The Early Language Environment and 9 Month “Hyper-Vocalizing” in Infants at Risk for Autism

Background:  Studies on normative development have shown the substantial impact of the early language environment on a child’s later cognitive development (Hart & Risley, 1995; Ramirez-Esparza et al., 2014). Previous research on infants at-risk for ASD have been inconsistent with some reporting high-risk infants vocalizing less in the first year of life when compared to low-risk infants (Paul et al., 2010), while others report no group differences (Talbott et al., 2015; Northrup & Iverson, 2015). However, methodology has been limited to videotaped interactions of parents and infants playing. 

Objectives:  The goal of the current project is to extend prior literature by examining the early home language environment of 9 month old infants at high and low familial risk for ASD using naturalistic day-long in-home recordings. 

Methods:  This study includes data from 41 infants at high-risk (HR) for ASD (66% male), and 20 infants low-risk (LR) for ASD (50% male). As part of a larger ongoing longitudinal study, the Infant Brain Imaging Study, 9-month old infants wore LENA speech recorders for 2 full days in the home (yielding 32 hours of recording). The number of adult words, child vocalizations, and adult-child conversational turns were calculated. Infants and their families were also assessed during 4 clinic visits where infants were assessed using the Mullen: an intake assessment at 3 or 6-months, two randomly assigned follow-up assessments (6, 9, 12, or 15 months), and a 24-month outcome visit. Data collection is ongoing hence this analysis focused on the 12 and 15 month time-points. 

Results:  LR and HR infants did not differ on chronological age at 9-month LENA recording (Figure 1). The groups did significantly differ on Mullen early learning composite scores and receptive language scores at 12-15 months, but not on other Mullen subscales. LR and HR infants significantly differed on number of child vocalizations with HR infants vocalizing at a higher rate, t(59)=2.76, p=.007. The groups did not significantly differ in the number of adult words or conversational turns at 9 months. Given that previous studies have found maternal education and sex of the child to contribute to emerging language skills, additional analyses were run to control for these variables. We also controlled for Mullen composite. Results revealed a significant main effect of maternal education, F(46)= 2.99, p=.05, and group, F(46)= 4.16, p=.04, but not sex of the child or Mullen composite. This significant group difference in vocalization rate appears to be driven by a subgroup (25%) of HR “hyper-vocalizing” infants who vocalized at an extremely high rate (2 SDabove the mean of LR infants, red diamonds in Figure 1).

Conclusions:  We found that infants at HR and LR for ASD experience strikingly similar language environments in regards to adult words and adult-child conversational turns. However, we uncovered significant group differences in child vocalizations. This difference was driven by a subgroup of HR hyper-vocalizers who vocalized at a high rate. We will explore whether 9 month hyper-vocalizing is stereotypic in nature and if it is predictive of later ASD diagnosis.