Enhancing Empirically-Supported Treatments for Autism Spectrum Disorders: A Case Study Using An Interactive Robot

Background: The use of robots in clinical therapy for children with autism spectrum disorders (ASD) has begun to receive research attention, although most of the research has focused on robot development, and very little attention has been given to integrating robots into empirically-supported treatments (ESTs).

Objectives: The objective of this case study is to identify the feasibility of incorporating an interactive robot in an ABA-based social-communication therapy.

Methods: The participant in the case study was a nine-year-old male with ASD. Diagnosis was confirmed using the Autism Diagnostic Observation Schedule and the Social Communication Questionnaire. The participant’s IQ and language abilities were both three standard deviations below the mean. At the beginning of the study, the child’s language was predominantly echolalic, but did contain some spontaneous communication. The participant attended one baseline visit, followed by 50-minute therapy sessions two times a week for 8 weeks with a certified behavior analyst and a robot, and one posttest visit. The interactive robot was a humanoid robot capable of online text-to-speech communication and movement that allowed for human-like social gestures. We developed scripts relevant to the child’s ABA goals of initiating/responding to greetings, answering questions about feelings, and answering/asking recall questions. During the session, the participant would alternate between interacting with the therapist and interacting with the robot. The therapy session was monitored by an experimenter in another room who also controlled the robot. Verbal communication performance was coded from videos by identifying correct and incorrect responses to scripted conversation. Pre/post parent report of verbal communication outside of the therapy sessions was also collected. 

Results: Results indicated an overall increase in the amount of correct verbal behavior inside and outside of therapy. At baseline, the participant used the appropriate communication behavior 29% of the time with the therapist. In the first session, the participant exhibited correct behaviors 25% of the time to the therapist and 22% of the time to the robot. At the midpoint of the study, the child exhibited an increase in correct behaviors with both the therapist (30%) and with the robot (30%). In the last session, the child was correctly responding to the therapist 59% of the time and only 9% to the robot. Additionally, parent pre/posttest data indicated a 62% increase in appropriate verbal behavior during interactions, suggesting that at least some of the skills developed in therapy may have been generalized outside of the sessions.

Conclusions: Our participant showed individual verbal communication gains both inside and outside of the sessions. The benefits of the robot for this participant’s ABA goals were short-term, and by the end of the therapy most of his correct behaviors were directed at the therapist. Our method and data suggest it is feasible to incorporate an interactive robot into ESTs such as ABA, and further suggest that the robot may allow for some scaffolding of verbal skills. Additional research is needed to validate our procedure and confirm its effects.