A man missing his lower leg has gained precise control over a prosthetic limb just by thinking about moving it — all because his unused nerves were preserved during the amputation and rerouted to his thigh, where they can be used to communicate with a robotic leg.
The man can now seamlessly switch from walking on level ground to climbing stairs, and he can even kick a ball around. (See video.)
During a traditional limb amputation, the main sensory nerves are severed and lose their function. In 2006, Todd Kuiken and his colleagues at the Rehabilitation Institute of Chicago realized they could preserve some of that functionality by carefully rerouting sensory nerves during an amputation and attaching them to another part of the body.
They could then use the rerouted nerve signals to control a robotic limb, allowing a person to control their prosthesis with the same nerves they originally used to control their real limb.
Kuiken’s team has performed TMR for the first time on a man with a leg amputation.
First, the team rerouted the two main branches of the man’s sciatic nerve to muscles in the thigh above the amputation. One branch controls the calf and some foot muscles, the other controls the muscle running down the outside leg and some more foot muscles. After a few months, the man could control his thigh muscles by thinking about using his missing leg. The next step was to link up a prosthesis.
The robot leg in question is a sophisticated prosthesis: It carries a number of mechanical sensors including gyroscopes and accelerometers, and it can be trained to use the information from these sensors to engage in certain walking styles. Kuiken’s team figured that the leg would perform even better if it could infer the user’s intended walking style with information from the sciatic nerve.
To do so, the researchers asked their volunteer to attempt to perform certain movements with his missing leg — for instance, flexing the foot — while they monitored the pattern of electric signals from the rerouted nerves in the thigh muscles. The researchers then programmed the robot leg to flex its foot whenever it detected that particular pattern of electrical activity.
Using just the mechanical sensor data, the robotic leg made the correct movement about 87 percent of the time. With additional data from the nerves, the success rate rose to 98 percent.