Patterns of Temperament Development in Infants Who Develop ASD

Background: Studies suggest young children with autism spectrum disorders (ASD) may exhibit extremes of temperament early in development (Gilbert et al., 1990).  Prospective studies of infants at high risk for ASD are inconclusive, reporting increased passivity in HR infants at 6 months, less surgency, or no clear differences (Zwaigenbaum et al., 2005; Clifford et al., 2012; Bolton et al., 2012).  Research is needed to determine whether the standard constructs used to measure temperament in normative populations (i.e., the IBQ-R, Gartstein & Rothbart, 2003) are applicable to infants at risk for ASD. It is possible differences in IBQ factor scores may mask true differences in the temperament profile of young children at risk for ASD. 

Objectives: We aim to characterize temperament in a longitudinal cohort of infants at high risk for autism. We will evaluate whether standard measures of temperament adequately fit data from infants at high risk for ASD.

Methods: Two hundred and seventy-three infants with older siblings with ASD (HR) and 150 infants with typically developing older siblings (LR) were assessed at 6, 12, and 24 months as part of a larger, multi-site, study of brain and behavioral development in ASD, the Infant Brain Imaging Study. Temperament was assessed at 6 and 12 months using the IBQ-R.  Autism symptoms were assessed at 24 months using the ADOS.   Preliminary data from a subset of infants from one clinical site is reported here (HR n=45, LR n=12). The full sample will be used for the final analysis. We will use standard factor scores (Surgency, which encompasses approach behaviors, Regulatory control, and Negative Affect) and subscales from the IBQ-R to compare HR vs LR infants, and will use ADOS scores to form groups based on ASD symptoms. We will then perform a factor analysis to examine whether the standard three factor solution adequately fits data gathered from this large, HR infant cohort.

Results: Infants from the smaller subset were grouped according to 24 month outcome using ADOS cut off scores (HR ASD positive, HR ASD negative, LR ASD negative). Using one-way ANOVA, no differences were found in 6 month factor scores.  At 12 months, Surgency was significantly reduced for the HR ASD positive group,  F (2,57) =4.73, p<.05. At 6 months, subscales revealed group differences on Smiling and Laughter, F(2,56)= 4. 986 ,p< .01,Low Pleasure F(2,56)= 5.628p< .01, and Vocal Reactivity F(2,56)= 3.776, p<.03. At 12 months, group differences were found on Smiling and Laughter F(2,57)= 7.213, p<.01, and Vocal Reactivity, F(2, 57)= 7.479, p<.01. We will examine these differences in the larger group of infants and conduct factor analyses on this larger data set.

Conclusions: Preliminary analyses suggest that infants who develop ASD at 2 years of age may smile and vocalize less than HR ASD negative and LR ASD negative infants. Despite no factor-level differences at 6 months, subscale analyses do reveal group differences at this age. Future research is needed to investigate whether alternative measurement strategies may better characterize temperament in infants at high risk of ASD.