Brain implant leverages AI, neural sensors to restore speech for ALS patient

Editor's Note

Researchers at UC Davis have developed a brain-computer interface (BCI) implant that has successfully restored speech in a patient with amyotrophic lateral sclerosis (ALS), whose ability to speak had deteriorated, Fierce Biotech reported August 15. The breakthrough was achieved using neural sensors from Blackrock Neurotech and AI-driven text-to-speech software, which replicates the patient’s own voice based on previous recordings.

The patient, 45-year-old Casey Harrell, expressed the technology made him sound like himself again. Unlike systems that translate thoughts, this BCI captures brain signals intended for speech muscles, converting them into phonemes and full words. The system has shown an impressive 97% accuracy, rivaling commercial speech-to-text software.

The BCI system uses Blackrock’s NeuroPort array, which includes 64 electrodes implanted into the brain’s cortex. Harrell, who underwent the procedure in July 2023, was able to communicate almost immediately during testing, with a vocabulary exceeding 125,000 words.

Previous speech BCI systems had struggled with frequent errors, hindering consistent communication. The new system, however, aims to enable users to be understood clearly whenever they wish to speak. The research, part of the BrainGate clinical trial consortium, was published in the New England Journal of Medicine.

This advancement builds on past successes, including enabling a paralyzed patient to type at 62 words per minute. The work underscores the dedication of participants like Harrell, who join these trials to help develop life-changing technology for others with paralysis.