An instantaneous voice-synthesis neuroprosthesis

The 60-second view

What problem?

Text output loses prosody, and delays between attempt and sound break the flow of conversation.

What did they do?

Decoded activity from 256 microelectrodes in ventral precentral gyrus straight to voice, giving a man with ALS immediate audio feedback of his own synthesised speech.

What changed?

Moves speech BCIs from producing transcripts to producing a voice the user can hear and modulate as they speak.

Why does it matter?

Hearing yourself as you speak is how people regulate speech. Closing that loop is a step toward conversation that feels natural to both sides.

Method

Participants

One man with ALS and severe dysarthria

Interface

256 intracortical microelectrodes, ventral precentral gyrus

Signal

Intracortical spiking activity

Task

Attempted speech with real-time audio feedback

Decoder

Causal brain-to-voice model decoding phonemic and paralinguistic features

Results

The published abstract does not state headline figures, so none are quoted here. See the primary source for quantitative results.

Strength of evidencePeer reviewed, one participant. Trained without ground-truth speech audio from the participant.

Figures are quoted from the published abstract and were checked on 3 Oct 2026.

Limitations

• One participant.
• Synthesised speech is intelligible but not yet equivalent to natural speech.

What this could enable

Expressive conversation: emphasis, questions, interjections and singing, not only words.