When someone loses a hand or leg, they don’t just lose the ability to grab objects or walk—they lose the ability to touch and sense their surroundings. Prosthetics can restore some motor control, but they typically can’t restore sensation. Now, a preliminary studyposted to the preprint server bioRxiv this month—shows that by mimicking the activity of nerves, a device implanted in the remaining part of the leg helps amputees “feel” as they walk, allowing them to move faster and with greater confidence.
“It's a really elegant study,” says Jacob George, neuroengineer at the University of Utah who was not involved with the research. Because the experiments go from a computational model to an animal model and then, finally humans, he says, “This work is really impactful, because it's one of the first studies that's done in a holistic way.”
Patients with prosthetics often have a hard time adapting. One big issue is that they can’t accurately control the device because they can’t feel the pressure that they’re exerting on an object. Hand and arm amputees, for example, are more prone to drop or break things. As a result, some amputees refuse to use such prosthetics.
In the past few years, researchers have been working on prosthetic limbs that provide more natural sensory feedback both to help control the device better and give them back a sense of agency over their robotic limb. In a critical study in 2019, George and his team showed that so-called biomimetic feedback, sensory information that aims to resemble the natural signals that occur with touch, allowed a patient who’d lost his hand to more precisely grip fragile objects such as eggs and grapes.
But such studies have been limited to single patients. They’ve also left many questions unanswered about how exactly this feedback helps with motor control and improves the use of the prosthetic.
So in the new work, researchers used a computer model that re-creates how nerves in the foot respond to different inputs, such as feeling pressure. The goal was to create natural patterns of neural activity that might occur when sensing something with the foot or walking.
The scientists first tested these patterns in experiments with cats. They used the model to deliver patterns of electrical pulses to peripheral nerves in the animals’ feet while recording the activity of spinal cord neurons that receive input from those nerves. As they hoped, the spinal neurons’ responses to biomimetic stimulation from the model looked similar to the cats’ response to actual touch—such as swabbing the cat’s paw with a cotton swab.
To test the approach in humans, the researchers implanted a device into the legs of three individuals with lower limb amputations. The device included a sensor that records pressure information from the robotic foot during walking and electrodes implanted into the peripheral nerves of the leg. It effectively “senses” the walking sensations and transmits that information into the amputee’s nervous system to try to mimic the sensations associated with walking.
To see which type of stimulation felt more natural, the researchers applied either the biomimetic stimulation from the model or a steady stream of electric pulses. Participants rated the biomimetic stimulation as feeling more natural. When they got the steady stream of pulses, “they would say things like ‘I felt like my leg was plugged into the electricity’, which we, of course, want to avoid,” says Stanisa Raspopovic, a neuroengineer at ETH Zürich and co-author of the study.
The amputees were able to walk up and down stairs faster with the biomimetic stimulation than with the constant stimulation, being able to complete about half a lap more of the circuit in each session. When asked after the task, participants said they felt more confident when walking with the natural stimulation. Finally, when the researchers asked the volunteers to spell a five-letter word backward while they were walking, their spelling was about 20% more accurate during biomimetic stimulation. That may be because attention is a limited resource, and having a more natural feeling prosthetic could open up the mind to attend to other things, says Gregory Clark, a neuroengineer at the University of Utah who was not involved in the research.
Raspopovic hopes his group and others will continue improving on this technology. George notes that current hand prostheses can only target a couple hundred of the more than 10,000 nerves in the hand, for example, limiting what they can do. “Imagine you're this brilliant painter, and you have this amazing vision, but the only thing you have to draw is black and white paint.”
A more detailed understanding of how the nervous system detects and communicates the different aspects of touch could help further refine such devices. Sensations such as pressure, pain, and temperature could help the researchers create a sensory experience that more closely resembles reality and eventually feels identical to the lost limb, for example, Clark says. “That's the idea—you want people to feel whole again.”
