VOICE Study

The VOICE Study is Neuralink's trial to restore speech for people who can no longer talk because of amyotrophic lateral sclerosis (ALS, a disease that destroys the nerves that control muscle), stroke, or spinal cord injury. It is one of Neuralink's 3 implant trials in the United States, alongside PRIME for computer control and CONVOY for robotic arm control.

What is the VOICE Study?

VOICE is registered on ClinicalTrials.gov, the United States government's trial registry, as NCT07224256. It uses Neuralink's N1 implant, the same coin-sized brain-computer interface used in PRIME, to decode intended speech instead of intended cursor motion. In May 2025 the United States Food and Drug Administration (FDA) gave Neuralink's speech-restoration device a Breakthrough Device Designation, a status that can speed FDA review.

Why the VOICE Study Exists

PRIME showed that a wireless implant can give people with severe paralysis control of a computer. Speech is the next channel. For someone who can no longer talk, a direct thought-to-speech path is the difference between typing on a screen and joining a conversation, and Neuralink's stated goal is 140 words per minute, the pace of normal conversation.

How it Works

In the PRIME Study the implant listens to the hand area of the brain. In VOICE the threads go into the speech motor region, the strip of brain that plans how the lips, tongue, and voice box move. A participant presses a key, thinks a sentence as if saying it aloud, and presses a second key to send it. Software turns the signals into words, and a voice rebuilt from recordings made before the participant lost speech says them aloud. The R1 surgical robot inserts the threads.

The Economics of the VOICE Study

The implant, the robot, and the speech software are investigational devices in a clinical trial, so they carry no price. Neuralink covers the device, the surgery, and study travel for each participant and a caregiver.

Current Status of the VOICE Study

The trial began on October 3, 2025 at UT Southwestern Medical Center in Dallas. Kenneth Shock, a former paramedic with ALS, received his implant there on January 5, 2026. Terence (Terry) Mandable, who has bulbar-onset ALS, the form that takes speech first, is the third VOICE participant, and his implant carries 3,000 channels, about 3 times the 1,024 in earlier devices. On September 16, 2026 Neuralink posted video of Terry telling his wife "I love you" in his own rebuilt voice, and 2 days later it posted the full story. Neuralink co-founder and president DJ Seo said the voice is trained and generated by Grok, the artificial intelligence model from SpaceXSI, a unit of SpaceX.

What Comes Next

The number to watch is VOICE's first published words-per-minute result and how close it gets to the conversational goal. PRIME participants already type up to 40 words per minute with the same implant family by imagining finger movements.

The Bottom Line

VOICE is the first Neuralink trial aimed at the voice instead of the hands, and it produced its first public result within a year of starting.