Establishing causal coherence across sentences: an ERP study.

Area of Science:

• Cognitive Neuroscience
• Psycholinguistics
• Computational Linguistics

Background:

• Understanding how readers establish causal relationships is crucial for discourse comprehension.
• Previous research has explored semantic and syntactic integration, but the role of causal inference in early processing stages remains less understood.

Purpose of the Study:

• To investigate the neural mechanisms underlying the establishment of causal relationships during online sentence comprehension.
• To determine if causal coherence influences early semantic processing, independent of lexico-semantic word relationships.

Main Methods:

• Event-related potentials (ERPs) were recorded while participants read three-sentence scenarios with varying degrees of causal relatedness (high, intermediate, unrelated).
• Latent Semantic Analysis (LSA) was used to match lexico-semantic co-occurrence across conditions.
• Participants judged the relatedness of the scenarios.

Main Results:

• A larger N400 component was observed for critical words in causally unrelated scenarios compared to highly and intermediately related scenarios.
• The N400 amplitude was attenuated for intermediately related sentence-final words, similar to highly related words at midline sites.
• No significant modulation of the P600 component was found across conditions.

Conclusions:

• Causal inferences, both simple and complex, significantly impact the earliest stages of semantic word processing.
• Causal coherence at the situational level influences incremental discourse comprehension, even when individual word semantics are controlled.
• These findings highlight the predictive nature of language comprehension and the integration of world knowledge during reading.