New Nerve-Reading Tech Could Make Prosthetic Legs Feel More Like the Real Thing
Reading Time: 6 minutes
___
Summary:
-
Researchers decoded leg movement intent directly from lower limb amputee peripheral nerves for the first time
-
AI-powered spiking neural networks interpret nerve signals the same way biological neurons communicate
-
Single bidirectional implant delivers both prosthetic motor control and sensory feedback; no multiple devices needed
-
Breakthrough targets above-knee amputees, where muscle-based prosthetic control is not possible
-
Next step: integrating neural decoding technology into a functional prosthetic leg for real-world testing
___
For people living with above-knee limb loss, prosthetic legs have long operated on borrowed signals, such as mechanical systems and built-in sensors that adapt to walking automatically, without any real input from the user. The leg moves, but the person wearing it isn’t truly in control. Now, a landmark study from researchers at Chalmers University of Technology in Sweden is changing what that relationship could look like.
For the first time, scientists have successfully decoded leg movement intentions directly from the peripheral nerves of people with above-knee limb loss, including something as precise as the will to wiggle a toe. The findings, published in Nature Communications in March 2026, represent a proof-of-concept breakthrough that could one day lead to prosthetic legs that respond to the brain’s own commands and return real sensation to the user.
Why Prosthetic Legs Have Lagged Behind Arm Prosthetics
To understand why this matters, it helps to understand the limitation it addresses. Arm and hand prostheses have made significant progress by tapping into residual muscle activity—muscles that are still present and receive nerve signals from the brain, even after amputation. The prosthesis reads those muscle signals and translates them into movement.
That approach doesn’t work for most leg amputees. In above-knee amputations, especially, the muscles needed to relay those signals are often gone. As a result, prosthetic legs have largely remained passive devices—smart in some ways, but not truly connected to the user’s nervous system.
Giacomo Valle, an assistant professor at Chalmers and one of the senior authors of the study, explained that when the body is instructed to move, signals travel through the nerves to the muscles that execute the action, even if the limb is no longer present. This indicates that all necessary information resides within those nerves, with the primary challenge being the extraction and understanding of the neural code behind it.
That challenge has stumped researchers for years. Nerve signals in the residual limb are faint and notoriously difficult to capture reliably. Prior attempts to read from peripheral nerves directly have been rare, and previous success involved upper-limb amputations. No one had cracked it for the leg—until now.
Reading Movement Signals Directly from Residual Nerves
The Chalmers team tackled the problem with two key innovations working in tandem: a specialized neural implant and an AI algorithm designed to speak the nervous system’s language.
The implant itself (developed at the University of Freiburg) consists of ultrathin electrodes, each roughly the width of a human hair, and is flexible enough to sit within nerve tissue without causing damage. Four of these electrodes were inserted into the tibial branch of the sciatic nerve, the nerve central to both leg movement and sensation, in two study participants. Previously, this type of implant had only been used to stimulate nerves, which delivered sensation back to the user. This study was the first to use it to precisely read outgoing signals from the nerve.
Once those signals were captured, the researchers needed a way to interpret them. That’s where the AI came in. The team used what are called Spiking Neural Networks (SNNs). Unlike standard AI, which processes continuous numerical data, SNNs work with discrete electrical impulses, or “spikes,” which are the same type of signals biological neurons use to communicate.
Elisa Donati, a professor at the University of Zurich and ETH Zürich and the study’s other senior author, explained that peripheral nerves transmit information via discrete electrical impulses, making spiking neural networks naturally suited to processing such signals. She added that by aligning computational models more closely with biological systems, it is possible to efficiently extract movement intent using compact models and relatively limited data.
The result was a system that could interpret specific, detailed movements, like knee bends, ankle rotations, and toe flexion, with high accuracy, even in participants whose amputation had removed most of the leg.
Valle remarked that it was astonishing to observe how electrodes positioned high on the residual leg could interpret attempts to move the toes.
One Implant for Motor Control and Sensory Feedback
Perhaps the most consequential aspect of the research is what the implant can do beyond just reading movement intent. The same device used to decode motor signals can also deliver sensory feedback, meaning a single implant could, in theory, handle both functions that current neural prosthetics require separate devices to achieve.
Valle explained that the system is bidirectional. Once electrodes are implanted inside the nerve, they can communicate with the nervous system in both directions. This allows a single neurotechnology to provide both natural neural control and sensory feedback within the same implantable device.
For people living with limb loss, that bidirectionality is significant. Sensory feedback, such as knowing where your foot is, feeling pressure underfoot, and sensing the texture of the ground, is something most prosthetic users live without. Its absence affects balance and the sense that the prosthesis is truly part of the body. A device that can restore both control and sensation through a single surgical implant would be a major step forward.
From Lab Proof to Real Prosthetic Leg: What’s Next
The researchers are clear that this study is a proof-of-concept. The experiments were conducted outside of a functioning prosthesis, where the participants attempted movements, and the signals were recorded and decoded without connecting to an actual prosthetic leg. The critical next step is integrating this neural decoding system into an actual prosthetic device and testing it in real-world conditions.
The research team—which includes researchers from Chalmers University of Technology, the University of Zurich and ETH Zürich, the University of Belgrade, the University of Freiburg, and the Medical University of Vienna—also believes the approach could eventually extend beyond leg prostheses to other types of limb replacements.
For the amputee community, particularly the many people living with above-knee limb loss for whom truly intuitive prosthetic control has remained out of reach, the implications are meaningful. This research doesn’t promise a product yet. But it does something equally important: it demonstrates that the nervous system’s own signals can be read, understood, and acted upon—even from a limb that’s no longer there.
Related Reading:
