The development of auditory-visual speech perception across languages and age

To understand the now well-established auditory-visual nature of speech perception, it is necessary to understand how it develops. We know that young infants perceive speech auditory-visually by the fact that they perceive the auditory-visual illusion known as the McGurk effect; that visual information use increases over age in English-language children; and that Japanese-language adults use less visual information than do their English-language counterparts. Here we complete the developmental scene and probe the processes involved. In Experiment, with 6-, 8-, and 11-year-old and adult Japanese- and English-language participants tested on a McGurk task, while 6-year-olds from both language groups were equivalently influenced by visual speech information, there was a significant jump in auditory-visual speech perception between 6 and 8 years in English- but not Japanese-language participants. To in-vestigate this further, in Experiment 2 we gave English-speaking 5-, 6-, 7- and 8-year-olds and adults a McGurk ef-fect task as well as a language-specific speech perception (LSPP) test with native- and non-native speech sounds, and reading and articulation tests. For children, but not adults, visual-only speech perception (lipreading) ability and LSSP predicted McGurk performance" children with good auditory-visual speech perception tended to be those who focussed more on native than non-native speech sounds. In Experiment 3, with 3- and 4-year-olds tested for McGurk effect, LSSP, receptive vocabulary, and cognitive skill, regression analyses showed that auditory-only speech percep-tion and cognitive skill, but not LSSP, predicted auditory-visual speech performance. Together the results show that there is an increase in auditory-visual speech perception between 6 and 8 years in English- but not Japanese-language children, and in English-language children this is related to language specific speech perception processes specifically around that age (5, 6, 7, 8 years) and not before (3, 4 years) or after (adults). It is suggested that LSSP is most vari-able and most predictive of visual influence in speech perception in the presence of significant linguistic challenges, such as those at the onset of reading instruction.