SoylentNews

takyon writes:

Medicine has progressed a lot since the Civil War, but amputations haven't. Once a limb is sliced off, surgeons wrap muscle around the raw end, bury nerve endings, and often attach a fixed prosthesis that is nowhere near as agile as the flesh-and-blood original. Better robotic limbs are available, but engineers are still figuring out how to attach them to people and give users fine motor control. Now, a team of researchers and clinicians has developed a simple surgical technique that could lead to prosthetics that are almost as responsive as real limbs.

[...] The biggest barrier to lifelike limbs is that signals can no longer travel in an unbroken path from the brain to the limb and back. Scientists have developed several ways to bridge the gap. The simplest is to place electrodes on remaining muscle near the amputation site. For finer control, doctors can use severed nerves themselves to relay the signals, through electronic attachments. But when they aren't rejected by nerve tissue, such attachments tend to receive weak signals. A stronger signal comes from attaching nerve endings to small muscle grafts that amplify the signal and relay it using electrodes. But even this method fails to take advantage of a simple biological solution for joint control: the pairing of agonistic and antagonistic muscles. When you contract your biceps to bend your elbow, for example, your triceps on the other side of the joint stretches, providing resistance and feedback. Together, such opposing muscle pairs let you fluidly adjust a limb's force, position, and speed.

The new technique, developed at the Massachusetts Institute of Technology (MIT) in Cambridge, creates such a pairing for prosthetic joint control. It respects "the fundamental motor unit in biology, two muscles acting in opposition," says Hugh Herr, a biophysicist at MIT and co-developer of the method.

Source: http://www.sciencemag.org/news/2017/05/new-surgical-procedure-could-lead-lifelike-prosthetic-limbs

On prosthetic control: A regenerative agonist-antagonist myoneural interface (DOI: 10.1126/scirobotics.aan2971) (DX)

Original Submission