Advances in Brain-Computer Interface Technology Help One Man Find His Voice

BCI-Blog-Header

Not that long ago, it would have been extremely difficult for Casey Harrell to tell his story. In the five years since his first ALS symptoms appeared, the disease had made it nearly impossible for anyone other than his closest caregivers to understand the 46-year-old father of one.

“I had a long process of losing my ability to communicate, from losing the ability to sing a song to my daughter, from needing to repeat myself over and over to be understood by most people, to finally needing specific people to translate for me. It was very painful to lose the ability to communicate, especially with my daughter,” Harrell recalled during a recent Intention to Treat podcast produced by the New England Journal of Medicine.

Not being able to communicate is so frustrating and demoralizing. It is like you are trapped.”
Casey Harrell

But last summer, a team at UC Davis Health led by neurosurgeon Dr. David Brandman surgically implanted an array of 256 electrodes into Harrell’s brain, sensors that could be used to help translate his attempts to speak into words “spoken” by a computer.

Dr. Brandman, together with neuroscientist Dr. Sergey Stavisky, lead the BrainGate trials at UC Davis and are developing a “neurological prosthesis” to restore speech for people affected by paralysis or neurological conditions like ALS with the support of two ALS Association grants.

“We’re really detecting their attempt to move their muscles and talk,” explained Dr. Stavisky in a press release. “We are recording from the part of the brain that’s trying to send these commands to the muscles. And we are basically listening into that, and we’re translating those patterns of brain activity into a phoneme—like a syllable or the unit of speech—and then the words they’re trying to say.”

You're not imagining a conversation. This isn't the voice in your head rehearsing what you might say. You've already decided you want to speak.”
Dr. Sergey Stavisky

Dr. Stavisky was awarded a Seed Grant in 2022 to help pilot this speech-decoding technology, which allowed for the recruitment of Harrell into the trial.

Three weeks after the surgery, the researchers arrived at Harrell’s house to connect the implant to the team’s computers through two metal posts poking out from the top of his head.

After giving Harrell some basic instructions, the device began decoding what he was trying to say from a 50-word vocabulary with 99.6% accuracy. The decoded words were shown on a screen, and then amazingly, read aloud in Harrell’s own pre-ALS voice. This voice was created using AI software trained from old podcast interviews and other audio recordings made before Harrell’s diagnosis.

“The first time we tried the system, he cried with joy as the words he was trying to say correctly appeared on-screen. We all did,” Dr. Stavisky said.

During a second session, after a little less than an hour and a half of additional training, the vocabulary available to Harrell grew to 125,000 words, and the device achieved 90.2% accuracy. This accuracy increased to 97.5% with continued data collection. These results are described in paper published last week in the New England Journal of Medicine.

At this point, we can decode what Casey is trying to say correctly about 97% of the time, which is better than many commercially available smartphone applications that try to interpret a person’s voice.”
Dr. David Brandman

According to the researchers, with additional tweaks and innovations supported by a 2023 Assistive Technology Grant, Harrell is now able to use this system whenever he wants, not just during structured, scheduled research sessions. He has reported using the system multiple days per week for up to 12 hours per day to speak with his family and friends, participate in Zoom meetings, send emails, reply to text messages, and surf the Internet.

“It has also allowed me to continue to work full-time and more efficiently than before,” Harrell said during his podcast interview. “I am about half of the speed of a normal person who’s talking, and that is good enough for now to keep up with the flow of conversation. And this is important because communication is the lifeblood of life.”

Something like this technology will help people back into life and society.”
Casey Harrell

Dr. Kuldip Dave, senior vice president for research at the ALS Association, underscored the importance of developing state-of-the-art technologies like brain-computer interfaces to make ALS a livable disease.

“We want people with ALS to live their lives to the fullest. And over the last 10 years, we have been working toward that goal by funding the development of cutting-edge and innovative technologies that can help people with ALS move better, breathe better, and communicate better,” said Dr. Dave. “The quicker we can make these new technologies available to everyone who needs them, the faster we can make ALS more livable.”

Dr. Brandman and Dr. Stavisky are now looking to replicate Harrell’s success with other BrainGate participants before making the technology more widely available.

“I would love to get to the point where I can walk up to a patient and say: ‘I'm really sorry that you've been diagnosed with ALS. Your brain-computer interface surgery is scheduled for next week. We'll get you communicating in no time,’" said Dr. Brandman.

To learn more about this research, including how to participate, visit braingate.org or contact braingate@ucdavis.edu.

If you would like to receive monthly ALS research updates, 
SUBSCRIBE to our newsletter, Research Matters. 

To continue to follow stories about people living with ALS in the community and learn more about the disease, subscribe to receive our weekly blogs in your inbox HERE or follow us at als.org/blog.   

Subscribe

We Want to Hear From You!
Your feedback is important to us. Please let us know what you think about our blog – take this short survey and provide your input and feedback.

Join the conversation. Please comment below.

CAPTCHA
6 + 13 =
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.