Scientists have made progress in developing a brain-computer interface (BCI) that translates thoughts about speech into spoken words in real time, potentially aiding individuals unable to speak. This experimental device was tested on a 47-year-old woman with quadriplegia who had not spoken for 18 years due to a stroke. During a clinical trial, researchers implanted the device in her brain, aiming to convert her intent to communicate into coherent sentences.
The study, detailed in the journal Nature Neuroscience, highlights that existing BCIs often experience delays between a person's thoughts and the generated speech. These delays can hinder effective communication. The new approach aims to eliminate such interruptions. According to Gopala Anumanchipalli, a co-author of the study, the device operates by translating the user's neural activity into speech units in real time, which could enhance the natural flow of conversation.
Researchers in California recorded the participant’s brain activity while she silently formed sentences in her mind. Using her pre-injury voice samples, they trained an AI model to synthesize her speech. The device is designed to process speech in small segments, transmitting approximately half a syllable every 80 milliseconds without waiting for complete sentences.
Experts, including Jonathan Brumberg from the University of Kansas, have recognized this as a significant advancement in speech technology. While further research is necessary before the device can be widely implemented, Anumanchipalli believes that with continued investment, this technology could be available for patient use within the next decade. The project has received partial funding from the National Institutes of Health, although it has not been affected by recent budget cuts.