The loss of a limb is a life-changing and often devastating event. As a result, the use of artificial approximations as replacements has persisted for both cosmetic and functional reasons. Historically, the technology used to produce these important artificial limbs has always failed at mimicking either the lost limb’s appearance or movement. Artificial limbs even today are often difficult to use, painful, and of limited help to the user. In fact, the execution of many simple tasks is fatiguing and difficult simply because of the hobbling effect and primitive state of the artificial limb. Fortunately, with a recent surge of interest and funding, these barriers are beginning to crumble in the face of renewed research efforts driven in the United States by demand from wounded veterans of the Iraq War (1). Artificial limbs today are marching steadily toward a future where limb replacement means much more than merely dampening the traumatic effects of losing a limb, and truly means regaining functionality.
The research arm of the U.S. Department of Defense, known as the Defense Advanced Research Projects Agency (DARPA), has been awarding grants worth millions of dollars to various laboratories and companies as part of its “Revolutionizing Prosthetics” program. This program aims to create an artificial arm and hand that interfaces directly with the nervous system by 2009. This prospective prosthetic would be “fully functional”, replicating the natural arm’s range of movement and giving sensory feedback (2). It would thus allow for a near-perfect replacement of the lost limb (2).
Two of the most lucrative recent grants in this project have gone to DEKA Research and Development Corporation in Manchester, New Hampshire, and to Johns Hopkins University. DEKA was awarded an $18.1 million grant for the creation of a prosthetic arm that would mimic a real arm in both cosmetic appearance and strength, while researchers at the Applied Physics Laboratory at Johns Hopkins received a $30.4 million grant for their work (3).
The Applied Physics Laboratory at Johns Hopkins University rapidly developed a prototype called “Proto 1” within a single year of receiving the grant. Proto 1 provides sensory feedback and provides eight degrees of freedom in movement. Although Proto 1 represented a large leap ahead for prosthetics technology, there is more progress to be made. The Applied Physics Laboratory is almost ready to unveil a second prototype with 25 degrees of freedom, the speed and the strength of a natural arm, and many more sensors for feedback (4). Proto 1’s ability to provide sensory feedback and neural control of movement are made possible by a cutting-edge technique known as “Targeted Muscle Reinnervation” (4).
Dr. Todd Kuiken of the Rehabilitation Institute of Chicago developed Targeted Muscle Reinnervation in order to transplant nerves that had once innervated the amputated limbs to a new area. In the case of a patient named Jesse Sullivan, Dr. Kuiken redirected the nerves for the removed arm that were running through the shoulder to several different muscle groups across the proximal pectoralis muscles (4, 5). By six months after the surgery, electrical signals were detected from the transplanted nerves (1). Contractions in these muscle groups were also felt on the surface (5). The effectiveness of this transplantation was then tested by fitting the patient with a new experimental myoelectric prosthesis that would exploit the transplanted nerves. The patient subsequently showed increased dexterity and speed of movement with the prosthesis, and reported that the movement was much easier and more natural when compared to previous prostheses (5). In fact, the transition was almost effortless. The first test was conducted by doctors who simply asked the patient to try to open his absent left hand. The connected prosthetic test piece immediately uncurled its hand (1).
Both the Proto 1 prosthesis and the effectiveness of the Targeted Muscle Innervation technique were demonstrated during subsequent clinical evaluations. Jesse Sullivan immediately displayed an increased capability for fine muscle control. Sullivan was able to quickly manipulate the prosthetic hand and deftly remove a credit card from his pocket. He was also able to display the natural force feedback of Proto 1’s control interface by stacking cups with a gentle, carefully controlled grip. Lastly, but of no less importance, the cosmetic covering for Proto 1 featured a photorealistic layer that was fabricated according to images of the patient’s arm before amputation (4).
