Predictive Brain Activity Shows Congruent Semantic Specificity in Language Comprehension and Production

Area of Science:

• Neuroscience
• Cognitive Science
• Psycholinguistics

Background:

• Language processing involves predicting upcoming words.
• Prediction potentials (PPs) are slow waves observed before word onset, reflecting semantic anticipation.
• Existing models propose specific neural mechanisms for semantic prediction.

Purpose of the Study:

• To investigate if semantic PPs differ between language comprehension and production.
• To determine if PPs arise from localized areas or distributed semantic circuits.
• To test current predictive processing models in language.

Main Methods:

• Collected 128-channel EEG data from male and female participants.
• Analyzed slow wave activity (PPs) during sentence comprehension and production tasks.
• Examined cortical source localization of semantic PPs.

Main Results:

• Larger slow waves (PPs) were observed before predictable words in both comprehension and production.
• Cortical sources of semantic PPs were distributed and category-specific.
• Anticipation of animal words engaged posterior visual areas; tool words engaged frontocentral sensorimotor areas.
• PP size correlated with cloze probability and decreased with shorter sentence completion latencies.

Conclusions:

• Semantic prediction in language involves distributed, category-specific neuronal circuits.
• These circuits are active in both language comprehension and production.
• Findings support theories of distributed semantic processing underlying prediction.