Robotic Leg Users First Underestimate, Then Overestimate, Their Own Gait. Here’s Why It Matters.
Reading Time: 5 minutes
___
Summary:
-
NC State study finds robotic leg users consistently misjudge their own gait during training
-
Early users underestimate their walking ability; experienced users overestimate it
-
Perception gap persists even as actual lower limb amputation rehab performance improves
-
Users fixate on torso position and largely ignore prosthetic device behavior
-
Overconfidence after day four may slow progress in powered prosthetic leg training
___
When someone begins walking with a robotic prosthetic leg, their brain is working overtime not just to master new movement patterns but also to make sense of what those movements actually feel like. A new study from North Carolina State University reveals that this mental picture of one’s own gait is consistently wrong.
The research, published in the journal PNAS Nexus in February 2026, examined how people’s perception of their own movement changes as they learn to walk with a robotic lower-limb prosthesis.
The image above is a representation and does not illustrate the actual research discussed in this article.
A Perception Gap
Everyone carries a personal body image: an internalized sense of how their body is structured and how it moves through space. When learning a new physical skill, like dancing or a sport, a person’s mental image of their movement often diverges from what their body is actually doing. Over time, that mental image tends to align more closely with reality, and physical performance improves alongside it.
Researchers wanted to understand whether this same process plays out when the body being learned includes a robotic device, and whether incorporating that device into one’s body image affects how well someone learns to use it.
For the study, nine able-bodied participants were fitted with a robotic prosthetic leg attached at the knee, bent at a right angle, and tasked with walking on a treadmill as quickly as possible without holding the handrails. They trained with the device every day for four days. After each session, they were shown computer animations of a range of biomechanical walking gaits and asked to identify which most closely matched how they had just been moving.
Underestimating First, Overestimating Later
The pattern that emerged was consistent. According to Helen Huang, the study’s corresponding author, the participants initially felt that their gait was more off-balance and stilted than it actually was. By the end of the four-day study, participants believed their gait appeared more natural and fluid than it truly was.
In other words, the error didn’t simply shrink with practice—it flipped. Early-stage users sold themselves short. Later-stage users gave themselves too much credit. While performance did improve significantly across all participants over the four days, the gap between perceived and actual movement persisted throughout, just oriented in the opposite direction.
The Brain Focuses on the Torso
The researchers found that when participants assessed their own gait, they concentrated heavily on the position of their torso rather than on the behavior of the prosthetic device itself. Huang suggests this is likely because participants received very little direct feedback about how the device was actually moving. They couldn’t see themselves walk.
This finding points to a meaningful gap in how prosthetic training is currently structured. Without accurate, real-time information about what the device is doing, users are left to interpret their own movement through an incomplete internal lens.
The Overconfidence Problem
For lower-limb amputees in prosthetic rehabilitation, the implications go beyond curiosity. Huang noted that overconfidence in one’s movement skills can actually impede progress: if someone already believes they’re performing well, they’re less motivated to invest in the additional effort needed to improve, even when meaningful gains remain within reach.
This creates a subtle but tangible barrier in the rehabilitation process. A person who has trained with a powered prosthetic leg for several days may feel they’ve reached a functional plateau when, objectively, they haven’t. Without external calibration, that false sense of mastery can slow down what is often a long and demanding adaptation process.
The Case for Better Feedback Tools
Huang suggested improving training outcomes by providing users with visual or other forms of feedback to help them calibrate their body image while working with the prosthetic device.
The study is described as the first to examine this perception phenomenon specifically in people learning to use lower-limb robotic prosthetics, and the researchers see it as opening a set of questions that could improve how people train and adapt to these devices.
For the limb loss community, this research adds important nuance to how rehabilitation progress is understood and measured. The subjective experience of walking with a prosthetic leg doesn’t reliably match the objective performance, particularly in the early weeks of use. That mismatch, researchers suggest, is not a personal failing but a predictable feature of how the brain integrates a new device into its model of the body.
Future work will likely explore which kinds of real-time feedback—visual displays, haptic signals, or other sensory tools—might help users develop a more accurate sense of their own movement.
Related Reading:
Are Powered Prosthetic Legs Finally Ready for Everyday Use?
New Nerve-Reading Tech Could Make Prosthetic Legs Feel More Like the Real Thing
Neural-Controlled Prosthetic Leg Recreates Natural Walking Experience
