Objectively Measured Social Communicative Behaviors during the ADOS-2

Background: Autism spectrum disorder (ASD) is defined in part by persistent disturbances of social communication and interaction, such as eye contact (social gaze), facial expressions (smiles), and vocal turn-taking. In current best practice, ASD is recognized on the basis of expert clinician judgment, which is informed by gold-standard measures such as the Autism Diagnostic Observation Schedule-2 (ADOS-2; Lord et al., 2012). To date, observational research addressing social communication disturbances in children with ASD is rare in part because the field has lacked efficient methods for measuring behavior (Chin et al., 2018, Rehg et al., 2014, Swanson et al., 2017). Yet, improving social communication and reducing interaction disturbances depend on an objective understanding of how children with ASD behave during clinical assessments. The current study aims to improve objective understanding of social communicative behaviors among children with ASD during the ADOS-2.

Objectives: To investigate associations between objective measures of child’s social gaze and social smile measured from adults’ first-person video, child-adult vocal interaction measured from audio recording, and the traditional examiner-scored symptom measure ADOS-2 calibrated severity scores (CSS).

Methods: Fifteen children (M age=35.93mo, SD=6.61) were administered the ADOS-2 assessments. The ADOS-2 yields a total CSS as well as subscales for Social Affect (SA CSS) and restricted/repetitive behavior (RRB CSS). Children’s total duration of smiles, socially-directed smiles, and social gaze during the assessment were recorded with a small Pivothead camera worn in glasses by the examiner and the parent. iMotions software was used to detect child smile expressions and gaze during the assessment as recorded on the parent and examiner Pivothead camera. Audio from wall-mounted camera recordings was processed with a Language ENvironment Analysis (LENA) software, which provided automated vocalization detection yielding child and adult vocal initiation counts and turn counts within conversational vocal blocks. Measures included proportion of gaze and smile at examiner and parent, total duration of gaze and smile at adult, rate of child and adult vocal initiation, and ADOS-2 CSS (total, SA, and RRB).

Results: The mean duration of smiles recorded by both examiner and parent head-mounted cameras was negatively associated with the total CSS (r=-.53, p<.05) and RRB CSS (r=-.52, p<.05), but not significantly associated with SA CSS (r=-.39, p>.05). The association between the proportion of gaze at the examiner and the total CSS was marginally significant (r=-.493, p=.06). Higher RRB CSS was significantly correlated with lower proportion of gaze (r=-.62, p<.05) and social smile (r=-.62, p<.01) at the examiner. In boys (n=11), more adult vocal initiation was associated with fewer smiles at examiner (r=-.67, p<.05), but this association was not significant for the group as a whole.

Conclusions: Objective measurements of children’s smiles and gaze during the ADOS-2 converge with examiner-scored ADOS-2 calibrated severity scores. Although these dimensions of social communication are indexed by ADOS-2 SA CSS, objectively measured social behaviors were unexpectedly associated with RRB CSS, which may limit social interaction. The objective measurement of key behavioral features of ASD appears to have the potential to produce quantitative indices of ASD symptoms.