A powered prosthetic leg that predicts when the wearer is about to take steps on flat or inclined surfaces, or climb stairs, helped improve prosthesis control in a small new study.
The control system, which uses electrodes in the cuff of the prosthesis to pick up signals from muscle contractions, is still in the testing stages but should be available on the market in a few more years, according to lead author Levi J. Hargrove of the Center for Bionic Medicine at the Rehabilitation Institute of Chicago.
His team is trying to allow people who use the artificial legs "to move exactly how they were moving previously,” Hargrove told Reuters Health.
“They have the potential to make it easier for people to walk, and use less energy when walking,” he said. “They also allow for more natural walking patterns, and have the potential to recover from trips or stumbles.”
About 115,000 people in the U.S. had major lower limb amputation due to trauma or cancer in 2005, and most prosthetic lower limbs are still “passive” and do not provide power, he and his coauthors write in JAMA.
But, they add, powered limbs are becoming available, and the new control system helps the motorized prosthetic device predict whether the wearer needs to walk on level ground, on a ramp, or up or down stairs.
Currently, people who use powered lower limbs need to slow down, stop and press buttons to switch between flat surfaces, ramps and stairs.
The new intuitive control system uses signals generated during muscle contractions - called electromyographic (EMG) signals - to predict movement.
For the new study, the researchers recruited seven people with a lower limb amputated at or above the knee who were currently able to use a passive prosthetic to get around at home. The participants wore EMG electrodes on the remaining muscles of the leg. They also wore a mechanical limb fitted with sensors of its own.
They wore the limbs while completing 20 walking and climbing trials.
While the participants moved, two computer algorithms predicted their steps: one used only the mechanical data from the prosthetic limb and the other used both mechanical data and EMG data from muscle contractions.
They found the algorithm combining the mechanical and EMG data reduced the proportion of incorrect predictions from 6 percent to 3 percent.
In a real-time trial, where the programs actually controlled the movement of the prosthetic instead of just predicting what it would do next, the EMG data reduced the error rate from 14 percent of steps to about 8 percent.
Some errors weren't noticeable, but some caused moderate impact, like stubbing a toe. Less than 1 percent of errors caused substantial impacts, which would have resulted in a trip or fall, Hargrove said.
It is important to eliminate those substantial impacts and reduce the other errors as much as possible, he said.
This technology is already available in powered arm prostheses.
While it's easier to predict leg movements compared to arm movements, Hargrove said the technology is more difficult to incorporate into lower limb prostheses, since walking around is a critical function in life and a mistake can lead to falls or injuries.
“Over the next 5 to 10 years, I expect powered devices will start to emerge more fully, and I suspect (like the smartphone) people will find that they offer considerably more functionality and freedom than previously available devices,” said Michael Goldfarb, an expert in mechanical and electrical engineering and computer science at Vanderbilt University in Nashville, Tennessee.
The powered prosthetics will be slightly more expensive than passive devices, but the benefits will justify the increase in cost, Goldfarb, who wasn't involved with the new study, told Reuters Health by email.
When they are available, the experience of wearing the new prostheses should be the same as that for artificial limbs now, and no more invasive, Hargrove said.
“I don’t think that we’re ever going to develop one prosthetic leg or technology optimized for all patients,” but this should be a great option for some, he said.