UC Davis researchers debut highly accurate brain-to-speech technology

UC Davis creates a BCI translating brain signals to text with sub-3% error rate, restoring ALS patient's communication.

: UC Davis introduced a brain-computer interface translating brain signals to readable text with a less than 3% error rate. This technology restored the ability to communicate for Casey Harrell, an ALS patient, using microelectrode arrays and the NeuroPort system. The study presented by Sergey Stavisky and David Brandman achieved a sustained accuracy of over 97.5% after training.
2

A team from UC Davis introduced a breakthrough brain-computer interface that translates brain signals to text with remarkable accuracy. This advancement, featuring a less than 3% error rate, was detailed in the New England Journal of Medicine on August 14, 2024.

Casey Harrell, a 45-year-old ALS patient, regained his ability to communicate using this technology. After implanting microelectrode arrays, Harrell's communication accuracy improved dramatically, reaching a 99.6% accuracy with a 50-word vocabulary in initial tests.

Further training expanded the vocabulary to over 125,000 words, achieving a sustained accuracy of 97.5%. UC Davis neuroscientist Sergey Stavisky and neurosurgeon David Brandman led this promising study, signaling a significant advancement for disabled individuals.

Brain Computerspeech recognitionhealthcare