Representation and processing of multi-word expressions in the brain.

Area of Science:

• Cognitive Neuroscience
• Psycholinguistics
• Computational Linguistics

Background:

• Language comprehension relies on predicting upcoming information for efficient communication.
• While discourse-level prediction is studied, phrase-level prediction remains under-explored.
• Binomials represent predictable multi-word expressions crucial for language processing.

Purpose of the Study:

• Investigate the neural mechanisms of phrase-level prediction in English.
• Examine event-related potentials (ERPs) during the comprehension of binomials.
• Differentiate processing of predictable phrases from infrequent associates and semantic violations.

Main Methods:

• Experiment 1a: Participants read binomials (e.g., 'knife and fork'), associated phrases, and semantic violations.
• Experiment 1b: Participants read the same stimuli but without the conjunction 'and'.
• Measured event-related potentials (ERPs) to analyze brain responses during comprehension.

Main Results:

• Experiment 1a showed binomials elicited larger P300s and smaller N400s compared to other conditions.
• P300 amplitude suggests template activation for predictable information.
• N400 reduction indicates easier semantic integration for binomials.

Conclusions:

• Distinct neural mechanisms support the processing of predictable versus novel word sequences.
• The presence of 'and' in binomials is critical for observing enhanced processing.
• Findings highlight the role of phrase predictability in language comprehension.