Using a learning-by-teaching approach, the study found that most participating children showed both strong engagement and measurable handwriting improvement, closely tied to the robot’s socially responsive behavior. The findings build on growing interest in the use of social robots to support personalized learning in special education contexts.
Background
Children with neurodevelopmental disorders (NDDs) often face academic challenges, including dysgraphia, a handwriting impairment that can affect both learning and self-confidence. While socially assistive robots have shown promise in supporting children with such needs, most prior studies have been short-term and conducted in tightly controlled lab environments, limiting their relevance in everyday educational settings.
This study aimed to fill that gap by testing the R2C3 system in a real-world daycare center. The research team set out to understand whether a socially interactive, realistic intervention could meaningfully improve handwriting skills and self-esteem in children with complex NDDs, especially when the robot’s behavior was tailored to each child by a caregiver.
Study Design and Methodology
Eighteen children participated in nine 20-minute sessions involving the R2C3 system, which created a triadic learning environment between a social robot (QTRobot), the child, and a caregiver.
At the core of the sessions was a learning-by-teaching setup: children were asked to teach the robot how to perform handwriting tasks using the Dynamilis serious game app. The robot's behavior varied across two conditions—either personalized and controlled by the caregiver (using a Wizard of Oz interface), or operating autonomously with minimal, stereotyped social cues.
The researchers focused on three core questions: Did the intervention improve handwriting skills (measured by the BHK assessment) and self-esteem (measured by the SPPC questionnaire)? What types of interaction patterns were linked to improvement? And how did the level of robot personalization influence engagement?
To answer these questions, the team collected detailed multimodal data using synchronized cameras and audio recorders. They tracked gaze direction, speech activity, and robot behavior frequency. Statistical analyses included Wilcoxon rank tests to assess score changes and linear mixed-effects models to examine correlations between observed behaviors and clinical outcomes.
Results and Findings
Out of 15 children with complete data, 60 % improved their handwriting scores by at least one standard deviation in either speed or quality. However, there were no statistically significant differences between the group’s pre- and post-test scores, pointing to high individual variability. Encouragingly, the children with the most severe baseline difficulties, particularly in writing speed and self-esteem, showed the greatest improvement.
When researchers looked more closely at interaction patterns, they found that lower speech activity during sessions was significantly associated with better handwriting outcomes. There were also trends suggesting that children who spent more time in silence and less time looking at the tablet achieved greater gains, indicating a deeper cognitive focus during task engagement.
A comparison of the two robot conditions added another layer of insight. Although the autonomous robot exhibited more frequent behaviors overall, the personalized, caregiver-controlled robot drove better outcomes. Sessions in this condition resulted in richer verbal interaction and greater engagement with the handwriting tasks, as children spent more time actively participating in the serious game.
Discussion and Implications
Taken together, the results suggest that socially assistive robots can play a meaningful role in motivating children with complex NDDs to engage in therapeutic activities. The high retention rate across sessions reinforces the system’s appeal. While the study didn’t find group-wide statistical gains in handwriting or self-esteem, many children made significant personal progress.
Importantly, the findings highlight two critical factors in driving improvement: social engagement through conversation, and task engagement through sustained focus. The personalized robot guided by a caregiver proved to be far more effective than one using repetitive, pre-scripted behaviors. This supports the idea that human-guided personalization is essential to unlocking the full potential of social robots in education.
The authors noted several limitations, including a small sample size, single-site implementation, and the complexity of conducting multimodal interaction analysis in naturalistic settings. Still, the study offers valuable evidence that socially interactive, customizable robots can complement traditional educational and therapeutic interventions for children with dysgraphia.
Conclusion
This research demonstrates that a socially assistive robot using a learning-by-teaching approach can effectively engage children with NDDs in handwriting re-education. Conducted in a real-world setting, the study underscores the importance of caregiver-guided personalization in fostering meaningful interaction and sustained engagement. As social robots become more sophisticated and accessible, they may serve as valuable tools to support individualized learning and therapy for children with diverse educational needs.
Journal Reference
Zou, J., Gauthier, S., Archambault, D., Chetouani, M., Cohen, D., & Anzalone, S. M. (2025). Robot-Assisted Handwriting Training: An Intervention for Children with Neurodevelopmental Disorders. Computers in Human Behavior Reports, 100799. DOI:10.1016/j.chbr.2025.100799. https://www.sciencedirect.com/science/article/pii/S2451958825002143
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.