Altered Functional Connectivity during Social Interaction in Children with Autism Spectrum Disorder

Background: Individuals with autism spectrum disorder (ASD) have difficulties engaging in reciprocal social interaction, which might be due to both atypical reward processing and mental state reasoning, or mentalizing (Chevallier et al., 2012; Pelphrey et al., 2011). While previous explorations of social processing in ASD were mainly focused on activation and conducted in non-interactive contexts, recent studies have begun to address this question during real-time social interaction (Jasmin et al., 2018; Redcay et al., 2013). However, research has not examined the integration within and between mentalizing and reward brain networks during social interaction in children with ASD, especially in middle childhood, an important period for social and cognitive development.

Objectives: We aimed to investigate functional correlations within and between the mentalizing and reward networks during social interaction in children with ASD compared to neurotypical (NT) children.

Methods: Preliminary data from an ongoing study are reported from 12 children with ASD (10 males, 12.3 ± 0.91 years) and 12 NT children (11 males, 13.01 ± 1.44 years), matched on age, gender, and IQ. We used an interactive game in which children predicted responses of either a social partner or a story character based on either mental or non-mental state information (Alkire et al., 2018).

Data preprocessing was performed using AFNI (Cox, 1996), with functional images normalized to a child template (Fonov et al., 2011) and effects of head motion minimized (i.e., censor volumes exceeding 1 mm framewise displacement (FD); remove runs >10% censored volumes or mean FD > 0.5 mm). Controlling for head motion, functional connectivity was calculated by estimating trial-wise beta series (Mumford et al., 2012; Rissman et al., 2004) of regions in the mentalizing and reward networks (Neurosynth; Yarkoni et al., 2011) for each condition (Figure 1). We examined effects of group (ASD vs. NT), social context (social partner vs. story character), and their interaction on the mean connectivity within and between the mentalizing and reward networks. Further, we verified specificity of the effects within and between the mentalizing and reward networks through a supplemental analysis within the mirror neuron and salience networks, which are also involved in social interaction (Redcay and Warnell, 2018).

Results: We observed group*social context interactions on mean connectivity within the mentalizing network (F(2,23) = 4, p = .0496) and between the mentalizing and reward networks (F(2,23) = 4.01, p = .0495). These effects were driven by greater connectivity during social interaction in NT children (mentalizing network: F(1,11) = 9.38, p = .004; between networks: F(1,11) = 6. 41, p = .017) but not in children with ASD. Our specificity analyses demonstrated no effects within the mirror neuron or salience networks.

Conclusions: We demonstrated that brain functional connectivity within and between the mentalizing and reward networks is modulated by social-interactive context in NT children but not in children with ASD. These preliminary results suggest reduced neural sensitivity to social interaction in children with ASD compared to NT children, which provides new insights into the brain basis underlying social difficulties in ASD during real-world social interaction.