7 March 2022 in Nature Neuroscience
Departing from traditional linguistic models, advances in deep learning have resulted in a new type of predictive (autoregressive) deep language models (DLMs). Using a self-supervised next-word prediction task, these models generate appropriate linguistic responses in a given context. In the current study, nine participants listened to a 30-min podcast while their brain responses were recorded using electrocorticography (ECoG). We provide empirical evidence that the human brain and autoregressive DLMs share three fundamental computational principles as they process the same natural narrative: (1) both are engaged in continuous next-word prediction before word onset; (2) both match their pre-onset predictions to the incoming word to calculate post-onset surprise; (3) both rely on contextual embeddings to represent words in natural contexts. Together, our findings suggest that autoregressive DLMs provide a new and biologically feasible computational framework for studying the neural basis of language.
30 March 2024 in Nature Communications
AGAvigail Grinstein-Dabush
+16
Contextual embeddings, derived from deep language models (DLMs), provide a continuous vectorial representation of language. This embedding space differs fundamentally from the symbolic representations posited by traditional psycholinguistics. We hypothesize that language areas in the human brain, similar to DLMs, rely on a continuous embedding space to represent language. To test this hypothesis, we densely record the neural activity patterns in the inferior frontal gyrus (IFG) of three participants using dense intracranial arrays while they listened to a 30-minute podcast. From these fine-grained spatiotemporal neural recordings, we derive a continuous vectorial representation for each word (i.e., a brain embedding) in each patient. Using stringent zero-shot mapping we demonstrate that brain embeddings in the IFG and the DLM contextual embedding space have common geometric patterns. The common geometric patterns allow us to predict the brain embedding in IFG of a given left-out word based solely on its geometrical relationship to other non-overlapping words in the podcast. Furthermore, we show that contextual embeddings capture the geometry of IFG embeddings better than static word embeddings. The continuous brain embedding space exposes a vector-based neural code for natural language processing in the human brain.
7 March 2025 in Nature Human Behaviour
This study introduces a unified computational framework connecting acoustic, speech and word-level linguistic structures to study the neural basis of everyday conversations in the human brain. We used electrocorticography to record neural signals across 100 h of speech production and comprehension as participants engaged in open-ended real-life conversations. We extracted low-level acoustic, mid-level speech and contextual word embeddings from a multimodal speech-to-text model (Whisper). We developed encoding models that linearly map these embeddings onto brain activity during speech production and comprehension. Remarkably, this model accurately predicts neural activity at each level of the language processing hierarchy across hours of new conversations not used in training the model. The internal processing hierarchy in the model is aligned with the cortical hierarchy for speech and language processing, where sensory and motor regions better align with the model’s speech embeddings, and higher-level language areas better align with the model’s language embeddings. The Whisper model captures the temporal sequence of language-to-speech encoding before word articulation (speech production) and speech-to-language encoding post articulation (speech comprehension). The embeddings learned by this model outperform symbolic models in capturing neural activity supporting natural speech and language. These findings support a paradigm shift towards unified computational models that capture the entire processing hierarchy for speech comprehension and production in real-world conversations.
26 November 2025 in Nature Communications
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.
3 July 2025 in Scientific Data
Naturalistic electrocorticography (ECoG) data are a rare but essential resource for studying the brain’s linguistic capabilities. ECoG offers high temporal resolution suitable for investigating processes at multiple temporal timescales and frequency bands. It also provides broad spatial coverage, often along critical language areas. Here, we share a dataset of nine ECoG participants with 1,330 electrodes listening to a 30-minute audio podcast. The richness of this naturalistic stimulus can be used for various research questions, from auditory perception to narrative integration. In addition to the neural data, we extracted linguistic features of the stimulus ranging from phonetic information to large language model word embeddings. We use these linguistic features in encoding models that relate stimulus properties to neural activity. Finally, we provide detailed tutorials for preprocessing raw data, extracting stimulus features, and running encoding analyses that can serve as a pedagogical resource or a springboard for new research.