Neural and Behavioral Predictors of Friendship in Children with Autism Spectrum Disorder

Background: Peer friendships are fundamental for social and affective functioning throughout the lifespan. In a meta-analysis, Mendelson et al. (2016) proposed a model of friendship success in children with autism spectrum disorder (ASD): efficient social information processing speed (SIPS), social cognition, and social motivation may underlie increased friendship quality and quantity and improve psychosocial functioning in individuals ASD who are able to form and maintain successful friendships. The present study utilized N170 latency to dynamic mutual eye contact as an index of SIPS. We hypothesized SIPS, social cognition, social motivation, and psychosocial functioning would relate to friendship (interest, quantity, and quality) in children with ASD.

Objectives: To empirically investigate and extend a model of successful friendship in ASD by examining the behavioral and neural correlates of friendship.

Methods: Children with ASD and average intelligence participated in a gaze-contingent experiment with co-registered electroencephalography and eye-tracking (N=72, 17 females, mean age=14 years-old, mean IQ=105). Participants viewed faces that responded to their gaze in four conditions: fixate on eyes, eyes open (mutual eye contact); fixate on eyes, mouth opens; fixate on mouth, eyes open; fixate on mouth, mouth opens. The amplitude and latency of the N170 were extracted from selected electrodes over the right and left occipitotemporal scalp. Friendship, social motivation, social cognition, and psychosocial functioning were measured with standardized interviews and questionnaires (see Table 1). Two ordinal regressions were conducted to examine the predictors of friendship quantity and interest, respectively, with N170 latency, social cognition and social motivation entered into the model as independent variables. A third multiple regression was run to assess whether friendship quantity and interest predicted psychosocial functioning. Final analyses will assess friendship quality as measured by coding from the Autism Diagnostic Observation Schedule-2.

Results: The first model significantly predicted friendship quantity over and above the intercept-only model [χ²(3)=16.91, p=.001]. Shorter N170 to mutual eye contact [Wald χ²(1)=12.34, p<.001] and higher social motivation [Wald χ²(1)=4.31, p=.04] significantly predicted greater number of friends. The second model significantly predicted friendship interest over and above the intercept-only model [χ²(3)=11.64, p=.009]. In this model, only higher social motivation emerged as a significant predictor of greater friendship interest [Wald χ²(1)=7.15, p=.04]. A third regression model examined friendship quantity and interest as predictors of psychosocial functioning. The overall regression model was significant [R²=.24, p=.03].

Conclusions: Associations between friendship and SIPS, social motivation, and psychosocial functioning in ASD are consistent with the model proposed by Mendelson et al. (2016). Results suggest neural differences in SIPS and social motivation in children with ASD are associated with enhanced ability to form and maintain friendships and thereby positively impact psychosocial functioning. Moreover, this study extends Mendelson et al. (2016)’s model, demonstrating that improved neural SIPS in response to mutual eye contact in ASD is associated with greater number of friendships and that social cognition is not associated with friendship. These clinically significant findings suggest that targeting SIPS and social motivation may lead to improved friendships in ASD and, eventually, improved psychosocial functioning.