Audiovisual Speech Integration in Infancy: Evidence from the McGurk Effect in a Perception-Based Behavioural Task

Abstract

Multisensory speech perception plays a critical role in language acquisition and later socio-cognitive and socio-emotional development. Recent evidence suggests that infants are able to integrate auditory and visual speech information early in life, and that this integration capacity is modulated by early visual experience with the familiar-race faces. Using the McGurk effect as an index of visual speech influence on auditory perception, we tested infants aged 6 to 12 months to examine (1) whether they could integrate audiovisual speech cues and whether this capacity strengthens with age, and (2) whether this integration is modulated by face-race familiarity. Infants participated in a perception-based behavioural task modeled on the Stimulus-Alternation Preference Procedure (SAPP; Best & Jones, 1998), and their looking times in response to McGurk and non-McGurk syllable pairs enacted by own-race and other-race faces were recorded using eye-tracking. When viewing own-race faces, infants demonstrated a robust and stable audiovisual integration capacity during the tested age range. However, other-race faces significantly disrupted this capacity from 6 months onward, indicating an early-emerging other-race effect (ORE). These findings support the view that speech is presented bimodally and processed in an integrative manner from early in life, and that perceptual tuning is a cross-modality, pan-sensory phenomenon. They further contribute to the argument that multisensory perceptual development involves a regressive reorganization that calibrates infants’ perceptual systems to the most ecologically relevant information.