12 Jun 2025

Brain-computer interface synthesises a participant's voice in real time, including intonation

Neurotechnology and human augmentationMedical technology and bioelectronics
AI illustration · Illustrative scene, not a photograph of the specific device or experiment.

What was demonstrated

Using 256 microelectrodes implanted in the ventral precentral gyrus of one man with ALS and severe dysarthria, the team decoded neural activity directly into synthesised voice with closed-loop audio feedback, rather than into text. The decoder also captured paralinguistic features, letting the participant change intonation and sing short melodies through the system.

Why it matters

Instant voice output, rather than text, could restore more natural conversation, including the tone and timing that text-based systems lose.

Caveats

  • Single participant with ALS; replication in more people and other conditions such as stroke is still needed.
  • There was no ground-truth speech from the participant to train on, so the method relies on inferred targets.
  • Invasive implant within a research trial, not an approved device.

Technology profiles

Sources

  1. peer reviewed paper
    An instantaneous voice-synthesis neuroprosthesis
    Maitreyee Wairagkar, Nicholas S. Card, Tyler Singer-Clark et al.. Nature, 12 Jun 2025. DOI: 10.1038/s41586-025-09127-3.

Last reviewed 20 Oct 2026.