Proto 1 is just the beginning for this method of prosthetic control. Sullivan’s chest muscles hold much potential at this point. The nerve currently used to control the closing of the artificial hand actually innervates at least 20 muscles. Those 20 muscles are currently serving to propagate only two signals. With more development and higher resolution, this nerve (and the other three that were reinnervated) can serve many more purposes (1). In fact, a second much more advanced prototype is already nearing completion. The researchers at the Johns Hopkin’s Applied Physics Laboratory have been hard at work on “Proto 2,”, which will feature over 25 degrees of freedom in movement, as well as strength and speed approximating a human arm, and over 80 sensors to provide feedback for touch, temperature, and position in space. Proto 2 may also feature the use of “Injectible MyoElectric Sensors” rather than the surface electrodes used to control Proto 1. These sensors will allow Proto 2 to interface with the necessary nerves through implanted or injected devices rather than exposed surface electrodes, and will also help ensure the reliable transmission of commands from nerve to prosthetic (4).
Technological advances toward this kind of neural control of machines have been long in coming. While science-fiction writers in the past have often dreamed of the ability to directly interface with machines for augmentation, replacement, or other purposes, it was only recently that monkey studies proved that these types of direct neural linkages were possible. For instance, one study performed at the Duke University School of Medicine and published in 2000 showed that electrodes implanted in an owl monkey’s brain were able to impart upon the monkey the ability to control a robotic limb. The owl monkey’s brain signals were monitored in order to initially identify the particular brain signals correlated with specific arm movements. Once this had been accomplished, a computer and robotic arm were attached. The robotic arm was directed by a processing computer that monitored the owl monkey’s brain signals for the previously identified patterns of neural firing. When a pattern was identified, the processing computer would relay the appropriate instructions to the robotic arm to cause a similar movement. This system successfully allowed the robot arm to mimic the actual arm movements of the owl monkey (6).
Although Dr. Kuiken’s aforementioned muscle reinnervation technique works wonders for upcoming prosthetics of similar design – the method used for the owl monkey – connecting the artificial limbs directly to the brain – holds much potential. Brown University’s Brain Science Program director, John Donoghue, also serves as the chief scientific officer for Cyberkinetics Neurotechnology Systems, located in Foxboro, Massachusetts. Cyberkinetics Systems is currently developing a small square chip, two millimeters on each side, for implantation in the primary motor cortex. This chip, known as “BrainGate”, sends motor cortex signals to an external processor for interpretation. This processor in turn operates the prosthetic. The system has been successfully tested on a paralyzed man named Matt Nagle. Nagle was able to use the BrainGate interface to move a cursor on a screen, and even to open the hand of a prosthetic arm. However, BrainGate still requires wireless capability and practical portable power sources to be truly useful (1).
While upper-body prosthetics require fine motor control for handling different objects, prosthetics for the legs require careful and timely responses in order to maintain balance, navigate different types of terrain, and allow for natural movement. In order to truly maintain balance, the wearer must also be informed of the prosthetic’s location in space. Feedback for the legs is just as important as feedback for the arms. Possibly more important for legs, however, is the issue of power. Without a boost from the prosthetic, even walking becomes a very difficult and exhausting task. Motorized prostheses would ideally be able to dampen the forces received and modulate the power delivered in order to fit the terrain or the task, whether the wearer is walking uphill, going up stairs, or even running (1).
The multiple and complex requirements necessary for leg prostheses have driven one leading company, Ossur, to develop knee and foot prostheses separately. Ossur’s Power Knee and Proprio Foot products both include motorized movement to help propel the user, and both are capable of swinging naturally to help impart a natural gait. The two devices work in sync to perform the complex tasks normally performed by the leg. The Power Knee provides the motorized support to lift the user from seated positions, up stairs, and up inclines as the Proprio Foot adjusts to these differing types of terrain and shifts beneath the user as they stand up. Both the Power Knee and the Proprio Foot work to lift the foot a proper height off the ground when walking. Ossur appears to have put much thought into the teamwork necessary for a natural walking motion (7, 8).
