Robotic prosthetic taps spinal nerve signals
While the act of picking up an object is something most of us take for granted, for prostheses users, it can be an exercise in frustration. For all their promise, brain-controlled bionic arms, both invasive and non-invasive, are still not ready to leave the lab. Could a new robotic prosthetic arm that detects signals from spinal nerve cells nudge researchers closer to creating an artificial limb that resembles the real thing?
The problem with muscle-controlled robotic arms is that they rely on twitches from amputated limbs, the fibers of which are often damaged.
"When an arm is amputated the nerve fibres and muscles are also severed, which means that it is very difficult to get meaningful signals from them to operate a prosthetic," explains Dario Farina, a professor of bioengineering at Imperial College London.
This limits the number of tasks the artificial limb can perform and explains why up to 50 percent of users end up dumping them in frustration, he claims. But what if the prosthetic made use of the nervous system instead, so that the signals sent from the motor neurons, which control muscle movement, could be deciphered more clearly? Could this lead to the development of more intuitive robotic arms? This is what Farina and his team wanted to find out.
To test their hypothesis, they developed a sensor that uses the electrical signals sent from spinal motor neurons as commands and enlisted six amputees to test it. In order to amplify the signals and make it easier for the sensor to detect them, the subjects had the nerves associated with hand and arm movements surgically re-routed to healthy muscles in their chest or biceps.
Next, the researchers decoded and mapped the information in these signals, comparing them to those of healthy patients. The idea here is that by decoding the meaning behind all the signals sent from these motor neurons, they would eventually be able to create a full suite of commands for arm and hand functions in the prosthetic so it could function like the real deal.
For this study, the researchers encoded specific motor neuron signals as commands into the design of the prosthetic, after which they placed a sensor patch on the muscle that had been operated on, which was in turn connected to the prosthetic.
As the researchers reported in their study, the results were encouraging: after working with physiotherapists to learn how to control the device, which entailed picturing themselves controlling a phantom arm and imagining simple actions, the amputees were able to perform a wider range of movements with the prototype compared to those using classic muscle-controlled robotic prosthetics. These actions included lifting their arms up and down, and moving their wrist from side-to-side.
For some observers, what is novel about this study is that the computer algorithms could be used to control prosthetic limbs in future. Since the sensor is not embedded inside the body like an implant, there is no need for additional surgery either, unlike mind-controlled prostheses that require users to be set up with invasive brain implants.
"You don't need the sensor inside of the body to understand what the nerves are doing. This is the really innovative part of the study," says Levi Hargrove, a research scientist with the Rehabilitation Institute's Center for Bionic Medicine in Chicago, who was not involved in the study.
The quest to create artificial limbs that can put the brain's commands into action is one that spans decades. In the US alone, 185,000 amputations occur each year, with vascular disease being one of the major causes. War is another factor that adds to the statistic. As of 2012, the number of wounded US combatants in Afghanistan and Iraq included more than 1,500 amputees. Since 2006, DARPA has spent US$153 million on its Revolutionizing Prosthetics program but despite innovations such as the LUKE arm, there's still plenty of room for progress in this area.
Having achieved proof-of-concept with this study, the next step for Farina and his team is to employ the technology in a larger clinical trial and subject it to rigorous testing. If all goes well, the researchers estimate that it could be on the market within three years.
The study was published in Nature Biomedical Engineering.
Source: Imperial College London