Calendrier

Abstract:

Embodied theories of cognition consider language as grounded in the sensorimotor system. Converging evidence shows that speech perception induces activations of sensorimotor brain areas that are involved in speech production. Notably, motor activity is elicited during both native and non-native phoneme perception, and it is somatotopically organized depending on the place of articulation of native phonemes. Specific motor activity is elicited depending on distinct articulatory features: listening to bilabial and dental consonants activates the cortical motor representations of the lips and the tongue, respectively. However, some studies have not replicated this precise somatotopic mapping within the motor cortex for native speech perception, while others have shown premotor recruitment only when speech sounds are degraded yet identifiable. Thus, the necessity and role of motor activation in processing both native and non-native, as well as clear and degraded, speech sounds remain uncertain. When speech in the native language is distorted by noise, or when sounds come from a foreign language, motor regions seem to support speech perception by retrieving articulatory features that are grounded in the speech production system. Given this motor contribution, the question arises of whether training that engages motor regions might enhance the learning and processing of non-native phonemes.

This thesis aims to explore sensorimotor processing in speech perception through an experimental study using functional magnetic resonance imaging (fMRI) and a review of foreign language learning paradigms with sensorimotor training. In the experimental study, we combined behavioral and fMRI measures and conducted Multivariate Pattern Analyses (MVPA) to investigate how sensorimotor regions are activated as a function of, and encode the phonetic features of native and foreign phonemes under optimal and degraded perceptual conditions. We show that neural activity for lip and tongue articulatory movements in the precentral gyrus predicts neural activation for bilabial and dental degraded native consonants, thus highlighting somatotopic coding of articulatory features in the motor cortex. Moreover, sensorimotor neural patterns associated with the perception of native and non-native phonemes reflect the phonetic similarity between speech sounds both within and between the two language repertoires. In the review article, we first outline current findings on speech processing in the context of embodied cognition, highlighting the functional contribution of the motor cortex to speech perception. We discuss the contribution of sensorimotor activity to foreign language learning, especially for phonemes, through gestural training that strengthens the embodiment of foreign articulatory features. Finally, we propose potential developments in training paradigms and directions for future neuroimaging studies to address existing gaps in the literature.

This thesis therefore contributes to ongoing discussions on motor resonance in native speech perception, especially under challenging conditions, as well as addresses gaps in understanding the motor contribution to non-native phoneme perception. Our fMRI study provides solid empirical evidence that the motor system, in conjunction with the auditory system, is fundamental to speech perception. We corroborate the sensorimotor nature of speech processing, both for native and non-native languages, thus opening pathways for advancements in foreign language learning. By offering an overview of sensorimotor training paradigms, together with neuroimaging evidence for central sensorimotor processes in speech perception, our research lays a foundation for future investigations exploiting the embodied nature of language processing.

Keywords: Speech Perception, Motor Cortex, Sensorimotor Integration, Phonetic Features, Foreign Phonemes, fMRI