Task-related electrodermal activity dynamics in robot-mediated interaction: a six-case study of children with neurodevelopmental disorders

Socially assistive robots have been explored as platforms to support structured interaction with children with neurodevelopmental disorders; however, most systems are designed for individuals with autism spectrum disorder (ASD) and may not be generalizable to other conditions. Emotional engagement and physiological arousal often differ across diagnostic categories and symptom severities, highlighting the challenge of capturing heterogeneous arousal patterns observed in small, diverse participant samples. This study presents a six case exploratory analysis of electrodermal activity (EDA) responses during four task-based emotional interaction sessions in robot-assisted training involving children diagnosed with ASD, pervasive developmental disorder, developmental coordination disorder, and intellectual disability. A multimodal platform integrating a NAO robot, an E4 wristband, facial expression tracking, and therapist control was used to synchronize behavioral and physiological data streams using unified timestamps; physiological signals were recorded for offline analysis. EDA signals were interpreted as task-related arousal dynamics rather than discrete emotional states. Descriptive time series inspection and vector autoregressive impulse response visualization were applied to characterize individual variability. The findings revealed heterogeneous arousal trajectories shaped by task type and individual characteristics, reflecting pronounced individual differences in task-related arousal dynamics. Overall, this study presents a timestamp synchronized multimodal data logging platform and an exploratory six-case analysis of task-related EDA arousal dynamics during robot-mediated interaction, emphasizing pronounced individual variability.