Lizarazu, M. 1 , Lallier, M. 1 , Bourguignon, M. 1 , Carreiras, M. . 1, 2 & Molinaro, N. 1, 2
1 BCBL
2 Ikerbasque
Speech comprises hierarchically organized rhythmic components that represent prosody (delta band), syllables (theta band) and phonemes (gamma band). During speech pre-processing steps, neural oscillations within a fronto-temporo-parietal network track these quasi-rhythmic modulations through different mechanisms (de-multiplexing and segmentation steps). Speech processing models associate neural computations in temporal regions to perceptual processes, while operations in frontal-parietal regions are linked to higher-order processes (attention). However, there is no comprehensive view of the neural dynamics that allow perceptual and attentional processes to interact before extracting meaning from speech. In the present study, we analyzed MEG data from 20 participants while hearing continuous speech. First, we determined how different brain areas within the fronto-temporo-parietal network deal with the de-multiplexing (Coherence analysis) and the segmentation (Phase-Amplitude -Coupling analysis) pre-processing steps. Then, we showed how low-frequency (delta and theta band) neural oscillations bidirectionally connect (Transfer Entropy analysis) temporal and fronto-parietal areas. Our results indicate that, during speech processing, cortical oscillations represent an ideal medium to deal with perceptual and attentional neural computations, as well as to control interactions between these cognitive operations. Understanding the oscillatory mechanisms underlying speech pre-processing could help to better characterize language developmental disorders, e.g. dyslexia and specific language impairment (SLI).