The time course of semantic and relational processing during verbal analogical reasoning

Abstract

Analogy is an important ability that allows humans to discover relationships between information domains that often vary in surface and relational characteristics. Cognitive neuroscience studies of analogy have demonstrated the importance of the prefrontal cortex during relational comparisons, but little is known about how semantic and relational similarity interact throughout its time course. We used scalp electroencephalography (EEG) analyzed with event-related potentials (ERPs) to examine the neural time course of analogical reasoning while 16 participants solved four-term verbal analogies. Semantic similarity was manipulated by increasing the semantic distance between source and target analogs creating semantically near and far analogies. Relational similarity was manipulated by creating relationally valid and invalid analogies. Only valid analogies were impacted by semantic distance such that far analogies were solved slower and less accurately than near analogies. Correctly solving near analogies elicited more positive waveforms at the N400 and during later relational processing. However, valid analogies elicited more positive signals during only later relational processing and not during the N400. These results suggest that semantic information impacts both early semantic and late relational comparison stages, while relational properties exert more influence in later stages of analogical reasoning. The degree of semantic similarity shared between knowledge domains demonstrated a potent effect throughout the time course of analogy that affected not only semantic access, but also the mapping of relational structures.