Longitudinal Stability of Quantitative Autistic Traits in Toddler Twins

Thursday, May 12, 2016: 2:21 PM
Room 308 (Baltimore Convention Center)
N. Marrus1, Y. Zhang2, A. Glowinski3, T. Jacob4, S. Kennon-McGill3, S. Sant5, T. Gray5, A. Haider2 and J. N. Constantino2, (1)Washington University School of Medicine, Webster Groves, MO, (2)Washington University School of Medicine, Saint Louis, MO, (3)Washington University in St. Louis, St. Louis, MO, (4)Family Research Center, VA Palo Alto Health Care System, Menlo Park, CA, (5)Washington University School of Medicine, St. Louis, MO

The ability to measure the early trajectory of core autistic traits has significant implications for elucidating developmental mechanisms of heterogeneity in ASD and tracking incremental responses to early interventions.  Previously (Marrus et al., J Child Psychol Psychiatry 2015), we demonstrated that a video-referenced rating of Reciprocal Social Behavior (vr-RSB), a novel toddler measure in which caregivers rate their child’s level of RSB compared to a typically developing video anchor, shows strong psychometric properties including a continuous, unimodal distribution of autistic trait severity (as indexed through levels of RSB), high heritability, excellent test-retest reliability, and marked impairments in RSB in toddlers with ASD.


To investigate in a general population sample whether quantitative autistic traits (QAT) measured using the vr-RSB at age 18 months predict variation in QAT at age 36 months, the latter measured using the preschool version of the Social Responsiveness Scale (SRS), which has been extensively validated in prior autism research (Baranek et al., Autism 2013; Duku et al., J Autism Dev Disord 2013).


Parents of 252 epidemiologically representative toddler twins [monozygotic (MZ)=31 pairs, dizygotic (DZ)=95 pairs] participated in the Early Reciprocal Social Behavior Study (ERSB), a longitudinal study of early social development. Parents rated their 18-month-old twins on the vr-RSB, and subsequently completed the preschool SRS on their children at age 36 months. Pearson’s correlations were performed for total RSB scores at 18 and 36 month time points, as well as sub scale scores for social communication (SC) and restricted, repetitive behavior (RRB). Comparisons of intraclass correlations (ICCs) for MZ vs. DZ twins provided estimates of heritability at each developmental juncture.  An exploratory principal components analysis (PCA) using a varimax rotation was conducted at each time point, based on common items across the vr-RSB and SRS-2.


Total RSB scores were strongly correlated between 18 and 36 months, r=.666, p<.01 (Fig. 1).  Similarly high correlations were observed for SCI and RRB scores across the two time points as well. ICCs for MZ twins were greater than DZ twins for total RSB scores, SC, and RRB scores at 36 months (Table 2).  PCA of 18 month data demonstrated a first principal factor accounting for 19% of the variance which mapped well to the principal factor derived from PCA of 36 month data.


Levels of RSB, and by extension autistic traits, demonstrated strong longitudinal stability in a general population sample over the critical period from 18 to 36 months, when interventions typically begin for children affected by ASD. The ability to serially implement reliable quantitative trait measurements during this period offers the opportunity to track nuanced changes in core features of the autistic syndrome over time and in response to intervention.