Characterizing Directional Dynamics of Semantic Prediction Based on Inter-regional Temporal Generalization.

Area of Science:

• Neuroscience
• Cognitive Science
• Psycholinguistics

Background:

• The N400 component (N400m) is a key neural marker for semantic prediction in language.
• Top-down signaling from the inferior frontal gyrus (IFG) to perceptual areas is theorized to generate the N400m.
• Estimating causal connectivity for this process has been methodologically challenging.

Purpose of the Study:

• To investigate the causal role of the IFG in semantic prediction during speech comprehension.
• To test a predictive model of speech processing involving feedback from IFG to temporal regions.
• To explore dynamic information flow between brain areas during semantic prediction.

Main Methods:

• Utilized magnetoencephalography (MEG) data from 21 participants performing a semantic prediction task with German sentences.
• Implemented a machine learning approach with temporal generalization using a support vector machine (SVM) classifier.
• Trained and tested SVM classifiers on activity from the IFG and superior/middle temporal gyri (STG/MTG) to assess information flow and causality.

Main Results:

• Significant decoding accuracy was observed in a bottom-up model (STG/MTG to IFG).
• Crucially, decoding accuracy also significantly exceeded chance levels when classifiers trained on IFG activity predicted subsequent STG/MTG activity.
• This demonstrates a causal influence of IFG on temporal regions during semantic prediction.

Conclusions:

• Findings support a predictive model of speech comprehension with feedback from IFG to temporal areas.
• Evidence for dynamic, bidirectional (top-down and bottom-up) information flow between IFG and temporal regions during semantic prediction.
• This study advances our understanding of the neural mechanisms underlying predictive language processing.