ИСТИНА

Speech disfluencies are an important component of natural discourse production and comprehension that may play a role of signals controlling the communication flow. However, the neural circuitry underlying the speech disfluencies is almost an unknown territory. In our recent pilot fMRI study (Smirnova et al., in press) we explored perception of four types of speech disfluencies and their clusters annotated in the “Russian Pear Chats and Stories” corpus: silent pauses, filled pauses, lengthenings (phoneme prolongation as hesitation marker), and self-repairs, or breaks. The present study continues the previous work with a larger sample (N=30) and a focus on neural correlates in connectivity rather than activation domain. Using data-driven approach, we compared the intrinsic connectivity contrast between speech disfluencies of different type and fluent speech. Functional connectivity changes were revealed for three conditions that implicated lengthening alone or in different combinations with other types of speech disfluency and different portions of the left parietal cortex. Our results indicate the promising targets for further studies of the functional connectivity and neural networks in speech disfluency perception. From a broader perspective our study demonstrates how corpus linguistics and neuroimaging may join their efforts for interdisciplinary research of brain and language.