Neural Correlates of Semantic Prediction and Resolution in Sentence Processing

Area of Science:

• Neuroscience
• Cognitive Science
• Linguistics

Background:

• Most brain imaging studies examine language comprehension after stimuli presentation.
• Predictive mechanisms are crucial for brain functions, including language processing.
• Direct brain measures of specific semantic predictions are scarce.

Purpose of the Study:

• To investigate neurophysiological indices of semantic prediction before critical word presentation.
• To explore how sentence context influences semantic predictions.
• To examine the neural basis of predicting specific word meanings and sentence types (affirmative vs. negated).

Main Methods:

• High-density electroencephalography (EEG) in adult human participants.
• Analysis of brain activity preceding critical words in varying sentence contexts.
• Source localization of neurophysiological signals to identify brain regions involved.

Main Results:

• A neurophysiological index, the semantic readiness potential (SRP), reflects context-induced semantic predictions before word onset.
• SRPs were observed in high-constraint contexts but not low-constraint contexts.
• Specific semantic predictions were localized to motor system areas (dorsolateral for hand-related, ventral for face-related words).
• Negated sentences showed distinct activation patterns in medial prefrontal and motor areas, suggesting computation of alternatives.
• The N400 component showed differential effects for affirmative versus negated high-constraint contexts.

Conclusions:

• Semantic predictions are reflected in brain activity prior to word recognition.
• The brain exhibits fine-grained predictive processing, distinguishing between specific semantic features and sentence structures.
• These findings illuminate the interplay between predictive semantic computation and meaning resolution in language understanding.