However, for leg prostheses, there are more upcoming options than such bulky and robotic attachments. Current research at Walter Reed Army Medical Center and Arizona State University is focusing on a project that uses lightweight springs to store energy. The prosthesis, dubbed “SPARKy” for “Spring Ankle with Regenerative Kinetics”, is meant to store enough energy and provide proper ankle motion to make it comparable to a natural leg. The teams researching SPARKy have also put much effort into learning about natural gait. The simplicity of SPARKy is a direct result of the research that reduces walking to a controlled “series of falls” (9). One heel swings forward and touches the ground. As the load is transferred to this leg, the springs in SPARKy store the energy. As the body’s weight continues forward, the springs begin to release the stored energy, providing a forward and upward push as the heel leaves the ground again. This heel will swing forward to catch the falling weight of the body to start the cycle anew. All that is needed is a small motor to tune the springs for optimal performance. SPARKy weights only about two pounds, and has been successfully tested and shown to provide a natural walking gait. By 2009, it should be completed with additional functionality to allow for everyday use (9).
Focusing on the leg would be of limited use without similar focus on the intricacies of the foot. At the moment, “no prosthetic foot has yet been produced that can imitate the natural sequence of movements during walking,” according to Dr. Urs Schneider of the Fraunhofer Technology Development Group TEG of Stuttgart, Germany (10). Even the Ossur Proprio Foot mentioned earlier does not have the three-axis flexibility of the natural foot; instead, it replaces the entire foot with a flexing pad that cannot reliably duplicate the flexibility and adaptability of the natural foot. As a result, Dr. Schneider and his team have worked to develop a complex mechanical prosthetic foot that can imitate the shifting of the foot during movement without the aid of any computers. This reliable reproduction of the foot’s actions promotes natural walking and reduces the time necessary for users to acclimate to their prosthetic. In fact, testing of the device has been so successful that outside observers generally do not recognize that the user is wearing a prosthetic foot (10).
The challenge of leg prostheses grows more complex, however, with the question of feedback information. Despite all the motorized, computerized, or spring-loaded technology of the lower prosthetic, without sensory feedback mechanisms, the user lacks the necessary innate sense of where the leg is in space (1). Without this particular sense, a natural gait is much more difficult to achieve. It becomes difficult to coordinate one leg with the opposite limb when the natural gait is achieved by automatic swinging motions of the artificial leg. Presumably, it would be not be possible to turn on the organic leg and expect the mechanical version to properly place itself in the new direction. It therefore may not even be possible to turn sharply without first stopping. At the least, it seems that turning will be difficult at best. The solution to this problem may lie in further developments in feedback and increased articulation in the ankles.
Leg and arm prostheses also share the complication of attachment points. Most weight-bearing attachment points today for leg prostheses are painful sockets that fit rather poorly and often impede natural movement. Arm prostheses may be strapped awkwardly around the body, and again may fit poorly or come loose due to the amount of motion. Despite innovative socket designs that attempt to compensate for poor fit, the very existence of the socket creates unnecessary complications (1).
The solution to these complications is to simply fuse the prosthetic directly to the remaining bone. However, although the technology to fuse bone to the metal of the attachment rod exists, the attachment point that must protrude out of the skin prevents this solution from being realized on a large scale. The skin generally fails to heal around the protruding attachment point, and infections subsequently invade the area (1). This is perhaps the largest weakness for future prosthetics. No matter how complex the prosthetics may become, they will always have to compensate for poor attachment. There is still hope, however, that a material or method will be found that resolves this issue. Jeffrey Morgan, a molecular biologist at Brown University, notes that metal and skin should be able to seal naturally, as they do with pierced noses (1).
In the future, prosthetics may be limited in their advancement by the advent of new technologies. Left alone, the prosthetics industry could eventually produce prostheses that are nearly indistinguishable from their organic counterparts, but they may yet be supplanted by an even better replacement: organic replacements grown from stem cells. Although it still sounds far-fetched, lab-grown replacement limbs would spell the downfall of the prosthetics industry. Even now, with hand transplants and face transplants being successfully performed around the world, technology is approaching the day where a lost limb may be grown to order or regenerated.
Whether the prosthetics industry would fully disappear upon development of the lab-grown replacement limb is debatable. Initially, of course, prosthetics would clearly be the more economical choice, but as the price for replacement limbs falls, the decision becomes a matter personal choice. Some would likely prefer the natural choice of regaining an organic limb, particularly if issues surrounding rejection could be resolved. However, others may prefer a prosthetic because recovery from a limb replacement surgery could be extensive and difficult. It is also possible that prosthetics will become more than just replacements for lost limbs, but rather augmentations of the human body, granting greater capabilities to the user.
