Young children rely on visual information to process degraded speech: Evidence from behavioural and neuroimaging measures

Young children acquire language in environments where speech is often acoustically degraded, yet little is known about how developing brains adapt to reduced speech intelligibility. Using a combination of eye-tracking and functional near-infrared spectroscopy (fNIRS), we investigated young childrens attentional allocation to a speaking face at varying levels of speech intelligibility and the brain activity supporting this behaviour during development. Infants (8-10 months) and toddlers (27-30 months) viewed videos of a speaker in three conditions: producing clear speech, spectrally degraded (vocoded) speech, and silent (audio muted) speech. Visual attention to the speakers mouth increased when speech was degraded relative to clear speech in both age groups, indicating an early-emerging compensatory strategy. However, this shared behavioural response was supported by brain activity that differed by age. Degraded speech elicited greater recruitment of prefrontal regions associated with effortful listening, particularly in infants, whereas toddlers showed stronger engagement of posterior temporal regions implicated in audiovisual integration. In response to silent speech, there was no evidence for increased visual attention to the mouth compared to the clear speech condition, but there was reduced temporal activation and increased prefrontal brain responses, especially in infants. Together, these findings suggest that experience with audiovisual correspondences and linguistic maturity contribute to a more efficient processing of speech, particularly relevant when speech is degraded. By combining behavioural and neuroimaging measures, this study advances mechanistic accounts of audiovisual speech processing and provides insights relevant to populations experiencing spectrally degraded input, such as children using cochlear implants.