November 20, 2003
"Biomedical engineers and surgeons have attached a bionic arm that can be controlled by thought."
ARLINGTON, Va., Oct. 27, 2003 – Biomedical engineers and surgeons at the Rehabilitation Institute of Chicago have attached a bionic arm that can be controlled by thought.
Because the patient lost both arms to his shoulders, doctors grafted existing nerve endings from one of his shoulders onto his chest muscle. After those nerves took root, electrodes placed on the skin over the graft could pick up the electrical impulses generated when the patient thought about moving his arm. Those impulses animate the artificial arm.
“Improving the control of artificial arms remains a considerable challenge, especially when most or all of the arm is removed,” says Todd Kuiken, M.D., Ph.D., a doctor and bioengineer leading the project. He explains that the nerve-muscle grafts can produce independent electromyographic (EMG) control signals that can simultaneously control different degrees of motion in the arm and hand, including bending the elbow, rotating the wrist, and opening and closing the hand.
After amputation, the nerves that conducted control signals from the brain to the limb before the amputation remain healthy and intact near the site of the amputation, says Kuiken. But because there is no limb, there is nowhere for the signals to go. His bionic arm, the first powered prosthesis to use a nerve-muscle graft, taps into these lost signals to greatly improve the patient’s control. This gives the nerve-muscle graft system its main advantage, he says, because the electrical signals are more directly related to the natural signals used to control the original arm, thus reducing the conscious effort required by the amputee and making the prosthesis seem easier and more natural to control.
If the amputee thinks, “close hand,” the neural control signal travels from the brain down to the nerve graft, which is now rooted on the chest muscle, and causes a part of the muscle to contract. Surface electrodes stuck on the patients chest pick up the EMG from the contracting chest muscle and send a control signal via wires to the powered prosthetic, telling it to close the hand. If the amputee thinks, “bend elbow,” the neural control signal from the brain to the nerve graft would instead contract a different area of the chest muscle, creating a slightly different EMG telling the arm to bend at the elbow.
A key part of the nerve-muscle graft concept is being able to record the EMG signals from the different nerve-muscle grafts independently, says Kuiken. If there is too much cross-talk or blending of signals, then the prosthesis will not work correctly. Although recording control signals directly from the remaining nerves without grafting them to a muscle might obtain clearer, more distinct signals, that would require permanent openings in the patient’s skin to allow wires through. The removable chest electrodes in Kuiken’s system eliminate any risk of infection and allow the patient to easily remove or put on the artificial arm.
A Whitaker Foundation Biomedical Engineering Research Grant supported Kuiken’s work to study how EMG signals propagate in the upper arm. He developed “some of the most sophisticated computer models ever made” for tracking EMG signals in arms. “We have shown what the effect of skin, fat and bone has on the EMG signal. We have analyzed different types of electrodes, how far apart recording sites need to be and what the pick up range of the electrodes should be,” he says. “The information gained from this work has allowed us to move forward with confidence in the clinical application of the nerve-muscle graft technique.”