New Bionic Arm Lets Amputees Feel Touch, Retrofits to Existing Prosthetics

We’ve come a long, long way from hooks and peg legs: Scientists from the Lerner Research Institute at the Cleveland Clinic have developed bionic arms that can directly interface with the motor and sensory neurons of amputees, affording a truly unprecedented degree of control and sensory fidelity. Even cooler, they did it with commercially available hardware retrofitted to existing prosthetics. That’s important in a field where new technological developments are often accompanied by large price increases for already expensive prosthetics.

Using existing COTS (Commercially Available Off-the-Shelf) components simplifies issues related to cost and availability but creates problems of its own. Trying to get good DEX stats out of a conventional electromechanical prosthetic arm is like playing QWOP or Hand Simulator. It’s a hypothetically manageable control problem that turns into real labor when you try to do it at faster speeds. (For those of you who haven’t played QWOP or Hand Simulator, these are games in which fighting managing the control scheme creates much of the game’s difficulty.)

Imagine you had to use the Hand Simulator control scheme to control the movement of your own arm and hand, but instead of holding the controller with two hands, you have to run it with one elbow. It takes conscious thought and attention, and it requires the person to learn a whole new and different way to coordinate the movements of their new QWOP hand. Better than nothing, by a long shot — but there’s also definitely room for improvement.

When a limb is amputated, the axons of the person’s sensory and motor neurons are severed, but their upstream parts are still there, and the nerves don’t get “unplugged” from the brain. This is part of why some amputees experience phantom limb pain. It’s also why mirror therapy works for some of those folks.

The researchers in this study retrofitted off-the-shelf force sensors into a prosthetic limb compatible with a surgical procedure known as “targeted reinnervation.” In targeted reinnevervation, surgeons take the severed ends of the patient’s existing sensory or motor neurons that would have governed the missing limb, and connect them to the axons of sacrificial sensory neurons that serve intact places on the patient’s skin, or intact muscles elsewhere in the patient’s body. This creates spots on the skin that the brain believes belong to the missing parts of the patient’s amputated limb.

Creating these spots does involve a tradeoff because scientists are essentially disconnecting one neuron and hooking up another of the same type. Signals that correspond to sensations on a person’s fingers, hand, and wrist are transferred to the patch of skin, while the person also loses a small amount of their original input/fidelity in the same physical area. This is generally seen as a worthwhile tradeoff because the benefit of being able to control an artificial limb with something like direct feedback is much more useful than a slightly more sensitive bit of forearm.

The COTS tech behind the bionic arm itself is straightforward. With the motor and sensory nerves reconnected and healed up, the team stuck a listening electrode on the motor neuron and put a tactor buzzer (just like the ones in VR haptics suits) against the new sensory patch to deliver tactile feedback. When participants moved to grasp objects, the degree of hand opening or closure on the bionic arm was reported in part through the use of a thing called a Bowden cable, which mechanically translates linear motion into rotational motion.

To quantify exactly how much of an improvement their build produced, the researchers tested the prostheses for detail discrimination by having participants fish around inside a container to get out objects of the right tactile stiffness (think baseball versus stress ball versus shower pouf). Test participants wore frosted goggles and noise-proofing earbuds to prevent them from using other external cues to determine which object they were holding.

The two participants in the trial regained varying levels of functionality, in part because of the nature of their amputations. But after getting rigged up with the new bionic arm, one patient consistently did well enough on grip dexterity indices to classify into the “able-bodied” cohort on mechanical limb function tests. This doesn’t mean the artificial limb actually performed identically to a human arm or hand. But in this specific case, the individual in question was able to accurately pick objects of the right firmness out from being mixed in with a bunch of distractors. She was also able to pick up a pasta box with the right amount of force, and then look at where she intended to put it down rather than following the artificial hand with her eyes at all times. This makes for a much more fluid, intuitive experience of movement: more flow, less QWOP.

Interfacing the human nervous system with artificial components is the very definition of a cyborg, and bionic arms with kinesthetic feedback thanks to surgical reinnervation of the interface site are a great example of how robot augments can help people in real life. But while these retrofitted bionics are already workable for mere mortals today, real-life cyborgs with major implants, à la Neuromancer or The Matrix, are a tough lift for scientists at the moment. The workability of “hardwiring” anything into the nervous system is somewhat dependent on whether its design breaks the skin or not.

If an indwelling object can heal back up and let the skin close on itself, that’s easier on the body’s inflammation response; transdermal implants are also at greater risk from being knocked around during daily life. It’s more likely that removable prostheses or augments will work more like what we’ve seen here: contact electrodes or inductive/mag-safe chargers pose less risk than the surgically inset transdermal SIM trays and augments of William Gibson’s dystopian timeline. Until that version of the future arrives, incremental advances like this can restore some normalcy to the lives of amputees, one delicate touch at a time.

Continue reading

No Flying Cars Yet, But How About a $300 Toaster With a Touch Screen?

As 2020 draws to a close, there's still no word on flying cars, but don't worry: We found something even better. For a certain definition of the word "better."

PlayStation 5 Hits Game Data Compression Ratios the Xbox Doesn’t Touch

Sony's PlayStation 5 offers much smaller downloads for some games compared with the Xbox Series S|X. We dive into the compression technology behind the company's achievements.

MIT Develops Inflatable Bionic Hand That Senses Touch

Medical prosthetics have come a long way but can cost many thousands of dollars, and they're heavy, rigid, and prone to mechanical failures. Researchers have created a prototype prosthetic hand that's the opposite: light, soft, and potentially very, very cheap.

Microsoft Says 20 Percent of Xbox Cloud Gaming Subscribers Use Touch Controls Only

Microsoft has been pushing to add full touch controls to as many Xbox games as possible, and that approach seems to be paying off. According to a new Xbox blog post, some 20 percent of Xbox Cloud subscribers are only using touch controls to play games, foregoing controllers altogether.