Just over 15 years ago, Johnny Matheny was diagnosed with fibrosarcoma, a rare form of cancer that attacked the connective tissue around the bones of his left hand. His doctors gave him a choice: Amputate the hand or die. He chose amputation. He could still feel his hand after it was gone, a so-called phantom limb caused by signals firing through the nerves that once connected his hand to his brain. Now fully in remission from the cancer, Matheny has spent years volunteering for experiments, test-driving the fanciest, most advanced prosthetics available. But while these devices offered high-tech motor control and vibration feedback, some things were still missing, including the ability to feel temperature.

Three years ago, researchers at the Johns Hopkins Applied Physics Laboratory (APL) began testing a new stimulator that would restore feelings of cold to people who have experienced limb loss, and Matheny became one of four volunteers. The device, a bandage-thin, dime-sized square that can be placed on the skin or embedded in wearable fabric, pumps heat from one side to the other to create a cooling sensation. As part of the lab’s experiments, Matheny wore a bionic arm equipped with a temperature sensor and the new stimulator, then was asked to grab a series of soda cans. He remembers picking one up and chuckling, wide-eyed and beaming: It was cold. “I hadn’t felt stuff like that since before they took my arm off,” he says. “It was just unreal.”

Although state-of-the-art prosthetic limbs can already provide fine motor control and basic sensory feedback, the nuances of touch that make people feel like their limbs are truly their own—like temperature—have remained out of reach. Now, in a study published in July in Nature Biomedical Engineering, the APL researchers showed that their cooling patch could give people their perception of temperature back. It offers hope that soon, artificial hands may allow users to enjoy a cold beer or the warm touch of a loved one.

Advanced bionic hands are getting better at moving, but “we’ve long recognized the need to provide sensory feedback as well,” says Doug Weber, a professor of mechanical engineering and neuroscience at Carnegie Mellon University who was not involved in this study. “It’s not only important to be able to make things move but to feel the consequences of those actions as well.” Without sensory feedback, it’s impossible to tell where your hand is without looking at it, or whether you’re about to crush something in your grasp.

But touch is complicated. “It’s more than just pressure or force on your fingertip,” says Luke Osborn, a neuroengineering researcher at APL and the paper’s lead author. “It encompasses all of these other complex sensations, such as temperature.”

Temperature perception seems to have special meaning in our social lives, says Emily Graczyk, a biomedical engineering professor at Case Western Reserve University who was not involved in this study. In her research, Graczyk found that providing sensory feedback helps prosthesis wearers feel more confident and comfortable interacting with others and makes their artificial limbs feel more like a part of themselves. “Nothing would make your prosthesis more human than if it actually felt the warmth of someone’s grasp,” Weber adds.

In an amputation, the nerves that once relayed information between the brain and the limb are severed. But the ends of these nerves can regenerate. As they regrow, they innervate whatever tissue they can latch onto, like the skin on the residual limb. For a person with an upper limb amputation, that might be the area just above the wrist or the elbow. Zapping that skin with electricity can actually feel like a zap to the hand. This pathway, from the residual arm to the brain, serves as a “window to deliver inputs,” Weber says.

“Our traditional approach for restoring sensations for amputees is electrical activation,” stimulating a patch of skin with a small electrode, says Osborn. Nerve fibers that respond to mechanical aspects of touch, like pressure and vibration, are fat and insulated by a myelin sheath that keeps current from leaking out, making them easy to activate. But nerves that carry information related to temperature are tiny and don’t usually respond to electrical stimulation. Graczyk says that the best way to make someone feel temperature is the old-fashioned way: using a hot or cold thing to activate the full range of skin receptors.

For prosthesis wearers to feel a chill, something must transmit an ultrafast temperature signal (in under half a second) with sub-centimeter precision from the prosthesis to the skin to activate the nerves that once corresponded to the person’s fingers. Rama Venkatasubramanian, the chief technologist for thermoelectrics at the Applied Physics Lab, was up for the challenge. He has spent the past 25 years developing thermoelectric cooling devices for infrared sensors and satellites, but he was especially excited to build something that could deliver cooling sensations to people. “Nothing compares to the idea of enabling human capability,” he says.

