Robot-Mediated Interventions for Social Anxiety in Individuals with Autism Spectrum Disorder
Objectives: We propose the initial application of robot-mediated telecommunication using an android robot for targeted interaction to improve communication skills in individuals with social anxiety and ASD.
Methods: The inclusion criteria were chronological age of 18 to 24 years and previous diagnosis of high-functioning ASD (IQ ≥ 70) and social anxiety. This experiment was conducted in a familiar room that was used often by the participants for various activities. The participants typed on the keyboard, and the android spoke. The android was positioned at one end of the room within plain view of the participants. There was a camera image before the participant, which the participant was able to watch. The android robot used in this study was ACTROID-F (Kokoro Co. Ltd.), a female-type humanoid robot with an appearance similar to that of a real person. Its artificial body has the same proportions, facial features, hair color, and hairstyle as a human. At first sight and from a distance, it is difficult to distinguish this android robot from a live adult. ACTROID-F is capable of a range of movements (moving limbs up and down and turning the head from side to side). We adopted the remote control system conventionally used in robotics studies in covert ways. ACTROID-F is tele-operated to perform semi-structured conversations with a person for consultation. Emotion was measured using the Stress Response Scale (SRS).
Results: Twenty-three individuals diagnosed with ASD and social anxiety participated. Based on video recordings of the interactions, quantitative and qualitative analyses were conducted. All participants were able to respond and type in any situation. In all cases, when they typed on the keyboard, their expressions became soft. Before they typed, “enter,” they were stiff, but after they typed, “enter,” they were relieved. They seemed to be deep in conversation. Total score and some item scores on the SRS (i.e., become irritable, mind everything, sink deeply, and cannot concentrate on something) improved from before to after intervention (p < 0.01).
Conclusions: Despite the limited effectiveness of the intervention, it is suggested that intervention using an android could be used as a tool for the development of communication. Future research needs to examine the issue using a larger, more diverse sample and multiple levels of analysis of behavioral change.