The neural architecture of language: Integrative modeling converges on predictive processing
Нейронная архитектура языка: интегративное моделирование сходится к предиктивной обработке
2021-11-04
SCID: 54.1/v9uqfx9z
Discuss with AI
language processingneural responsesnext-word predictionpredictive processingtransformer models
Figures from the paper
Abstract (AI)
The neuroscience of perception has recently been revolutionized with an integrative modeling approach in which computation, brain function, and behavior are linked across many datasets and many computational models. By revealing trends across models, this approach yields novel insights into cognitive and neural mechanisms in the target domain. We here present a systematic study taking this approach to higher-level cognition: human language processing, our species' signature cognitive skill. We find that the most powerful "transformer" models predict nearly 100% of explainable variance in neural responses to sentences and generalize across different datasets and imaging modalities (functional MRI and electrocorticography). Models' neural fits ("brain score") and fits to behavioral responses are both strongly correlated with model accuracy on the next-word prediction task (but not other language tasks). Model architecture appears to substantially contribute to neural fit. These results provide computationally explicit evidence that predictive processing fundamentally shapes the language comprehension mechanisms in the human brain.
Key Findings
1
Model architecture substantially contributes to how well computational models fit neural responses.
2
Models’ neural brain scores and behavioral-response fits strongly correlate with next-word prediction accuracy, but not with performance on other language tasks.
3
The findings provide computational evidence that predictive processing fundamentally shapes human language comprehension.
4
The neural predictions generalize across independent datasets and imaging modalities, including functional MRI and electrocorticography.
5
Transformer models predict nearly 100% of explainable variance in neural responses to sentences.
Research Object
human language processing in the human brain
Research Subject
the role of predictive processing in language comprehension mechanisms, including neural and behavioral responses to sentences
Publication Details
Publication Date
2021-11-04
Journal
Publisher
ISSN
Cited by
555
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai6
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer2019
Distributed Representations of Words and Phrases and their Compositionality2013
Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank2013
Whatever next? Predictive brains, situated agents, and the future of cognitive science2013
Preface: Cerebral Cortex Has Come of Age1991
HuggingFace's Transformers: State-of-the-art Natural Language Processing2019