Skip to content
DraconiticResearch
Search Sources API
220,782 studies31 sourcesUpdated 39 min ago
Peer-reviewedCC BY-NC-NDNature

Mapping the neuronal building blocks of human language with language models

  • Jing Cai1
  • Yoav Kfir2
  • Mohsen Jamali2
  • Hesen Huang2
  • Young Joon Kim2
  • Sydney S. Cash3,4,5,6
  • Ziv M. Williams7,8,9,10
10 affiliations
  1. 1Department of Neurosurgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. [email protected]
  2. 2Department of Neurosurgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
  3. 3Center for Neurotechnology and Neurorecovery, Department of Neurology, Massachusetts General Hospital, Boston, MA, USA. [email protected]
  4. 4Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. [email protected]
  5. 5Harvard Medical School, Program in Neuroscience, Boston, MA, USA. [email protected]
  6. 6Harvard-MIT Division of Health Sciences and Technology, Boston, MA, USA. [email protected]
  7. 7Center for Neurotechnology and Neurorecovery, Department of Neurology, Massachusetts General Hospital, Boston, MA, USA. [email protected]
  8. 8Department of Neurosurgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. [email protected]
  9. 9Harvard Medical School, Program in Neuroscience, Boston, MA, USA. [email protected]
  10. 10Harvard-MIT Division of Health Sciences and Technology, Boston, MA, USA. [email protected]
Open at sourceDOICite

Abstract

Humans can convey new and highly diverse information through language. This ability to form and combine words into elaborate phrases and sentences enables us to express inexhaustible meanings and is fundamental to human cognition1–5. However, understanding the microscopic cellular building blocks and cortical landscape that precisely underlie human language has remained a challenge. Here we used wide-scale single-neuronal recordings combined with natural language processing models to identify fine-grained linguistic representations across the human frontotemporal cortex during language production. We find that, whereas certain neurons represented the detailed grammatical relationships between words or their parts of speech, others tracked the sentences’ higher-order syntactic structure, their phrase transitions and sequence. Collectively, these neurons reliably captured the words’ syntactic and semantic properties but also dynamically incorporated their specific sentence contexts, therefore enabling them to encode information combinatorially and at highly granular levels of detail. We show how these cell populations were locally organized and how their microscale representations differed from that of their wider field potential patterns. We also show how these neurons were distributed broadly across the frontotemporal cortex, but how their ability to encode linguistic information was left-lateralized and varied between cortical regions. Together, these findings identify some of the most basic cellular building blocks by which linguistic information is encoded in humans and begin to define the cortical landscape of language at a combined micro (cellular), meso (local population) and macro (regional) scale. Wide-scale recordings reveal neurons in the human brain that encode fundamental components of language such as the grammatical relationships between words, their parts of speech and the higher-order syntactic structure of phrases and sentences.

Related

  1. Volumetric building blocks: A simple and modular approach to create perfusable models via assembly of volumetrically printed subunits
    M. Bartolf-Kopp, I. Geroneit, C. Gergely +6
    bioRxivPreprintCC BY-NCFull textSystems & synthetic biologybioRxiv
  2. Sensory and Motor Building Blocks of an Extra Robotic Digit Representation
    L. Dowdall, E. Amoruso, G. Dominijanni +4
    bioRxivPreprintCC BYFull textNeurosciencebioRxiv
  3. Cryptic Genes as Building Blocks of Antimicrobial Resistance: Evidence from a Novel Marine Imipenem-Hydrolyzing Metallo-β-Lactamase
    P. Gupta, P. Yadav, J. Kishan +8
    bioRxivPreprintCC BY-NCFull textMicrobiologybioRxiv
  4. The building blocks of social structure: simulating constraints of social network analysis for inference about group-level properties
    J. Brooks, G. Badihi, L. Samuni
    bioRxivPreprintCC BYFull textZoology & animal behaviourbioRxiv
  5. Genome-Wide Mapping of Major Histone Modifications Reveals Distinct Epigenetic Regulatory States in a Reef-Building Coral
    G. Warfel, S. L. Davidson, L. Sadalla +6
    bioRxivPreprintAll rights reservedFull textGenetics & genomicsbioRxiv

Cite

BibTeXRIS
BibTeX
@article{cai2026mapping,
  title = {{Mapping the neuronal building blocks of human language with language models}},
  author = {Jing Cai and Yoav Kfir and Mohsen Jamali and Hesen Huang and Young Joon Kim and Sydney S. Cash and Ziv M. Williams},
  journal = {Nature},
  year = {2026},
  month = jun,
  doi = {10.1038/s41586-026-10691-5},
  url = {https://doi.org/10.1038/s41586-026-10691-5},
  note = {CC BY-NC-ND},
}
RIS
TY  - JOUR
TI  - Mapping the neuronal building blocks of human language with language models
AU  - Jing Cai
AU  - Yoav Kfir
AU  - Mohsen Jamali
AU  - Hesen Huang
AU  - Young Joon Kim
AU  - Sydney S. Cash
AU  - Ziv M. Williams
PY  - 2026
DA  - 2026/06/17
JO  - Nature
DO  - 10.1038/s41586-026-10691-5
UR  - https://doi.org/10.1038/s41586-026-10691-5
AB  - Humans can convey new and highly diverse information through language. This ability to form and combine words into elaborate phrases and sentences enables us to express inexhaustible meanings and is fundamental to human cognition1–5. However, understanding the microscopic cellular building blocks and cortical landscape that precisely underlie human language has remained a challenge. Here we used wide-scale single-neuronal recordings combined with natural language processing models to identify fine-grained linguistic representations across the human frontotemporal cortex during language production. We find that, whereas certain neurons represented the detailed grammatical relationships between words or their parts of speech, others tracked the sentences’ higher-order syntactic structure, their phrase transitions and sequence. Collectively, these neurons reliably captured the words’ syntactic and semantic properties but also dynamically incorporated their specific sentence contexts, therefore enabling them to encode information combinatorially and at highly granular levels of detail. We show how these cell populations were locally organized and how their microscale representations differed from that of their wider field potential patterns. We also show how these neurons were distributed broadly across the frontotemporal cortex, but how their ability to encode linguistic information was left-lateralized and varied between cortical regions. Together, these findings identify some of the most basic cellular building blocks by which linguistic information is encoded in humans and begin to define the cortical landscape of language at a combined micro (cellular), meso (local population) and macro (regional) scale. Wide-scale recordings reveal neurons in the human brain that encode fundamental components of language such as the grammatical relationships between words, their parts of speech and the higher-order syntactic structure of phrases and sentences.
SN  - 0028-0836
SN  - 1476-4687
ER  -