FEATURE ARTICLE

May 2005

Casualties of War

Robotic, Biological Research Aiding Military Amputees

By Joe Pappalardo

Hugh Herr sits alongside a colleague, watching the man’s extended right foot rotate in mid-air. Herr, a double amputee who lost his legs to frostbite when he was 17 years old, attempts to replicate the motion, but the brain’s commands stop at the stump just below his knee. Nevertheless, Herr’s nervous system is still sending electrical signals as if the entire limb were there.

Researchers, the Massachusetts Institute of Technology assistant professor says, need to translate those brain signals into a language a robotic limb can understand. That way, advances in artificial limbs have the potential to approach the capabilities of a real leg. “In five to 10 years, I think we’ll see an unprecedented level of innovation,” Herr says.

Herr is a leading researcher in the Department of Veterans Affairs new five-year program, aimed at helping soldiers who lost limbs in combat. At the core of this program are new technologies meant to seamlessly fuse prosthetics with the human body.

Herr’s role in this new effort is creating ankles and knees that are controlled by the amputee’s nervous system and possess their own self-generated power. His blueprint for this is the human body: “We’re trying to steal nature’s secrets.”

It is the first time that the many disciplines needed to create a bio-robotic hybrid limb have been gathered under one structure. Success depends on developing new ways to grow bone, fuse robotics with the body parts and control artificial limbs using commands from the brain.

The VA’s renewed emphasis on limb loss is its answer to the casualties of wars in Afghanistan and Iraq. According to statistics from the New England Journal of Medicine, for every soldier killed in Iraq, nine others have been wounded and survived, the highest rate of any war in U.S. history.

“Thanks to advances in battlefield first aid and protective body armor, many soldiers or Marines who otherwise would have been killed in action are able to return home,” says Stephen Fihn, the VA’s chief research and development officer. “Unfortunately, many of them must undergo amputation of the feet, hands, arms or legs, and in many cases, multiple limbs.”

The VA’s research and development service in late 2004 dedicated $4.7 million to establishing a center that is dedicated to researching these cutting-edge technologies to help make prosthetics more effective. The project involves partners such as Herr at MIT and medical researchers at Brown University.

Roy Aaron serves as the director of the Restorative and Rehabilitation Center that was created specifically for the project. It is located at the VA Medical Center in Providence, R.I. Aaron says that the project now just exists as a sign on a door, but he hopes it will evolve into two physical centers for research and rehabilitation.

For now, all research and clinical care is being done at the research labs and hospitals, a “center without walls.” It is his job to integrate the many sciences into actual products that are used by amputees. In February, center officials visited Walter Reed Army Medical Center to coordinate the research and assess the needs of their patients.

“It’s important to recognize this as more than just a research grant,” Aaron tells National Defense. Some of the technologies will be ready before others, and the goal is to transition useable products as soon as possible. He anticipates having advances reach VA hospitals within three years. “Most of the research is that close.”

For example, advances involving material science, such as improved titanium, will be developed years before tissue-engineering techniques are perfected.

Known as “osseo-integration,” the process of connecting bone to metal is a vexing one. Recent advances indicate that coating the titanium with a porous mixture at the spot where it contacts bone helps the bone grow into the metal, Aaron says.

In other cases, the metal juts from the skin to attach with the prosthetic. The point of contact where the skin meets the metal is prone to infection. If the skin could be coaxed into growing into the titanium, there would be no place for an infection to set in.

Scientists at Brown University are altering skin cells, called keratinocytes and fibroblasts, which make up skin layers, so they are able to bond to metal as they multiply. That aspect of the center’s work teams an engineer with a biotechnology expert to pursue a breakthrough.

Other forms of tissue engineering involve damaged muscle and cartilage. “The largest hurdle is regrowing as much biotissue as we can,” said Aaron.

One goal is to regenerate cartilage by releasing hormone-like proteins that inspire cellular growth. However, they have to be released in the right order, and in the right place inside the body.

Researchers are using biodegradable beads that are smaller than a pinhead and inserting them inside a joint to release the protein. The cells that grow have the benefit of artificial scaffolding that is inserted to protect them. Other methods involve genetically altering living cells to make and release growth proteins. Those cells could be placed within the micro-beads and likewise be placed inside the joint.

Aaron’s center is also perfecting a technique for lengthening bones that has existed for decades. In the 1940s, a Russian surgeon from Siberia, Gavriel Ilizarov, began inserting pins into stunted or deformed bones, producing small fractures. When the bone heals, it creates more bone in the direction of the fracture.

By manipulating this process with an external frame, a leg bone can be extended. For amputees, this means expending less energy to use a joint and increase the range of movement. The surgeon-in-chief of the project, Michael Erlich, will join Aaron in researching cellular studies to boost a bone’s rate of healing. New techniques of limb lengthening, officials at Brown say, likely will be the first procedures tested in humans.

For the foreseeable future, robotics will remain crucial in creating advanced prosthetics. In this, Herr has the awkward position with the VA center of making his newly commercialized effort, the Rheo Knee, obsolete.

The Rheo is the first artificially intelligent knee system that has the ability to learn and adapt to its user’s movements. The leg uses magnetorheological (MR) fluid between sliding steel plates to generate motion. MR fluid is an oily liquid approximately three times denser than water. However, the fluid instantly turns to a near-solid as soon as a magnetic field is applied. By controlling the fields, one can control the thickness. The product is applied in nautilus equipment, car shock absorbers, washing machines—and now prosthetics.

In the Rheo, the fluid’s thickness determines the resistance in the knee. This makes it bio-mechanically closer to the way a real knee works, which constantly adjusts to pressures. The MR actuator acts at the rate of 1,000 times per second.

