Concurrent Associations between FYI Attentional Constructs and ADOS Scores in 12-Month High-Risk Siblings

Background: Deficits in social attention as well as the presence of atypical sensory attention constitute some of the core symptoms of ASD observed at an early age (Chawarska et al., 2014; Baranek et al., 2013). Current work has focused on identifying behavioral markers of ASD as early as 12 months (Reznick et al., 2006). Recently, researchers used the First Year Inventory (FYI), designed to identify risk for ASD at 12 months, to create new composite scores based on attentional constructs: responding to social attention (RSA), initiating social attention (ISA) and nonsocial sensory attention (NSA; Stephens et al., 2017). Their preliminary analyses revealed that the new algorithm scores predicted some variance in social responsiveness scores at 3 years. However, little research has examined these constructs in relation to concurrent ADOS scores in high-risk siblings.

Objectives: To examine if infants at high and low familial risk for ASD differ in mean scores on attentional constructs and to investigate concurrent associations between attentional construct scores and severity of autism symptoms indexed by the ADOS-Toddler.

Methods: Participants included infants at high-risk (HR) for ASD (N=122; M_age= 12.3 months) and at low-risk (LR) for ASD (N=70; M_age=12.3 months). At the 12-month visit, parents completed the FYI and clinicians administered the ADOS-Toddler. Differences in mean scores between HR and LR groups on the attentional constructs were tested. Stepwise regression analyses tested the predictive value of the attentional constructs on concurrent ADOS-Toddler Total scores, Social Affect (SA) scores and Restricted, Repetitive Behavior (RRB) scores.

Results: Independent t tests revealed that HR infants showed more deficits in ISA (M=1.90, SD=.60) than LR infants (M=1.64, SD=.55), t(190)=2.97, p<.01) and more deficits in RSA (M=1.57, SD=.35) than LR infants (M=1.44, SD==.33); t(190)=2.53, p<.05. HR infants also presented with more atypical NSA scores (M=1.62, SD=.34) compared to LR (M=1.78, SD=.43); t(190)=-2.86, p<.01. All three attentional constructs correlated significantly with concurrent ADOS-T Total scores, r=.38(p<.001), r=.31(p<.001), and r=.23(p=.001) for ISA, RSA and NSA, respectively. There were significant correlations between ADOS SA scores and ISA (r=.38, p<.001), RSA (r=.31, p<.001) and NSA (r=.20, p<.01). Stepwise regression analyses showed that only ISA contributed uniquely to ADOS SA scores (F(1,189)=31.763, p<.001), accounting for 14.4% of the variance. While there were significant correlations between ADOS RRB scores and NSA (r=.225, p=.001) as well as ISA (r=.18, p<.01), only NSA was significantly associated with ADOS RRB scores (F(1,189)=13.135, p<.01) accounting for 5.1% of the variance.

Conclusions: The attentional constructs derived from the FYI differentiate between HR and LR infants, with HR infants’ scores indicative of fewer social initiations and responses and more atypical sensory attention. Levels of responding to social attention was not related to severity of autism symptoms. However, parent-reported indices of initiating social attention as well as nonsocial sensory attention were directly related to ADOS algorithm scores. The initiating social attention construct may be pivotal in detecting early deficits in social attention and the NSA construct shows promise in capturing early RRB behaviors.