Casey Harrell, 47, lost his voice to amyotrophic lateral sclerosis, or ALS, in his early 40s. Six years after his diagnosis, he talks with his family, answers emails and keeps a full-time job, using a brain implant that turns his attempted speech into words spoken in a voice built from recordings of his own. The brain implants ALS patients are testing in a trial at the University of California, Davis do not cure the disease or restore natural speech, but Harrell has used his system at home for nearly two years and communicated close to 2 million words with it.
Brain implants ALS patients use to speak again
Harrell began to lose his ability to talk shortly after his diagnosis in 2020. “My voice is the most important thing I have lost in this whole disease,” he said. ALS kills nerve cells in the brain and spinal cord that control muscles, and it affects an estimated 35,000 people in the United States. As the connections between Harrell’s brain and body faltered, fewer and fewer people could understand him. He learned about the UC Davis trial through a family member and enrolled in BrainGate2, a pilot clinical trial testing brain-computer interfaces in people with severe speech impairments.
In 2023, UC Davis neurosurgeon David Brandman implanted four microelectrode arrays on the left side of Harrell’s brain, on the speech motor cortex along the precentral gyrus, the stretch of brain that controls the movements needed for speech. The brain implants ALS patients receive in the trial record activity when Harrell tries to speak, even when he does not move his mouth. Software decodes the signals into text, and a synthetic voice reads the words aloud in a voice modelled on recordings Harrell made before ALS took his speech.
How brain implants ALS patients use the system at home
The brain implants ALS patients use at home work without researchers present to run them. Harrell plugs in each morning and uses his system for about 10 to 14 hours a day, and he calls himself “the world’s biggest power user of the brain-to-computer interface known to man.” Researchers at UC Davis reported in June in Nature Medicine that he had communicated 183,060 sentences and 1,960,163 words through the system, at an average rate of 56 words per minute. He sends emails, browses the internet and works full time.
The UC Davis team built the system with collaborators at Brown University and the Mass General Brigham Neuroscience Institute. A separate UC Davis study, published in Nature in 2025, showed the same approach could turn brain activity into voice within 10 milliseconds of the signal, with changes in intonation and pitch, and let the participant sing simple melodies. The brain implants ALS patients test in these studies are still early versions of the technology, and the researchers note that the findings so far come from a single participant.
The brain implants ALS patients are testing have clear limits. The system is bulky, relatively slow and limited in the nonword sounds it can decode. Leigh Hochberg, a neuroscientist at Brown University who works with Harrell, called it “a beginning” and said there is a lot of research still to happen. David Brandman said the technology has the potential to dramatically change the conversation for people who cannot speak.
What comes next for brain implants ALS patients
Other teams are working on the same problem. Neuralink’s VOICE trial is testing a thought-to-speech implant in people with severe speech impairment, including ALS. One participant, Kenneth Shock, was diagnosed with ALS in January 2024 and received a Neuralink N1 implant in January 2026. Decoding software maps his intended speech to phonemes and produces words in a voice rebuilt from recordings made before his symptoms progressed. The device is investigational and is not approved by the Food and Drug Administration.
The brain implants ALS patients are testing across these trials share the same aim: a way to speak that does not depend on muscles that no longer respond. Matthew Willsey, a neurosurgeon and biomedical engineer at the University of Michigan, said scientists once did not know whether decoding speech would work at all, and a body of research results now points to yes. For Harrell, the result is already practical. He talks with his wife and daughter at home, and his daughter, who barely remembers his old voice, hears it again each day.