The effects of human-computer interaction-based interventions for autism spectrum disorder: a meta-analysis

Interventions utilizing human-computer interaction (HCI) in autism rehabilitation have gained popularity, yet their efficacy has not been fully elucidated. This meta-analysis assessed 30 international studies from the past decade that employed HCIs for autism intervention. The findings revealed a moderate positive effect (g = 0.484) on improving social skills, reducing repetitive behaviors, and mitigating emotional disturbances in individuals with autism. Common HCI technologies, such as applications, robots, and virtual reality, exhibit moderate effects. Adult participants benefitted the most, followed by adolescents, children, and infants. This discrepancy may be related to the stability and engagement of adults in terms of cognition and behavior. Individual-format interventions surpass group formats, and interventions lasting less than one month tend to be more focused and goal-oriented, demonstrating the most significant effects on ASD. Furthermore, research findings indicate no publication bias; however, due to the variability among individuals with ASD and the diversity of intervention measures, further analysis is needed to identify the underlying factors influencing the effectiveness of these interventions.