Predictive entrainment of natural speech through two frontomotor top-down channels
Hyojin Park1,2, Robin, A.A. Ince1, Gregor Thut1, Joachim Gross1,3; 1Institute of Neuroscience and Psychology, University of Glasgow, Glasgow, United Kingdom, 2School of Psychology, University of Birmingham, Birmingham, United Kingdom, 3Institute for Biomagnetism and Biosignalanalysis, University of Munster, Munster, Germany
Natural communication between interlocutors are enabled by the ability to predict upcoming speech inputs at a given context. In our previous studies, we showed that low-frequency brain oscillations in delta and theta bands in auditory and temporal areas track external auditory speech envelope when speech is intelligible compared to when it is not, and this mechanism was modulated by the same frequency activity in inferior frontal and motor cortices by top-down manner. However, the direct relationship between these frontal top-down signals and speech inputs still remains to be understood, particularly its temporal aspect of when these top-down signals predict upcoming speech streams that is to be received by primary sensory cortex. We obtained MEG data while participants were listening to intelligible and unintelligible continuous speech. By applying transfer entropy analysis, we identified low-frequency delta/theta (<7 Hz) phase information in left inferior frontal cortex predicts the same frequency phase information of upcoming speech signal with a time lag of 200-250 ms when speech is intelligible. Furthermore, beta (19-30 Hz) power in left supplementary motor area and superior/middle frontal cortex predicts delta/theta phase information of upcoming speech signal with the same time lag (200-250 ms). Interestingly, this temporal characteristic matches a typical turn-taking gap in natural human communication in which its temporal property represents one of important aspects of predictive comprehension of speech. This study identifies that top-down modulation from frontomotor areas is temporally structured at predicting upcoming streams of intelligible speech and this mechanism is implemented through two different frequency channels.
Topic Area: LANGUAGE: Other