Project Details
Projekt Print View

Cross-modal integration of language and face cues in German and French infants

Subject Area Developmental and Educational Psychology
Term from 2011 to 2016
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 186144755
 
Final Report Year 2016

Final Report Abstract

It was the aim of our CrossMod ANR-DFG project to determine when the binding between auditory and visual facial cues emerges during infancy, how language-specific it is and whether it narrows down to the native language toward the end of the first year. The originally planned experiments were successfully carried out and have already resulted in a number of publications. We have examined the processing of audio-visual fluent speech information by German and French infants, with four series of experiments: (i) the first part of the project has shown that infants at 3.5 months of age cannot discriminate between French and German based on visual cues only, but that 6-month-old German infants can. Surprisingly, however, German 6-month-olds were shown to prefer looking at the French stimuli, while French 6-month-olds did not show any preference when presentation times were short (30 sec) but also present a preference toward the French stimuli when presentation times were prolonged (60 sec). (ii) Then we tested audio-visual integration of speech and showed that German infants as young as 4.5 months of age can cross-modally match audiovisual fluent speech of their native and non-native languages, and that this capacity narrows down to the infants’ native language by the age of 6 months. French infants tested at 3.5, 6, and 12 months were not found to show a multisensory response, which raises a number of questions on language specific and stimulus-specific effects on the development of audio-visual integration. (iii) The third part of the project tested the role of prosodic cues on the perception of audio-visual speech. Kubicek et al. (2014) compared infants’ response to infant-directed (ID) vs. adult-directed (AD) speech and showed that 12-month-old infants were not capable of audio-visual integration with AD speech whereas they were with ID speech. In another series of experiments, it was showed that audio-visual integration is hampered when the two languages tested are from the same rhythmic class (here: German and English, both stress-timed). Both results prove that infants rely on salient prosodic cues for audio-visual integration, and suggest that their audio-visual integration is mainly driven by rhythmic cues. (iv) Finally, we observed the effect of visual and audio-visual cues on face scanning by 12-month-old infants, and showed that for the native language audio-visual information induced more scanning of the eyes than visual information only, which draws more attention to the mouth. For the non-native language, infants looked more at the mouth region. In sum, the CrossMod project provided a number of original results on audio-visual integration of speech in infants.

Publications

 
 

Additional Information

Textvergrößerung und Kontrastanpassung