Over the last few decades, we have seen robots created with astonishing capabilities—a humanoid robot that can walk up and down stairs, a robot that can rove on Mars, collect samples, and send images back to Earth, and robots that are used to help perform surgery. With all the amazing robotic technology that surrounds us, one might wonder why prosthetic limbs (used to replace human arms and legs) aren’t more like the real thing.
In fact, many amazing advancements in prosthetic technology have been made in recent years, with more currently underway. Unfortunately, prosthetic limbs still haven’t reached the level of Luke Skywalker’s flawless arm in Star Wars. Why? One reason is that we lack ways for people to naturally and intuitively tell their prosthetic limbs what to do.
In general, robots are either controlled through remote control or programmed to perform certain tasks at a given time. A prosthetic limb, however, cannot be preprogrammed— how could we know when someone would want to take a drink of water, or begin to walk up the stairs? And a remote control would take a great amount of concentration as well as the constant use of one or both hands.
There are two ways in which prosthetic arms are currently controlled. One way is to use other body motions, such as shrugging the shoulders, to pull on a series of cables and operate the hand, wrist, or elbow. These prostheses are called “body-powered” prostheses. The other way is to read the tiny electrical signals generated by muscles when they contract, and use these signals from remaining muscles to control a motorized arm.
These muscle signals are read by small electrical antennas called electrodes. For example, someone who has lost their arm above the elbow may use their biceps and triceps muscles to control their prosthetic hand. The prosthesis could be programmed to interpret signals from the biceps muscle as “open the hand” and signals from the triceps muscle as “close the hand.” These prostheses are called “myoelectric prostheses.”
Designers of myoelectric prostheses have an important problem to solve—there are many movements to control, and not enough remaining muscles to control them. Take the example above: if the biceps and triceps muscles are used to open and close the hand, what muscles can send the signals to bend and straighten the elbow? And what if the hand needs to rotate sideways in order to pick up an object—how will the user tell it to do this?
Researchers have tried many ways to get around these issues. They have implemented various types of switches, or instructed users to perform combinations of muscle contractions to act as a switch. However, the lack of available control signals makes it extremely difficult to move prosthetic limb technology forward.
The Neural Engineering Center for Artificial Limbs at the Rehabilitation Institute of Chicago is focused on this problem of prosthesis control. The lab centers around a surgical technique called targeted muscle reinnervation (TMR). The goal of TMR surgery is to utilize to the brain commands that still attempt to reach the missing limb. These commands travel down nerves that were severed during amputation and are no longer received by muscles. If these nerves are connected to different muscle sites, they can cause these other muscles to contract, producing the signals used to control myoelectric prostheses.
The first TMR patient was a man who lost both arms at the shoulder in an electrical accident. Without arms, his chest muscles were useless to him. So, on his left side, they cut the nerves which ran to the chest muscles and instead connected the nerves which had previously gone to his arm.
After a few months, attempts to move his missing left arm resulted in contractions of his chest muscles, with different contractions occurring for elbow bending, elbow straightening, hand opening and hand closing. Then, using an even larger number of surface electrodes to read the muscle contractions, researchers were eventually able to tell the difference between attempted movements of the elbow, wrist and hand, including different types of hand grasps. Not only did this help with the problem of control, letting the patient independently move different joints, but it allowed him to control his prosthesis intuitively. His attempts to move his missing arm now resulted in similar movements of the prosthesis.
Work is now underway to further refine myoelectric control. Rather than using individual muscle sites to control each movement, a computer can be trained to recognize different movements from the pattern of electrical signals recorded from all reinnervated muscles. It does not require the muscle signals to be perfectly isolated, and can adapt to extra signals, such as a heartbeat. In addition, the placement of electrodes is not required to be exact, as they would with conventional control, because the system is retrained each time the system is used. This technology, called “pattern recognition,” is currently only used as a research tool. However, it’s possible that it could one day be incorporated into prosthetic devices for even more precise control.
Further improvement of prosthetic technology may be possible because of an additional and unexpected result of TMR surgery. Several months after his procedure, the first TMR patient reported that being touched on his chest actually felt like he was being touched on his missing arm–the sensation nerves had grown into his chest skin. Further investigation has shown that patients are able to perceive different temperatures, sharpness of objects, vibrations, and pressures on their reinnervated skin, and that they feel like all of these stimuli are being applied to their missing limb.
Researchers have been able to “map” these perceptions, or see what locations on the reinnervated skin correspond to the pinky finger, palm, forearm, etc. Surprisingly, these maps appear to remain stable over time. By learning more about these perceptions, we hope to someday provide feedback from the prosthesis to the reinnervated skin that will allow users to “feel” with their artificial arm as if it were their own hand. This could go a long way in improving patients’ ability to control their prostheses, and even improve their quality of life.