Measuring Meaning Composition in the Human Brain with Composition Scores from Large Language Models
This provides a novel tool for neurolinguistics to quantify meaning composition, addressing a gap in computational metrics for brain research.
The paper tackled the lack of a computational metric for meaning composition in human sentence comprehension by introducing the Composition Score, a model-based metric derived from transformer feed-forward networks, and found it correlates with brain clusters related to word frequency, structural processing, and word sensitivity.
The process of meaning composition, wherein smaller units like morphemes or words combine to form the meaning of phrases and sentences, is essential for human sentence comprehension. Despite extensive neurolinguistic research into the brain regions involved in meaning composition, a computational metric to quantify the extent of composition is still lacking. Drawing on the key-value memory interpretation of transformer feed-forward network blocks, we introduce the Composition Score, a novel model-based metric designed to quantify the degree of meaning composition during sentence comprehension. Experimental findings show that this metric correlates with brain clusters associated with word frequency, structural processing, and general sensitivity to words, suggesting the multifaceted nature of meaning composition during human sentence comprehension.