Comparing Severity of Restricted and Repetitive Behaviors Between High Risk Baby Siblings with Different Temperament Profiles

Friday, May 12, 2017: 5:00 PM-6:30 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
J. Chen, J. D. Burke, M. Barton and D. A. Fein, Psychological Sciences, University of Connecticut, Storrs, CT

Early self-regulatory behavior is a key component of temperament (Rothbart, Posner, & Kieras, 2006). High risk baby siblings with ASD demonstrate early signs of repetitive behaviors, as well as difficulties with self or other regulation (Bryson et al., 2007). Poor behavior regulation also predicts repetitive behaviors in high functioning school-aged children with ASD (Boyd, et al., 2009). The current study further explores the relationship between early behavior regulation and repetitive behaviors in high risk baby siblings who screened positive for ASD.


To compare severity of restricted and repetitive behaviors (RRBs) in high risk baby siblings with distinctive temperament profiles at age 2.


Baby siblings of children with ASD were screened for autism using the M-CHAT(-R) at age 2. Parents of children who screened positive completed the Toddler Temperament Scale (TTS) and Repetitive Behavior Scale – Revised (RBS-R) during a diagnostic evaluation. Mean scores on the Activity, Rhythmicity, Approach, and Distractibility subscales of the TTS (with acceptable internal reliability; i.e., Cronbach’s α ≥ 7; Chen, Barton & Fein, 2016) were dichotomized to contrast the top quartile against the remainder, and entered into a latent class analysis. Model fit indicators suggested a two-class solution as the best model. The two latent classes were subsequently labeled Poorly Regulated and Attuned (PRA) and Well Regulated and Attuned (WRA; see Figure 1). The PRA group was relatively more likely to endorse high activity, arrhythmicity, and poor approach, and less likely to endorse high distractibility. Chi-square tests were used to compare demographics between the PRA (n = 87) and WRA (n = 40) groups. Logarithmic or square root transformations of RBS-R Total and subscale scores were performed to address skewness in the data. Diagnostic group (ASD, non-ASD) was entered into a hierarchical linear regression before entering the dummy-coded latent classes as predictors of RBS-R scores.


Chi-square tests indicated no significant differences in age or gender between the PRA and WRA groups. The PRA group had a significantly higher proportion of baby siblings diagnosed with ASD. Logarithmic transformations of the RBS-R Total, Stereotyped Behavior, and Sameness Behavior scores successfully addressed data skewness. Transformations of the other domain scores were unsuccessful due to severe positive skew. Linear regression results indicate a significant improvement in variance predicted by temperament latent class above variance predicted by diagnostic group alone for RBS-R Total Score (p < .001, ΔR2 = .118), Stereotyped Behavior (p < .001, ΔR2 = .100), and Sameness Behavior (p < .001, ΔR2 = .083).


High risk baby siblings who screen positive for ASD at age 2, and who demonstrate different temperament profiles in early patterns of behavioral regulation and attunement to others, differ significantly in the severity of their RRB symptoms. More severe RRBs were associated with high activity, poor rhythmicity, poor approach, but less distractibility, suggesting that perseverative attention may contribute to RRBs. These differences exist above and beyond having a diagnosis of ASD. Early regulatory behaviors may provide important information about ability to regulate maladaptive repetitive behaviors.