AI researchers have made a groundbreaking advancement in the field of neuroprosthetics by developing a speech neuroprosthesis that can decode brain impulses in real-time and translate them into speech. This innovative technology, created by scientists at UC Berkeley and UC San Francisco, aims to help individuals who have lost the ability to speak due to paralysis or neurological impairments.
The neuroprosthesis, featured in the March edition of Nature Neuroscience, works by bypassing the vocal cords and translating electrical brain activity from the speech-motor cortex directly into synthesized speech. This allows for spontaneous and natural conversation, a feat that previous systems have struggled to achieve due to long delays in decoding speech.
Unlike older models that required a full thought or sentence before producing speech, this new system delivers output in 80-millisecond chunks, enabling users to speak at rates of up to 90 words per minute. The device uses a dense electrode array implanted on the brain’s surface to record neural activity from the speech-motor cortex, which controls mouth and vocal tract movements.
One of the key features of this neuroprosthesis is its ability to distinguish between when a person is attempting to speak and when they are at rest, preventing unwanted mind reading or accidental output. The system only generates speech when the user consciously tries to speak, with no false positives detected during testing.
While the current system still requires invasive electrodes for accurate decoding, the research team is optimistic about its future applications. They believe that clinical use could be possible within 3-5 years, with potential for wider applications beyond medical conditions. The AI model developed for this technology could also enhance voice assistants like Siri or Alexa by processing silent or non-verbal communication.
Looking ahead, the researchers aim to make the system more closed-loop, allowing the AI to learn and adapt in tandem with the user. This shift towards co-adaptive systems represents a significant advancement in brain-computer interfaces, moving from mere tools to restore lost human functions. With continued research and development, this technology has the potential to revolutionize communication for individuals with speech impairments.
