Investigation of Physiological Features by Age Groups in Children with Autism

This paper presents the computerized analysis of physiological signals collected from children with autism from different age groups in 4 different countries during a child-robot interaction project. While the child interacts with the KASPAR humanoid robot, physiological signals (blood volume pulse (BVP), skin conductance (EDA), and temperature (ST)) were recorded using an Empatica E4 wristband to explore the emotions or stress of children in future studies. In this study, parametric statistical tests and feature selection methods have been used to investigate whether there are significant differences between age groups. The statistical tests revealed that there are specific subsets of features derived from BVP, EDA, and ST signals causing a significant difference between the signals of children from different age groups. Additionally, the feature selection study showed the set of most distinctive features for each age group.