HMN 2026: What is the brain-computer interface that works with—not against—the brain

It might soon be “game over” for the video game controller. Yale researchers have developed a new kind of brain-computer interface (BCI) that lets humans play video games directly with their brains. Using real-time fMRI (functional MRI), they confirmed that the technology could help humans control a computer with their brain activity in a highly efficient way. The study appears in the journal Nature Neuroscience.

How effective are human BCIs?

A BCI is technology that allows a human to control a computer with brain activity. Historically, they have not been effective. BCIs built using real-time neurofeedback from fMRI—a type of MRI scan showing which areas of the brain are most active over time—require up to 10 long training sessions per person, and even then the learning effects are modest. About a third of users never gain control, regardless of how many hours they practice.

What is the key to learning how to use a BCI?

Activity in the brain travels along well-worn routes. Working with those routes, rather than against them, was the key to learning how to use a BCI, the researchers found. When a BCI is built around these routes, people gain rapid control, and their brain activity reorganizes to support the learning. A BCI that ignores this natural geometry, by contrast, produces little or no learning.

“The implications are broad, from helping people with motor or communication disorders to developing treatments for depression or anxiety to building the next generation of consumer games and technologies: interventions designed around the brain’s natural geometry are likely to be faster, more effective and more accessible,” said Erica Busch, the first author of the study, who recently completed her Ph.D. at Yale.

Busch and her colleagues suspected that this was a consequence of how brain activity was measured and trained: The BCIs were asking the brain to learn something poorly matched to its natural geometry. They hypothesized that more sensitive tools—specifically, ones that tailor neurofeedback around this geometry—could dramatically improve the speed and effectiveness of BCI learning.

“Could we build a system smart enough to discover that geometry in real time, using noninvasive brain imaging?” asked Smita Krishnaswamy, associate professor of genetics at Yale School of Medicine (YSM) and of computer science at Yale School of Engineering & Applied Science (Yale Engineering) and a corresponding author of the study.

With the new study, the Yale researchers sought to answer that question. To do that, they asked healthy young adults to complete four fMRI scanning sessions. During the first session, participants played a video game—in which they steered an avatar through a virtual arena using a joystick—while the researchers recorded their brain activity. The researchers focused on a network of brain regions known to be involved in how people navigate the world.

They then used an algorithm called T-PHATE that they had developed in earlier work, a mathematical approach that learns the natural geometry of each person’s brain activity, their individual “neural manifold.”

From that manifold, the researchers designed three different ways of mapping a person’s brain activity to the avatar’s movement in the game: one that worked with the brain’s most natural patterns (the most well-traveled neural route; the “intuitive mapping”); one that used less dominant but still natural patterns (another regularly traveled route; the “within-manifold perturbation”); and one that required patterns the brain doesn’t naturally produce (like paving an entirely new road; the “outside-manifold perturbation”).

The researchers then built a closed-loop system that reads a new brain scan every two seconds and instantly translated it into the avatar’s movement direction. During the final three sessions, participants tried to control the avatar purely by thinking, with one session for each mapping, and the researchers measured how quickly and accurately they learned.

Game-changer

Through this process, the researchers found that participants successfully learned to control the avatar with less than an hour of training—and sometimes much faster—when the BCI mapping followed the brain’s natural manifold. When the mapping veered away from the manifold, the participants couldn’t learn it at all during that same timeframe.

But BCI learning didn’t just change behavior: The brain itself reorganized under the hood, shifting its activity to better align with what the BCI was asking for. In some conditions, this reorganization predicted how well each person performed, and it spread to brain regions outside the targeted areas, suggesting that BCI learning ripples across different parts of the brain.

“The manifold is both a constraint and an opportunity: it determines what people can learn and how fast,” said Nick Turk-Browne, director of the Wu Tsai Institute and the Susan Nolen-Hoeksema Professor of Psychology in the Yale Faculty of Arts and Sciences (FAS) and a corresponding author of the study.

The findings, researchers say, suggest why certain things feel hard to learn. It might not come down to effort or ability, but to how well what a person is trying to learn fits their existing neural architecture.

These findings could have implications well beyond the lab, researchers say. In mental health, they suggest that symptoms associated with conditions like depression or anxiety—where the brain gets stuck in unhelpful grooves—might be better addressed using strategies that work incrementally with the brain’s existing architecture rather than attempting a complete overhaul.

For people with motor or communication difficulties, the findings could offer a path toward new BCIs that work more reliably for more people. For the broader population, it points toward the possibility of cognitive self-enhancement through training healthy people to think better.

“We spend tremendous resources trying to become better versions of ourselves through education, practice, therapy and more,” Busch said. “Understanding the structure of our own mind and brain may help us do that far more effectively.”

The content is provided for information purposes only.