Objective Measurement of Social Communication Behaviors in 3-Year-Olds With Suspected ASD During the Autism Diagnostic Observation Schedule-2

• In Event: Using computer technology to expand detection and intervention for children with developmental disorders

Abstract

Background: Autism spectrum disorder (ASD) is defined by persistent disturbances of social communication, as well as repetitive patterns of behavior. In current best practice, ASD is identified on the basis of expert, but subjective, clinician judgment, which is informed by gold standard measures such as the Autism Diagnostic Observation Schedule-2 (ADOS-2), in conjunction with parent report of developmental history and current functioning. Quantification of key social behavioral features of ASD using objective measurements would enrich scientific understanding of the disorder. The current study sought to improve objective understanding of children’s social communicative interaction with adults (both examiner and parent) present during the administration of the ADOS-2, by leveraging computer vision and audio signal processing that can identify a key set of objective measures of children’s social communication behaviors captured with adult-worn camera-embedded eyeglasses during the assessment.
Objective: To investigate associations between objective measures of child’s social communication behaviors (i.e. social gaze, social smile, and adult-child vocal turn-taking) measured from adults’ first-person video and the examiner-adjudicated symptom measure ADOS-2 total scores.
Methods: Sixty-six children (Mage=39.97 months, SD=10.58) with suspected ASD were administered the ADOS-2. A hypothesis-blind examiner provided total scores in two domains, Social Affect (SA) and Restricted and Repetitive Behavior (RRB), as well as the Overall (SA and RRB combined). Children’s total duration of social gaze and socially-directed smiles, as well as vocal interaction with adults, were recorded with a small camera worn in eye-glasses (Pivothead or Orca) by the examiner and parent. We processed first-person video from the cameras with Affectiva (iMotions) software to automatically quantify child gaze and smile expressions. Audio from the camera was processed with Language ENvironment Analysis (LENA) software, which provided automated vocalization detection yielding adult-child vocal turn-taking counts. Measures included ADOS-2 total scores (SA, RRB, and overall) and proportions of gaze, social smile at adult, and adult-child vocal turn-taking.
Results: We found negative associations between the ADOS-2 SA score and the proportions of social smile (r=-.34, p=.01). The RRB score was negatively correlated with the proportions of social gaze directed at adults (r=-.33, p=.01), social smile at adults (r=-.26, p=.04), and adult-child vocal turn-taking (r=-.29, p=.02). The overall (SA+RRB) score showed a negative association with the proportions of social smile directed at adults (r=-.37, p=.002) and adult-child vocal turn-taking (r=-.25, p=.04).
Conclusion: The current study provides an initial investigation of the objective detection of social communication behaviors during assessment of ASD symptoms. Objective measurements of children’s social communicative behaviors, during the ADOS-2 showed relatively low levels of association with the examiner-adjudicated ADOS-2 scores. The objective measurement of key behavioral features of ASD using first-person video and automated vocal interaction detection appears to have the potential to produce quantitative indices of ASD symptoms, which would provide useful information to clinicians and parents about children’s ASD symptoms and may allow for more efficient and objective monitoring of children’s development.

Authors

• Yeojin Amy Ahn, University of Miami
  Presenting Author

• Jacquelyn Michele Moffitt, University of Miami
  Non-Presenting Author

• Stephanie Custode, University of Miami
  Non-Presenting Author

• Amy Beaumont, University of Miami
  Non-Presenting Author

• Sandra Cardona, University of Miami
  Non-Presenting Author

• Meaghan Venezia Parlade, University of Miami
  Non-Presenting Author

• Jennifer Durocher, University of Miami
  Non-Presenting Author

• Melissa Hale, University of Miami
  Non-Presenting Author

• Michael Alessandri, University of Miami
  Non-Presenting Author

• Lynn K Perry, University of Miami
  Non-Presenting Author

• Daniel Messinger, University of Miami
  Non-Presenting Author