Article révisé par les pairs
Résumé : Low integration of speech sounds with the mouth movements likely contributes to language acquisition disabilities that frequently characterize young autistic children. However, the existing empirical evidence either relies on complex verbal instructions or merely focuses on preferential gaze on in-synch videos. The former method is clearly unadapted for young, minimally, or nonverbal autistic children, while the latter has several biases, making it difficult to interpret the data. We designed a Reinforced Preferential Gaze paradigm that allows to test multimodal integration in young, nonverbal autistic children and overcomes several of the methodological challenges faced by previous studies. We show that autistic children have difficulties in temporally binding the speech signal with the corresponding articulatory gestures. A condition with structurally similar nonsocial video stimuli suggests that atypical multimodal integration in autism is not limited to speech stimuli