Assessment of the communicative and coordination skills of children with Autism Spectrum Disorders and typically developing children using social signal processing

• Published in: Research in autism spectrum disorders Vol. 7; no. 6; pp. 741 - 756
• Main Authors: Delaherche, Emilie, Chetouani, Mohamed, Bigouret, Fabienne, Xavier, Jean, Plaza, Monique, Cohen, David
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.06.2013; Elsevier
• Subjects: Artificial Intelligence; Autism; Autism spectrum disorder; Children; Cognitive science; Communication Skills; Computer Science; Computer Vision and Pattern Recognition; Coordination; Counselors; Developmental Stages; Imitation; Interpersonal Communication; Neuroscience; Nonverbal Communication; Pervasive Developmental Disorders; Psychology; Robotics; Signal and Image Processing; Social Development; Social Influences; Social signal processing; Speech Communication; Coordination; Social signal processing; Autism spectrum disorder; Imitation
• ISSN: 1750-9467; 1878-0237
• DOI: 10.1016/j.rasd.2013.02.003

► Low-level automatic features discriminate typical and ASD children. ► The rhythm of the therapist interacting with the child predict the child clinical group. ► The duration of the therapist gestural pauses predict the child clinical group. ► The performance in the tasks also depends on the developmental age of the child whatever the group. To cooperate with a partner, it is essential to communicate by sharing information through all available avenues, including hand gestures, gazes, head gestures and naturally, speech. In this paper, we compare the communicative and coordination skills of children with typical development to those of children with Autism Spectrum Disorders (ASDs) in cooperative joint action tasks. Communicative skills were assessed using a pragmatic annotation grid. Coordination skills were assessed based on automatically extracted features that characterize interactive behavior (turn-taking, synchronized gestures). First, we tested the performance of the interactive features when discriminating between the two groups of children (typical vs. ASD). Features characterizing the gestural rhythms of the therapist and the duration of his gestural pauses were particularly accurate at discriminating between the two groups. Second, we tested the ability of these features for the continuous classification problem of predicting the developmental age of the child. The duration of the verbal interventions of the therapist were predictive of the age of the child in all tasks. Furthermore, more features were predictive of the age of the child when the child had to lead the task. We conclude that social signal processing is a promising tool for the study of communication and interaction in children with ASD because we showed that therapists adapt differentially in three different tasks according to age and clinical status.