Fact · Voice and AI
10milliseconds
In the 2025 Nature study, the decoder turned the participant's neural activity into voice causally, within 10 milliseconds.
Checked against its source on October 17, 2026
In the source's own words
“decoding his neural activity causally within 10 ms”
An instantaneous voice synthesis neuroprosthesis, Maitreyee Wairagkar et al., Nature, 2025, Abstract.
Open the original (DOI 10.1038/s41586-025-09127-3)- What it measures
- Causal means each sound is produced from the neural signal up to that moment, with no look-ahead. The paper contrasts this with text-to-speech approaches that are acausal and must wait for each word to be decoded.
- What it does not show
- This is the decoder's processing step. It is not the full delay a listener hears, and this fact does not claim an end-to-end latency.