Language Prediction Is Reflected by Coupling between Frontal Gamma and Posterior Alpha Oscillations

Area of Science:

• Neuroscience
• Cognitive Science
• Psycholinguistics

Background:

• Readers and listeners actively predict upcoming words during language processing.
• These predictions are crucial for integrating words into sentence context, involving language network interactions.

Purpose of the Study:

• To investigate the neural mechanisms underlying predictive language processing.
• To examine brain activity differences based on sentence predictability using magnetoencephalography (MEG).

Main Methods:

• Participants read sentences with varying contextual constraints, manipulating word predictability.
• Magnetoencephalography (MEG) was used to record brain activity.
• Analyzed alpha and gamma power in specific brain regions, including the left inferior frontal cortex, left posterior temporal region, and visual word form area.

Main Results:

• Higher alpha power suppression was observed in the left inferior frontal cortex, left posterior temporal region, and visual word form area for highly predictable sentences compared to low predictability sentences.
• A negative correlation was found between alpha power in temporal language areas and gamma power in the left frontal cortex for highly constraining sentences.

Conclusions:

• The findings suggest that decreased alpha power in temporal language areas and increased gamma power in the left frontal cortex reflect the initiation of anticipatory unification during predictive language processing.
• This neural activity supports the integration of predicted words into the ongoing sentence context.