Using Socially Assistive Robots in Speech-Language Therapy for Children with Language Impairments

• Published in: International journal of social robotics Vol. 15; no. 9-10; pp. 1525 - 1542
• Main Authors: Spitale, Micol, Silleresi, Silvia, Garzotto, Franca, Matarić, Maja J.
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.10.2023; Springer Nature B.V
• Subjects: Autism; Children; Control; Engineering; Head movement; Language; Linguistics; Mechatronics; Position measurement; Robotics; Service robots; Skills; Speech; Therapists; Therapy; Socially assistive robots; Language impairments; Speech-language therapy; Children
• ISSN: 1875-4791; 1875-4805
• DOI: 10.1007/s12369-023-01028-7

Socially assistive robots (SARs) have been shown to be promising therapy tools for children with primary or co-occurring language impairments (e.g., developmental language disorder and autism spectrum disorder), but only a few studies have explored the use of SARs in speech-language therapies. This work sought to address the following research goals: (1) explore the potential of using SAR for training linguistic skills of children with language impairments, targeting specific aspects of language and measuring their linguistic improvements in speech-language therapy; (2) explore children’s facial cues during SAR-supoported speech-language therapy; and (3) collect therapist perspectives on using SARs in speech-language therapy after having experienced it. Toward these goals, we conducted an 8-week between-subjects study involving 20 children with language impairments and 6 speech-language therapists who conducted the SAR-supported therapy. Children were randomly assigned to either a physical SAR or a virtual SAR condition; both provided the same language impairment therapy. We collected linguistic activity scores, video recordings, therapist questionnaires, and group interview data. The study results show that: (i) the study participants’ overall linguistic skills improved significantly in both conditions; (ii) participants who were engaged with the physical SAR (measured based on gaze direction and head position) were more likely to demonstrate linguistic skill improvements and had a significantly higher numbers of speech occurrences in the child-robot-therapist triads with the physical SAR; (iii) therapists reported skepticism about SAR efficacy in this context but believed that SAR could be beneficial for keeping children engaged, motivated, and positive during speech-language therapy.