Theta band (4~8 hz) oscillations reflect syllables processing in chinese spoken word production

Abstract

Languages may differ in the proximate units of phonological encoding in spoken word production. It has been demonstrated that syllables are proximate units of phonological encoding in Chinese speech production. Previous studies report that the θ band oscillations has been associated with syllables processing in language comprehension, however, it remains unknown what are the neural oscillations for syllables retrieval in speech production. The present study aims to investigate the neural oscillations of syllables retrieval at the stage of phonological encoding in Chinese spoken word production. We employed a masked priming paradigm and electrophysiological signals were recorded concurrently. In the task, participants were instructed to name pictures with disyllabic words which were preceded by briefly presented and masked prime words. Prime words were syllabically or phonemically related to the first syllable of targets or were unrelated, and the experimental design includes prime type (syllable overlapped vs. phoneme overlapped), relatedness (related vs. unrelated), and repetition (first vs. second). Behavioral data analysis showed a significant triple interaction among prime type, relatedness and repetition. In the first repetition, naming latencies were faster in syllabically related than unrelated condition, whereas latencies were longer in the phonemically related than unrelated condition. Time-frequency analysis also showed a significant triple interaction in the time window of 300~600 ms after pictures onset. Specifically, theta band power was lower for syllabically related than unrelated while no significant differences between phonemically related and related, in the first repetition. Cluster based permutation test showed that, in the first repetition, syllabically related condition elicited lower θ band power than unrelated in the time window of 270~460 ms, while phoneme relatedness produce marginally higher θ band power than phoneme unrelatedness in the 340~390 ms time window. These effects were absent in the second repetition. In sum, we found that syllable priming effect was reflected by the decrease of θ oscillation in spoken word production in Chinese. Time-frequency analysis also revealed an early syllable priming effect, and a late phonemic inhibition effect in Chinese spoken word production, which provides evidence for the proximate units principle.