However, despite the possibility that prosthetics will ultimately fade out in favor of organic replacements, prosthetics research today is still extremely worthwhile. Organic replacements, if feasible, would be complex due to both ethical dilemmas and scientific uncertainties. The concept of organic replacements is still too distance for patients to rely on. Amputees today still need the prosthetics industry, and the industry must continue to advance because much work needs to be done before prosthetics can truly approximate the limbs they supposedly replace.
The goals for prostheses in the near term are few and obvious. Unfortunately, each entails finding, researching, and implementing a myriad of complex solutions. Much information must travel through the prosthetic-organic junction, and so control interfaces and schemes must be devised to allow for the complex array of sensory feedback and simultaneous multidimensional motions possible with organic limbs. Only after this has begun to improve can prostheses begin to incorporate the additional sensors and degrees of freedom of movement necessary to be as useful as limb replacements. Additionally, prosthetics must be less painful and less exhausting to use. A fully functional prosthetic that causes severe pain and fatigues the user is fully functional in name only. Lastly, direct bone attachment may be a neat near-term solution for the problem of prosthetic attachment since it is used extensively for internal prostheses like joint replacements.
Fortunately, it appears that in the short-term, much of this work will be completed. It is likely that the discomfort of prosthetics will soon be resolved with direct bone attachment. The advent of both targeted muscle reinnervation and direct brain interfaces like BrainGate give great hope to the short-term possibility of instinctive, comfortable neural control of prostheses. These interfaces will also likely be able to handle much of the data flow for sensory feedback and simultaneous complex multidimensional movements. The Defense Advanced Research Projects Agency may not have their fully functional arm prosthesis by 2009, but the new prostheses that year will be very close to fully functional.
The future of prosthetics is bright, with many avenues of research left to explore. There is much to accomplish in terms of realistic appearance, natural range of movement, and the ever-complex sensory feedback. The ultimate goal, of course, is to replace the missing limb with a high degree of fidelity so that the user will be able to function normally. As science surges forward in the fields of mechanics, biomedical engineering, computer science, and neuroscience, bionics will follow. Hopefully, if we cannot learn to regenerate and heal such extensive injuries as the loss of a limb, at the least we can learn to produce a suitable replacement in the future.
1. S. Sataline, Popular Science. July 2006, p. 68.
2. DARPA Defense Sciences Office – Revolutionizing Prosthetics. Available at http://www.darpa.mil/dso/thrust/biosci/revprost.htm (25 May 2007).
3. D. Miles, DARPA’s Cutting-Edge Programs Revolutionize Prosthetics (2006). Available at http://www.defenselink.mil/news/newsarticle.aspx?id=14914 (25 May 2007).
4. New Prosthetic Limbs Allow For Eight Degrees of Freedom (2007). Available at http://www.news-medical.net/?id=24306 (25 May 2007).
5. T. Kuiken et al., Prosthetics and Orthotics International. 28(3), 245 (2004). Available at http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=15658637&dopt=Citation.
6. J. Stephenson, Journal of the American Medical Association. 284, 2987 (2000). Available at http://jama.ama-assn.org/cgi/content/full/284/23/2987-a.
7. Bionic Technology – Ossur Power Knee (2007). Available at http://www.ossur.com/bionictechnology/powerknee (25 May 2007).
8. Bionic Technology – Ossue Proprio Foot (2007). Available at http://www.ossur.com/bionictechnology/propriofoot (25 May 2007).
9. Next Generation of Powered Prosthetic Devices Based On Lightweight Energy Storing Springs (2007). Available at http://www.whatsnextnetwork.com/technology/index.php/2007/05/02/next_generation_of_powered_prosthetic_de (25 May 2007).
10. Artificial Limbs That Walk Naturally (2006). Available at http://www.gizmag.com/go/5298/ (25 May 2007).