Temporal structure of natural language processing in the human brain corresponds to layered hierarchy of large language models

Large Language Models (LLMs) offer a framework for understanding language processing in the human brain. Unlike traditional models, LLMs represent words and context through layered numerical embeddings. Here, we demonstrate that LLMs' layer hierarchy aligns with the temporal dynamics of language comprehension in the brain. Using electrocorticography (ECoG) data from participants listening to a 30-minute narrative, we show that deeper LLM layers correspond to later brain activity, particularly in Broca's area and other language-related regions. We extract contextual embeddings from GPT-2 XL and Llama-2 and use linear models to predict neural responses across time. Our results reveal a strong correlation between model depth and the brain's temporal receptive window during comprehension. We also compare LLM-based predictions with symbolic approaches, highlighting the advantages of deep learning models in capturing brain dynamics. We release our aligned neural and linguistic dataset as a public benchmark to test competing theories of language processing.

Authors

Institutions

Publication Details

Journal
Nature Communications
Published
2025-11-26
DOI
https://doi.org/10.1038/s41467-025-65518-0
Citations
1
Primary Topic
Neurobiology of Language and Bilingualism
Type
article
Field-Weighted Citation Impact
1.43

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Temporal structure of natural language processing in the human brain corresponds to layered hierarchy of large language models

Patricia Dugan, Michael P. Brenner, Orrin Devinsky, Zaid Zada et al.
1 citations
Nature Communications
Neurobiology of Language and Bilingualism
1.43
article

Temporal structure of natural language processing in the human brain corresponds to layered hierarchy of large language models

Patricia Dugan, Michael P. Brenner, Orrin Devinsky, Zaid Zada, Bobbi Aubrey, Noam Siegelman, Omer Levy, Eric Ham, Sasha Devore, Roi Reichart, Amir Feder, Samuel A. Nastase, Yossi Matias, Uri Hasson, Mariano Schain, Adeen Flinker, Daniel Friedman, Harshvardhan Gazula, Ariel Goldstein, Werner Doyle, Avinatan Hassidim, Avigail Grinstein-Dabush
article en
1 citations

Abstract

Large Language Models (LLMs) offer a framework for understanding language processing in the human brain. Unlike traditional models, LLMs represent words and context through layered numerical embeddings. Here, we demonstrate that LLMs' layer hierarchy aligns with the temporal dynamics of language comprehension in the brain. Using electrocorticography (ECoG) data from participants listening to a 30-minute narrative, we show that deeper LLM layers correspond to later brain activity, particularly in Broca's area and other language-related regions. We extract contextual embeddings from GPT-2 XL and Llama-2 and use linear models to predict neural responses across time. Our results reveal a strong correlation between model depth and the brain's temporal receptive window during comprehension. We also compare LLM-based predictions with symbolic approaches, highlighting the advantages of deep learning models in capturing brain dynamics. We release our aligned neural and linguistic dataset as a public benchmark to test competing theories of language processing.

Nature CommunicationsVol. 16(1)
Harvard University (US), University of California, Los Angeles (US), Tel Aviv University (IL), Technion – Israel Institute of Technology (IL), Hebrew University of Jerusalem (IL), Princeton University (US), Google (Israel) (IL), New York University (US)
National Institutes of Health, National Institute of Neurological Disorders and Stroke
Openalex Percentile: Top 14%
Neurobiology of Language and Bilingualism
1.43
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.