A Longitudinal ANALYSIS of Maternal Infant Directed Speech to Preverbal, at-RISK for ASD, Infants

Background:  Developments in language acquisition research are coalescing on earlier and earlier stages of infant development as the crucial times and the critical importance of speech addressed to infants during the preverbal period is well established.  Mother’s vocal/verbal input and imitation rates are predictive of later language development in both typically and atypically developing children (Meltzoff 2002; Masur & Olson 2008) yet comparatively little is known about the input received in this early pre-verbal stage before 12 months (Gros-Louis 2007).  The acoustic and prosodic characteristics of maternal infant directed speech have been well documented but little systematic research has been carried out into the syntactical and grammatical aspects of that speech input. 

Objectives:  To compare maternal Infant Directed Speech characteristics in face-to-face interaction with infants at genetic risk of autism with maternal speech to typically developing low-risk controls. 

Methods:  This paper presents one set of findings from a longitudinal study of mother-infant interaction over the course of a nine month period. Nineteen mother-infant dyads participated, of whom nine infants, including twins, are at genetic risk for autism spectrum disorder.  Using a prospective observational design, dyads were videorecorded in naturalistic face-to-face interaction every 4 weeks for a period of nine months when the infant was aged between 3 and 15 months.  142 five-minute samples of maternal speech and vocalisations were selected and transcribed and approximately 10,000 utterances were analysed on a range of measures of quantity, quality and complexity.

Results:  Results indicate differences in the linguistic environments of the high-risk infants in interaction, in terms of quantity of speech input, lexical diversity and clausal complexity of linguistic input.  Although there were few statistically significant differences for most characteristics, given the sample size, on many of the measures mothers of at-risk infants could be differentiated from mothers of TD infants.   The mothers of at-risk infants are consistently making additional modifications to their speech when interacting with their infant.

Conclusions:  Two conclusions are drawn.  First, typical patterns of maternal input and response are disrupted by the high-risk status of the infants, evidenced, for example, by increased use of compensatory, attention-getting, vocal but non-meaningful, linguistic devices, zero clause utterances, and very frequent use of the infant’s name.  This is at the expense of more complex and diverse linguistic constructions.  These attention getting-and-keeping devices may be a function of the mother’s heightened awareness of the infant’s risk status and her interaction with the older sibling with ASD.  It has been suggested (Johnson 2010) that this persistence in relation to engaging the infant in interaction may serve to promote joint attention development and this will be analysed in relation to the infants’ performance on the ESCS at 12 and 18 months.  Second, a subset of the high-risk infants themselves can be identified as behaving differently in interaction and are providing fewer feedback opportunities for mother to create tailored linguistic environments over time.