Brain oscillatory dynamics of lexical-semantic processing

Neuropsychological and imaging studies have shown that anterior and posterior brain areas are involved in the retrieval and selection of semantic representations, but it is not known how these areas dynamically interact during semantic processing. A candidate mechanism is through the synchronization of neuronal oscillations as they are likely involved in the creation of local and long-range functional neuronal networks. We hypothesized that low frequency coherence would reflect the coupling of anterior and posterior brain areas and examined the oscillatory dynamics of lexical-semantic processing using two paradigms, a semantic priming paradigm and a semantic ambiguity paradigm, which recruit this global network. For the first experiment, high density EEG was recorded while subjects read semantically-related or unrelated word pairs. Time-frequency analysis revealed facilitation of semantic processing for Related compared to Unrelated conditions, which resulted in shorter response times, reduced N400 amplitude, and reduced gamma (45-50 Hz) power from 150-450 ms. Moreover, reduced theta (4-7 Hz) coherence for Related compared to Unrelated conditions was observed between left anterior and posterior electrodes over the time windows 150-425 ms and 600-900 ms. We suggest that while gamma power reflects activation of local functional networks supporting semantic representations, the anterior-posterior topography of theta coherence indicates dynamic coupling of anterior and posterior areas for retrieval and post-retrieval processing of these semantic representations. This coupling appears much greater when retrieval demands are high and top-down retrieval is needed. In the second experiment, subjects read word pairs in which the first word biases the second, an ambiguous word, to either the dominant or subordinate meaning. When selecting the subordinate meaning over the prepotent, dominant meaning, selection demands are higher which slows response time and engages a similar anterior-posterior semantic network for top-down selection processing. We observed both increased reaction time and increased anterior-posterior theta coherence over 100-1000ms for the Subordinate condition compared with the Dominant condition. Thus we suggest that theta coherence reflects coupling within the anterior-posterior semantic selection network for selecting the subordinate meaning. These experiments provide converging evidence that theta coherence may serve as a mechanism for creating dynamic large-scale semantic networks.