Distributed cortico-subcortical networks enable robust speech state detection from sparse intracranial recordings
Chen Feng, En Zhang, Yifei Jia, Zhoule Zhu, Junming Zhu, Di Wu, Kedi Xu · 2026
Read Distributed cortico-subcortical networks enable robust speech state detection from sparse intracranial recordings by Chen Feng, En Zhang, Yifei Jia, Zhoule Zhu, Junming Zhu, Di Wu, Kedi Xu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13194996.txt.
Description
Chen Feng and co-authors show that the brain's 'about to speak' signal is spread far wider than the classic speech areas. A speech brain-computer interface must catch the moment speech starts, reliably and from few electrodes. Their 2026 study records from deep and surface sites in four Mandarin-speaking epilepsy patients as they read sentences aloud, then trains a neural network to sort rest from speech.
Accuracy beat chance in every patient and topped ninety percent in the best case. Mixing deep and surface signals beat either alone, and broad-band signals beat the usual high-frequency band. Dropping half the channels at random barely dented the scores. Accuracy stayed above seventy percent even with the classic speech regions removed, with frontal areas and the thalamus contributing strongly. The authors argue that sparse deep recordings alone could drive a practical speech detector.
Catalog record
| CREATOR | Chen Feng, En Zhang, Yifei Jia, Zhoule Zhu, Junming Zhu, Di Wu, Kedi Xu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |