Unprecedented brain implant allows paralyzed man to completely control his computer and “speak” independently

A recent study published in Nature Medicine provides evidence that a specialized brain implant can allow a person with severe paralysis to independently communicate and operate a computer at home. By translating brain signals into text and computer cursor movements, the system allowed the participant to converse and work without needing daily supervision from scientists. These findings represent a substantial step toward creating practical assistive devices for people who have lost the ability to speak or move.

Amyotrophic lateral sclerosis is a progressive neurological disease that gradually impairs a person’s ability to control their muscles. This condition often leads to a complete loss of speech and physical mobility. To help individuals with this condition, scientists have been developing brain-computer interfaces. These are systems that record electrical activity directly from the brain and translate it into digital commands.

In previous laboratory experiments, these interfaces have shown promise by allowing paralyzed individuals to type or control a cursor using their thoughts. However, these experimental systems usually required a team of technicians to set up the equipment and manually adjust the software to keep it working accurately. The neural signals recorded from the brain tend to shift slightly from day to day, requiring frequent recalibration.

“For years, BCIs have been proof-of-concept devices that lived in highly controlled research labs,” said David Brandman, co-senior author of the study and associate professor in the University of California, Davis, Department of Neurological Surgery. “This work shows that we may have crossed a threshold, by empowering a person with paralysis to speak on his own terms.” Brandman also serves as the co-director of the UC Davis Neuroprosthetics Lab.

The transition from a supervised laboratory demonstration to a practical communication device requires a system that a patient and their family can operate on their own. A collaborative team of scientists from the University of California, Davis, Brown University, and Mass General Brigham developed a modified interface designed specifically to overcome these barriers. The new system bypasses the need for constant expert supervision. They programmed the software to automatically update itself in the background, creating a user-friendly process for independent home use.

Free daily newsletter

The study involved a single participant, 47-year-old Casey Harrell, who enrolled in an ongoing clinical trial. Harrell has amyotrophic lateral sclerosis, which has caused severe weakness in his arms and legs, a condition known as tetraparesis. The disease has also made his speech very hard to understand, a symptom referred to as dysarthria.

In 2023, surgeons implanted four tiny sensor arrays, each containing 64 microscopic electrodes, into the surface of Harrell’s brain. These sensors were placed in the precentral gyrus, a specific area of the motor cortex responsible for coordinating speech. The 256 microscopic electrodes recorded the electrical firing of individual brain cells. This raw neural data was sent through a physical wired connection embedded in his skull to a nearby computing system.

The computing setup consisted of several networked computers mounted on a mobile cart in Harrell’s home. The software relied on advanced artificial intelligence models to decipher his intentions in real time. For speech decoding, the system continuously analyzed the brain signals while he attempted to silently mouth words. The software predicted the corresponding speech sounds, known as phonemes, and formed full sentences on a screen.

The system utilized a vast vocabulary of 125,000 words. “In our previous study, we showed 97% accurate word decoding. But Harrell could only use the neuroprosthesis when someone from our research team was there to set it up,” said Sergey Stavisky, co-senior author of the study and assistant professor in the UC Davis Department of Neurological Surgery. Stavisky also serves as the co-director of the UC Davis Neuroprosthetics Lab.

“Now we’ve made improvements that bring this medical technology closer to clinical usefulness: He can use it at home without researcher support,” Stavisky explained. “It’s even more accurate (99%), keeping up as he attempts to speak faster, and has been working very well for almost two years.”

For computer control, the system translated Harrell’s thoughts of moving his hand into the directional movements of a computer mouse pointer. A simple thought of squeezing his hand was translated into a digital mouse click. The researchers also integrated a commercial eye-tracking device. This allowed him to select on-screen buttons by simply looking at them for a fraction of a second.

After an initial testing phase, the research team allowed Harrell and his family to use the system on their own. Caregivers were taught how to connect the wires and turn on the software. This daily setup process took about twenty minutes. Once the system was running, he could use it continuously for up to nineteen hours without any assistance from the scientific team.

“Casey can use the system to communicate his own thoughts, not only while we’re there in a controlled environment, but whenever he wants,” said Nicholas Card, lead author of the study and postdoctoral scholar in the UC Davis Department of Neurological Surgery. “Sometimes, he would do that over 12 straight hours. The system worked well, was reliable and stable, and delivered consistent results.”

Card added that the success of the home setup is a major milestone for assistive technology. “This is one of the strongest demonstrations that BCIs can be practical and useful,” he noted. Over the course of nearly two years, Harrell used the interface for more than 3,800 hours. He operated the device independently on a near-daily basis.

During that time, Harrell communicated more than 183,000 sentences and close to two million words. In controlled testing, he rated 92 percent of his sentences as accurate or mostly correct. His average speaking rate reached 56 words per minute, a speed that increased significantly as he became more accustomed to the system.

“It is a life that is more full of dynamic action and with friends and family, with colleagues, and it is something that allows me to communicate more in my natural way of communicating than any other technology that I have experienced,” Harrell shared through the brain-computer interface system. The software even included an optional text-to-speech feature trained to match his voice from before his diagnosis. “It is very sweet to have the ability to look at my wife’s eyes when she hears my voice and conjures up a sweet memory and to explain to my daughter who does not really remember anything about when I was still talking to them and remind them of what I used to sound like,” he added.

Harrell also used the interface to gain full control over his personal computer. By combining the brain-to-text feature with the mind-controlled computer cursor, he browsed the internet and sent emails and text messages. He also participated in video calls and was able to maintain ongoing communication and employment despite his paralysis.

The immense amount of personal use time provided the researchers with an unprecedented amount of data. “In addition to testing a way to restore communication, this clinical trial is producing a wealth of unique data that we’re studying to better understand how the human brain produces speech,” Stavisky said.

“As far as we know, these 3,800 hours of brain recording as Casey used the system is by far the largest individual brain recording dataset with single neuron resolution,” Stavisky added. “This will help us develop even better therapies.” Future analysis of this data tends to yield new insights into the neurological mechanics of language.

While the outcomes are highly promising, the study has limitations that should be noted. The research involved only a single participant, meaning the results might not automatically apply to other individuals with different neurological conditions. The system also relies on physical wires passing through the skin. This setup carries a small risk of infection and requires daily maintenance.

The current computer equipment is quite bulky and is confined to a large mobile cart. This physical hardware requirement limits the user’s mobility and prevents the system from being used outside the home environment. Readers should also be aware of potential financial conflicts of interest among the research team, as several authors hold patents related to speech decoding technologies. Some team members also serve as advisors for neurotechnology companies that could benefit from related scientific advancements.

“This fundamental advance in BCI technology could not have been possible without the tireless dedication of participants in clinical trials,” Brandman said. “It is by working together with them that we have achieved so much. Thanks to them, the future will be brighter for people living with ALS, spinal cord injuries and other neurological conditions.”

The study, “Long-term independent use of an intracortical brain–computer interface for speech and cursor control,” was authored by Nicholas S. Card, Tyler Singer-Clark, Hamza Peracha, Carrina Iacobacci, Xianda Hou, Maitreyee Wairagkar, Zachery Fogg, Elena C. Offenberg, Leigh R. Hochberg, Sergey D. Stavisky, and David M. Brandman.