The human brain responds to the movements of others as if it were preparing to reproduce them itself. When we see someone reaching for a cup or curling their fingers, a response arises in our motor cortex — as though our body were trying the sight on for size. This phenomenon has been studied for decades, but a recent study reveals its unexpected complexity.
The concept of mirror neurons grew out of precisely such observations. These brain cells got their name thanks to their dual activity: they fire both during one's own action and when observing a similar action performed by another person. They were discovered in the 1990s by the Italian neuroscientist Giacomo Rizzolatti while studying the brains of macaques, when scientists noticed that neurons in the monkeys' premotor cortex activated both when a movement was performed and when it was observed being performed by a human.
But a new paper, published on 21 September 2026 in the journal Proceedings of the National Academy of Sciences, offers a more nuanced vision. Perhaps it is not only about individual specialized cells that mechanically "mirror" a movement. Instead, the process involves a distributed network of the motor cortex — an entire population of neurons whose activity scales depending on how human the brain perceives the one performing the action to be.
Researchers from Brown University, the Center for Neurotechnology and Neurorecovery at Mass General Brigham, and the VA Rehabilitation Research and Development Center gained a unique opportunity — direct observation of the activity of individual neurons in the human motor cortex. The study involved two people with tetraplegia who had previously, as part of the clinical BrainGate brain–computer interface program, had microelectrode arrays implanted in the region of the motor cortex responsible for controlling the hand.
Participants were shown the same grasping movement, but it was performed by different virtual "agents": a realistic human hand, an anthropomorphic robotic hand, a three-fingered mechanical device, a simple cube, and abstract shapes. The trajectory of the movement remained the same — only who performed it changed. And along with that, the brain's response changed too.
The strongest response arose when observing the realistic human hand. The further the image departed from a human appearance — a robot, a mechanism, a cube — the weaker the activity in the motor cortex became. This discovery reveals a key detail: the motor cortex does not respond merely to the shape and speed of movement. What matters to it is the morphology of the agent — how close its appearance is to the human body.
But this was only the beginning. The researchers discovered another critical feature. The effect is not concentrated in one place in the brain and is not tied to the behavior of a single type of neuron. On the contrary, it is distributed across the entire recorded neural population. The same groups of cells flexibly changed in the intensity of their response depending on the anthropomorphism of the moving object. This is a more flexible picture than the traditional image of a "mirror neuron" that fires like a switch at the sight of a familiar action. Here the motor cortex works more like a fine-tuning system — the more easily the brain recognizes a human action in the movement it sees, the more vividly its own motor system responds.
A particularly intriguing finding came in the second part of the experiment. Participants were shown movements composed only of dots — so-called point-light displays. Some easily resolved into the image of a moving hand, while others looked like an abstract cloud in motion.
Here is what happened: one of the participants at first perceived such a stimulus simply as chaotic moving dots. Then, at some moment, his consciousness seemed to click — and he suddenly recognized in them the movement of a human hand. The moment he noticed it, the activity of his motor cortex while viewing those very same images sharply intensified.
This observation concerns only one participant, so it cannot be called rigorous proof. But it takes the study to a particularly interesting level and points to a powerful mechanism: the motor system can respond not only to the physical characteristics of what is seen — shape and speed. What also matters is what a person understands beneath what he sees, what meaning he assigns to it.
The same picture can acquire a completely new meaning — and along with the meaning, as if by magic, the brain's work changes too. The perception of movement turns from a simple registration of form into a process of active interpretation. The brain recognizes an action, relates it to its own bodily experience, and seems to ask itself: "Does this movement look like something I could do?"
The next part of the experiment showed yet another facet of this mechanism — and directly touches on practical application. When participants themselves tried to perform a grasping movement — when they actively generated their own motor command — the differences between the human hand, the robot, and the abstract objects almost completely disappeared. One's own motor signal became dominant, drowning out the effect of observation.
In other words: anthropomorphism is critical precisely for observation. When the brain forms its own motor command, however, the internal signal comes to the fore, and the appearance of the controlled object no longer matters as much. This distinction has direct significance for creating prostheses and robots controlled via an interface.
BrainGate is not just a scientific project. It is a program creating implantable brain–computer interfaces for people who have lost the ability to move or speak because of spinal cord injuries, stroke, or neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS). Such technologies are already converting neural activity into commands for cursors, virtual objects, robotic manipulators, and speech interaction systems.
A new study suggests a non-obvious path to improving these technologies. If the motor cortex responds more strongly to anthropomorphic design, then the outward appearance of a controlled device is not merely cosmetics. It is a mode of dialogue between the brain and the apparatus. Anthropomorphic design can potentially ease the integration of an external prosthesis or a robotic manipulator into the user's consciousness — into their internal map of the body and movements. In the future, such knowledge will help develop interfaces that "embed" more naturally into the neural mechanism of control and the sense of one's own body.
At the same time, this study returns us to the very concept of mirror neurons with a renewed perspective. The phenomenon of observational activity remains quite real — motor areas do indeed respond to others' movements. But a more three-dimensional model appears before us than previously seemed. The response is born not from the work of specialized cells, but from the coordinated activity of an entire distributed network. This network is sensitive to context, changes along with the understanding of what is happening, and depends on what meaning the observed movement receives.
The brain unites vision, the experience of one's own body, and the understanding of what is happening into a single continuous process. When we see another person's movement, its echo arises within the motor cortex. But this echo reflects not only the gesture itself. It reflects whom we see, how much we recognize in them a body similar to our own, and what meaning we assign to their movement. The brain is not a passive mirror. It is an active interpreter.
