Seeing the Speaker's Face Enhances Second Language Shadowing: Neural and Behavioral Evidence

Abstract This functional magnetic resonance imaging (fMRI) study investigated how facial cues influence second language (L2) shadowing among 42 Japanese learners of English. Participants completed four conditions that varied by task type (listening vs. shadowing) and visual input (face vs. mosaic). Behaviorally, shadowing reproduction accuracy was higher when the speaker's face was visible. Neuroimaging revealed greater activation in the left posterior middle temporal gyrus (pMTG), right ventral pallidum, and left hippocampus during shadowing in the Face condition, reflecting enhanced audiovisual integration, engagement, and memory‐related processing. Learners with higher oral proficiency exhibited increased activation in speech‐integration areas, such as the posterior superior temporal gyrus, and those with higher listening proficiency showed reduced cerebellar engagement, suggesting proficiency‐dependent neural strategies for integrating facial cues during shadowing. These findings support embodied, multisensory, and socially grounded accounts of L2 learning, emphasizing the pedagogical importance of visible facial cues. Incorporating face‐based shadowing into L2 learning may help bridge perception and production and prepare learners for interaction‐oriented communication.