You can buy a thermoelectric cooling device online right now, if you’d like. They’re great for preventing gamer PCs from overheating but are too hefty and too slow to mimic speedy biological processes or to wear around all day. Venkatasubramanian wanted to design something quick, noninvasive, and lightweight. So he built a tiny device out of super-thin films (about 20 to 25 microns thick, less than half the width of a human hair), which use electrons to pump heat from one layer to the next, leaving coolness behind. It’s like a tiny refrigerator with remarkable power and neuron-like speed. (Imagine a semiconductor with the heat flux levels of a rocket nozzle, while the other end is kept cool, Venkatasubramanian says.) After confirming it via many benchtop experiments, he says, “this is the world’s fastest, most efficient and intense refrigeration near room temperature.” 

Then the team recruited four people with upper limb amputations to test Venkatasubramanian’s new device and compare it with an existing alternative. First, they used a larger commercially available cooling device to map out the spots on each person’s residual limb where a cold touch to the skin caused a chill somewhere on their phantom hand. Then, each participant placed the commercial cooling device on one of these spots while the experimenter cooled it from room temperature to 16 degrees Celsius (about 61 Fahrenheit). Each volunteer pressed a button as soon as they felt a temperature shift, then used a sliding scale to report how strong the shift felt.

Next, they repeated this part of the test with Venkatasubramanian’s newer, tinier device. It delivered cooling sensations about four times faster than a standard thermoelectric device, and they felt roughly twice as strong. Two people without amputations repeated the experiment, holding the new device to a fingertip, and felt the same thing: faster, stronger cooling.

Matheny helped the researchers take the experiment a step further. This time, instead of testing the stimulator on bare skin, they wanted to try sending cooling sensations from the fingertip of a prosthetic all the way to Matheny’s own nervous system. They used an advanced prosthesis developed by researchers at APL—which Matheny took for a yearlong real-world test drive in 2018—and embedded a temperature sensor in the pinky finger. This fed a signal to the stimulator Matheny was wearing on the patch of skin on his arm that received sensations from his phantom pinky finger.

In 19 trials, Matheny felt three identical soda cans: two at room temperature and one straight out of the fridge. He was able to identify the cold soda every time. “It felt like my own fingers were feeling it,” he remembers. At the end of one experimental trial caught on film, Matheny chuckles softly as he correctly points to the cold one, cracks it open, and toasts the camera. “I just loved it,” he recalls.

At least in theory, such a tiny, noninvasive device would be easy enough to embed in the fabric liner worn under the prosthetic, as long as it’s fitted tightly enough that it hits the right patches of skin every time. And since no surgery is required, the process for gaining US Food and Drug Administration approval as a prosthesis add-on will likely be less arduous than for other, more invasive medical devices.

Along with vibration feedback and fine motor control, temperature sensors could help artificial hands get closer to fully re-creating touch. Matheny’s dream prosthesis would match his remaining arm in all but one sense: “I know the one thing I don’t want to feel is pain,” he says. Although Osborn’s team has made bionic arms feel pain before, for ethical reasons researchers typically steer clear of causing it unless it’s the topic being studied. But Graczyk suspects that pain may be, to some degree, a necessary part of touch. “We know where our body ends because of the sense of pain,” she says. And to feel like a prosthesis is actually part of the wearer—not an inanimate object they’re lugging around—they may need to feel the boundaries that pain provides. That feeling of embodiment, Graczyk says, “can be desirable to people. It helps them feel like they are whole again.”

Sensing temperature may not only help people feel at home in their own bodies but also create a deeper sense of social connection. When receiving a tender arm squeeze or a pat on the back, “being able to actually feel somebody’s body temperature is actually a big part of it,” says Graczyk. “We feel things as more pleasant and more comforting when they’re at body temperature.”

Adding temperature sensation to existing devices is a step in the right direction, Osborn says. “Now, we can start pushing the boundaries of what we thought we could do with prosthetic arms,” he says. “We can give somebody the ability to accomplish something that they couldn’t do before.”