes could pick up only a fraction of the number of signals necessary for such fine control, which is why Kuniholm was training the computer to respond to his commands with preprogrammed grasps. The different movements depended not on only what muscles he flexed in what sequence, but at what speed he flexed them.

The team believes that, in the not-too-distant future, their system will make use of injectable myoelectric sensors, or IMES, which have been developed at the Rehabilitation Institute of Chicago. Embedded directly into the muscles they are meant to read, the IMES should theoretically attain a much higher degree of precision than noisy surface electrodes. As a bonus, they are impossible to pull loose, giving users more latitude of movement.

Researchers in Pittsburgh are also making great progress with electrodes implanted directly in the brain, which have the potential to pick up a user's intention to move a limb. Carnegie Mellon roboticist Andrew Schwartz and colleagues nearby at the University of Pittsburgh published the results of their experiments last year, along with a startling video. In it a rhesus monkey manipulated a robotic arm to pick up a morsel of food and bring it to its mouth, all by only using its thoughts. "We have expanded the capabilities of prosthetic devices through . . . closed-loop cortical [brain] control," the authors announced.

But that still leaves the matter of providing power to a prosthesis. The best controls and actuators will do little good if the user has to wheel around a heavy battery that fades after a few hours' use or needs to remain plugged into a wall socket. The electric motors in APL's Proto 2 were an obvious choice as readily available actuators. But when the Proto 2 upper arm and forearm were put through an amazingly lifelike sequence of movements on a mannequin at the APL, the effort involved betrayed the system's shortcomings. The arm, though elegant in appearance, maintaining the weight and form of an adult male limb, drew its power through cables partly concealed behind the mannequin's back, which plugged into a heavy-looking power supply box on the floor.

Vanderbilt University's Michael Goldfarb and his team hope to overcome the limitations of electrically powered activation with a novel monopropellant system. In the future, a user would plug a fresh hydrogen peroxide cartridge into his arm each morning; small bursts of mono- propellant would drive the fingers and other articulated joints of the arm by reacting with a catalyst to release steam. Goldfarb envisions that the cooled steam would seep through pores in the arm's cosmesis much like natural sweat in the skin.

Because Ling was anxious to rush his arms to the veterans who needed them, he also commissioned a "strap-and-go" arm to be completed in just two years, well ahead of the APL arm's four-year deadline. DEKA Research in Manchester, New Hampshire, won the contract to build it, and the Veterans Administration is now putting it through clinical trials. While the DEKA arm requires no implants or additional surgery, it does take some training to operate. Double arm amputee Chuck Hildreth modeled the DEKA at the 2007 DARPATech conference. In contrast to the form-fitting elegance of the APL prostheses, the DEKA featured an unapologetically robotic appearance, gripping Hildreth's body with computer-controlled airbags that shifted the system's weight in concert with his movements for maximum, long-wearing comfort. It drew its electric power from power packs on Hildreth's belt. He controlled it with specially designed joysticks, one in his shoe, the other attached to the atrophied remnant of his left upper arm. He reveled in performing activities denied him for the 26 years since he had lost his arms in a power line accident. He enthralled a crowd of military officers, DARPA staffers and contractors, and journalists by operating a drill. He even plucked a small piece of candy off a table and brought it deftly to his mouth.

While the DEKA arm appears within reach, the realization of Ling's ambitious goal for the APL arm seems much farther off. Yet the unprecedented $100 million investment in the Revolutionizing Prosthetics Program will continue to advance state-of-the-art arm prosthetics as never before. Kuniholm, for one, remains concerned that such heavy research investment may never come again. With only about 100,000 people in the United States missing one or more arms or hands, the market simply doesn't exist for private enterprise to continue with vigor when the government leaves off. "As an engineer, I'm thrilled to be involved in the creation and testing of this latest batch of prosthetic 'concept cars,'" reflects Kuniholm. "As a customer, my greatest fear is that of failure to complete the cycle of product development and actually bring it to market." And if the advanced arms do reach the market, he further worries that they'll be too expensive; even conventional myoelectric arms cost over $30,000.

That's why Kuniholm and his colleagues at Tackle Design have launched a prosthetics project of their own, based on the open source model of computer software development. Members of the far-flung Open Prosthetics Project (OPP) contribute their time to the goal of crafting affordable arm prostheses, including an improvement on the tried-and-true Trautman hook. Their goal is to mass produce an improved design, called the OPP T-Hook, for $150 each.

Kuniholm and members of OPP began by reverse engineering an old Trautman hook to create computer- aided design models of the three main parts. They then e-mailed the newly created design to a rapid prototyping service, which sent them back a 3D model made of composites. That enabled the team to test the parts for fit and movement before ordering metal prototypes for real- world testing by OPP members. Meanwhile, one enterprising OPP member, John Bergmann, went to work on a mechatronic hand of his own design made of LEGOs. The advantage of LEGOs? They're inexpensive, easy to come by, and offer lots of flexibility for building prototypes. To activate the fingers, Bergmann turned to ordinary zip ties, serving the same function as tendons in a natural hand. Next steps for other OPP members: developing an actuation system (motors) and a low-cost control system for the LEGO prototype. Even with all these good ideas flowing into the project, however, OPP has so far had a hard time finding commercial partners to take its design work into production. As in times past, the market represented by the relatively small number of amputees does not promote innovation.

Still, the unprecedented combination of a big government project expressly designed to drive major innovation in the marketplace and a cohesive community of dedicated amateur craftspeople working on affordable, low-tech solutions seems destined to advance prosthetics as never before. Whether advancement will come rapidly enough to suit young and understandably impatient amputees such as Kuniholm is another matter. This hit home for him in a very personal way as the APL Revolutionizing Prosthetics project wound toward its conclusion at the end of 2009 with funding for only two prototypes. Both of the prototypes are left arms—not the right one that he had hoped to test drive.