AI BriefingUC Berkeley and UCSF Study Reports One-Second Latency for First Sound in Brain-to-Speech Implant
·2026.11.04 02:21
Key point
Researchers at UC Berkeley and UCSF developed a streaming system that produces the first sound of a sentence within about one second of detecting a brain signal, compared to an eight-second delay for a full sentence in their earlier approach.
Details
- Experimental brain-to-speech implant achieved a latency of about one second for the first sound, compared to an eight-second delay for a sentence in the team's earlier approach.
- System tested with one participant, Ann, who has severe paralysis; it maps brain signals from the speech production area to speech sounds.
- AI model trained to decode silent attempts at sentences shown on a screen, using pre-injury recordings to mimic the participant's original voice.
- Study demonstrates decoding of attempted speech rather than reading arbitrary private thoughts.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Fish Audio Drama 3 Ranks #1 Open-Weight Model on Artificial Analysis Speech ArenasAnalysis: LLMs are not superhuman at reading and writing neural network weights