Novel in-Vivo Approach Using Smartphone Reporting to Map Real-World Social Interactions in ASD

Background: While considerable research demonstrates deficits in the quality of social interactions of adults with ASD, little work has examined the quantity and temporal patterning of such interactions. Prior work suggests that adults with ASD engage in few social interactions (Orsmond et al., 2004), however, these studies have almost exclusively utilized caregiver reports, which are subject to recall biases (Shiffman et al, 2008). Ecological momentary assessment (EMA) is designed to address these limitations by facilitating in-vivo data collection, providing a more accurate estimate of social behavior. This is particularly important for individuals with ASD who may have limited insight upon reflection (Damiano et al., 2014). Only one study has utilized an EMA approach to examine social interactions of adults with ASD, finding lower levels of social interaction relative to other activities (Chen et al, 2017). However, this study used beep-contingent reporting, restricting inferences about quantity and patterning of interactions. Finally, prior work has indicated that symptoms of alexithymia (an impairment recognizing and understanding emotion), not ASD, may account for some of the social deficits in ASD (Bird & Cook, 2013). The current study represents a novel EMA approach to quantifying in-vivo patterns of social interaction in adults with ASD and the first examination of temporal patterns and psychosocial correlates of real-world social interaction.

Objectives: Based on previous research, we hypothesized that, controlling for demographic variables, adults with ASD would exhibit fewer social interactions and a daily cycle of social interaction with a lower peak than TD adults. We also hypothesized that symptoms of alexithymia, above and beyond the effects of ASD, would be related to fewer social interactions, and an attenuated peak in daily cycle of social interactions.

Methods: Adults with and without ASD (N_ASD = 23, N_TD = 52), aged 18 to 47 years (M=22.15 years, SD=5.53; 33 males) reported their social interactions in-vivo via smartphone over one week. Participants completed measures of IQ, ASD symptom severity and alexithymia symptom severity. Cyclical multilevel models were used to account for dependency of observations within person.

Results: Results demonstrated a daily cyclical pattern of social interaction that was robust to symptoms of ASD and alexithymia (Figure 1). Adults with ASD did not report fewer social interactions than TD peers; however, alexithymia symptom severity was negatively related to social interactions regardless of ASD status (Table 1).

Conclusions: This is the first study to use an EMA approach to map quantities and patterns of real-life social interactions. Results suggest a temporal pattern of social interaction, comparable across groups, that could not otherwise be identified. Our findings are the first to link alexithymia to reduced social interactions and suggest that these symptoms, not ASD severity, may drive social isolation for adults with ASD. This implicates a unique subset of individuals with ASD who struggle with social interactions and suggests an avenue for improving personalized intervention. In conclusion, results highlight the utility of a smartphone-based EMA approach for capturing subtle details of daily life in ASD with potential to provide a novel real-world outcome measure.