Live Speech Production for the Paralyzed

This piece, by Onno Berkan, was published on 04/15/25. The original text, by Littlejohn et al., was published by Nature Neuroscience on 03/31/25.

This joint UCSF & Berkeley study developed a new brain-computer interface that allows a person with severe paralysis to communicate through synthesized speech in real-time by just thinking about speaking. The research focused on a 47-year-old woman who lost her ability to speak due to a brainstem stroke, leaving her with severe paralysis.

The researchers implanted an exceptional array of 253 electrodes on the surface of her brain, explicitly covering areas involved in speech production and comprehension. This device reads brain signals while the participant silently "mimes" or "mouths" what she wants to say, without making any sounds. The system uses advanced artificial intelligence and deep learning techniques to convert electrode readings into synthesized speech and text.

The crucial innovation of this system is its ability to work in near real-time, producing speech with minimal delay as the person thinks about speaking. Previous attempts at speech BCIs had significant delays, waiting until the end of a complete thought before producing any sound. The new system processes brain signals in tiny 80-millisecond chunks, allowing for more natural, flowing conversation.

The system was tested using two sentences: a smaller set of 50 common phrases related to daily care needs, and a more extensive set of over 12,000 sentences using 1,024 unique words. The results showed impressive communication speeds, achieving about 91 words per minute for the smaller phrase set and 48 words per minute for the more extensive vocabulary set. These speeds were significantly faster than previous approaches. 

Perhaps most remarkably, the system recreated speech using the participant's pre-injury voice, which was accomplished by training the system with a short voice recording before she lost the ability to speak. This technology also showed promise in working with other brain-recording devices and could help people with different speech impairments.

Want to submit a piece? Or trying to write a piece and struggling? Check out the guides here!

Thank you for reading. Reminder: Byte Sized is open to everyone! Feel free to submit your piece. Please read the guides first though.

All submissions to berkan@usc.edu with the header “Byte Sized Submission” in Word Doc format please. Thank you!