International Congress of Infant Studies, V-ICIS

Glasgow (Online)
6-9 July 2020

Poster: Testing the role of audio-visual integration in infant vocal imitation in a naturalistic setting.
Bastianello, T., Majorano, M., Vihman, M.M.

Early word learning can be explained by a two-way mechanism involving attention to the auditory-source and articulatory representations (the “dual-stream-model”: Hickok, & Poeppel, 2007). In line with the intersensory redundancy hypothesis (Bahrick & Lickliter, 2000), infant ability to integrate different cues from a talking face is related to vocal imitation (in particular, time spent looking toward the mouth: see Imafuku, Kanakogi, Butler, & Myowa, 2019). Several studies have shown this ability experimentally, but none have so far tested for effects in the course of naturalistic adult-child interaction. The aim of the present study is to assess children’s looking behaviour during a semi-structured vocal imitation task, using a camera mounted on the mother’s head to record the baby’s eye movements, in the home, in response to words and non-words. The relationship between the child’s ability to imitate the mother’s vocal patterns was also analysed. The study involves 12 infants in each of two age groups: (i) 12- month-olds, (ii) 15-18-month-olds. Each mother was first asked to complete the Italian short version of the McArthur-Bates Communicative Development Inventory (MB-CDI; Caselli, et al., 2017); this provided the basis for choosing the stimuli for each infant. Each mother then participated with her child in a vocal imitation task including four disyllabic words or non-words referring to four different objects. Each item was repeated six times. The stimuli consist of 1) one word (F , a word that the child is reported to comprehend and produce) 2) one familiar word (F-, a word that the child is reported to comprehend but not to produce) 3) one segmentally familiar non-word (F , a non-word including patterns that the child is able to produce, as reported by the mother) 4) one segmentally unfamiliar non-word (F-, a word that includes patterns that the child is not able to produce, as reported by the mother). All play sessions were video-recorded, and the imitations were analysed (i.e., any child attempt to produce a target form that includes at least two sounds). The child’s looking time toward the mother’s mouth was analysed (see Figure 1, based on the six children per age group analysed to date). The number of child vocal imitations in each condition was tallied (words, M=3.50, SE=.69; non-words, M= 1.67, SE=.88). A t-test showed a significant difference between the vocal imitations in the word- vs. non-word conditions (W = 53.50, p < .009). Correlational analysis showed that looking time towards the mother’s mouth in the non-word condition is positively related to the number of vocal imitations for that non-word (F : r = .620, p < .05 and F -: r = .733, p <.01, respectively). In sum, the results support the role of the mouth as a visual cue in the child’s vocal imitation between the ages of 12 and 18 months. This imitative behaviour supports the suggested role of audio-visual integration for the child’s word-learning process.