Herr says that the absence of high-pressure fluids eliminates the need for seals and valves that are common areas of breakdown and leaks. “We’re hoping this will be more durable than hydraulic systems,” he says. “At peak torque we’re still at ambient pressure.”

But as advanced as Rheo is, Herr sees a boom in progress coming. The next wave will feature something new, something closer to what nature designed. Current prosthetic knees have computer-controlled brakes. What they lack, notes Herr, is “a throttle, a gas pedal. That’s somewhere we need to go.”

Contemporary models use springs to help the patient push off when taking a step. The spring stores energy when the foot initially steps down, and releases it at the ball of the foot when the heel leaves the ground. The result is a more natural step than is possible with a conventional prosthetic.

But prosthetic ankles can’t be passive springs if they are to work like the real thing, Herr notes. Joints like knees and ankles respond and anticipate actions: stiffening and relaxing muscles to handle inclines, obstacles or even flat ground different speeds. Artificial legs require a higher level of energy from wearers because of this reason, and Herr wants to lower their metabolic cost.

The artificial knees and ankles will be sown with a slew of sensors for situational awareness and feedback control. A wireless microchip called Bion, developed by the Alfred Mann Foundation, will provide connection between the nerve endings in the leg and the artificial knee and ankle. The chip is 1 cm long and 2 mm thick and is inserted into existing leg muscle.

Information must be fed back to other parts of the artificial leg, and ultimately will be driven by signals from the central nervous system. It must also be predictive, not just responding to the first step, but changing the leg to anticipate it.

“Bion is just one of many sensors,” Herr says, adding that information about the terrain and location of the prosthesis will come from additional sensors. Some will be in the heel and at tip of the prosthetic foot.

Installing and networking sensors in an artificial limb is difficult. It requires a researcher to match the pattern of electromyogram (EMG) signals to a behavior. EMG records the electrical activity of muscles. When muscles are active, they produce an electrical current that are proportional to the level of the activity. But matching electrical signals to real-life movements is complex since there are so many variables at work. In such systems, it is preferred to build an adaptive system that can “learn” how to handle instructions from the user’s nervous system, says PhD candidate Waleed Farahat, who works in Herr’s biomechatronics group at MIT.

Using frog muscle hooked up to an electrical current, sensors and a small machine that slides a small metal bar against the muscle, Farahat can simulate the muscle’s reaction to a hard wall, soft ground, and a variety of environments, as dictated by the machine’s software. The piece of frog muscle, about the size of a fingernail, twitches and jumps. “We’ll use this information to create data-based models,” Farahat says.

To this end, the VA program is turning to a system called Braingate that was developed by Cyberkinetics Neurotechnology Systems Inc. Brain signals are translated into motion by relaying impulses from a sensor that is implanted in the motor cortex. The sensor is connected to two sets of neural processors by wires poked through the skull.

Braingate is being tested currently in a 25-year-old quadriplegic, who can switch on lights, change television channels and open email with mental commands.

In order for the system to work on artificial legs, officials at Brown say, the system needs to be smaller. It will feed brain signals via fiber-optic cable to a pacemaker-sized processor in the chest. Neural decoding is the other hurdle—making sure the electrical signals inspire the right motion, in real time. Another big challenge is making the information flow between brain and machine.

Many players in the VA effort also have taken part in other defense programs. Advanced prosthetics technology, mostly focusing on nervous system-machine connectivity, is now receiving a fresh focus at the Defense Advanced Research Projects Agency. The VA Center’s staff hosted DARPA in February to brief officials on their topics of study.

“From their point-of-view, they wonder why we have to limit limbs to just restore function,” notes Aaron. “Why not be stronger?”

The research agency is doing its own outreach on the topic of helping amputees, and building a foundation for possible future augmentation projects. DARPA has only one such effort, the Exoskeleton for Human Augmentation project. That project does not integrate robotics in the human body, but rather seeks to develop a wearable suit to make soldiers stronger.

In late January, DARPA conducted a prosthetics conference to discuss the best way emerging technologies could be used to advance prosthetics. DARPA brought together VA officials, National Institutes of Health staff, prosthetics industry representatives, doctors, surgeons and robotic engineers.

The meeting is one of the early signs that DARPA is moving past its Human Assisted Neural Devices (HAND) program, which investigated direct interfaces between man and machine.

The agency is currently planning an advanced prosthetics program that will use the techniques pioneered through the HAND program for mental control of prosthetics.

According to DARPA, the new prosthetics program will support research in microelectronics, control systems, materials and power. “Neural interface technology will be refined and incorporated into prosthetic applications to create a system of control that goes beyond archaic pulleys and springs,” a DARPA document states. “Through this program, amputees will eventually have prosthetics that mimic all the functions of their natural limbs.”

HAND featured programs at Duke University, where researchers wired Macaque monkeys’ motor cortexes to control computer cursors with their minds. By correlating the neural activity of the monkeys to joystick movements, the team was able to unplug the joystick, and let the monkeys control the cursor with their brain’s motor signals. The primates now are being trained to manipulate a robotic arm to receive a reward.

Another team is designing algorithms that will improve medical methods of measuring brain waves with electricity and magnetism that are called electro-encephalograms and magneto-encephalogram, respectively.

And doctors at the University of Southern California and Wake Forest University are examining the function of the hippocampus of rats as they learn complex tasks, and transferring brain signals onto a microchip. Research on this front could provide sensory feedback to users of advanced prosthetics in the future or to restore learning to damaged brains.

Back To Top