Supporting E-Learning with Emotion Regulation for Students with Autism Spectrum Disorder

Abstract

Students with Autism Spectrum Disorder (ASD) in general have been found to have significantly lower academic achievement relative to their level of ability. Research has shown that students’ emotional impairment with ASD severely interferes with their learning process, and academic emotions are domain-specific in nature. Therefore, the regulation of domain-specific academic emotions is an important approach to help students with ASD learn effectively. This study proposed an e-learning model that featured emotion recognition and emotion regulation to enhance mathematics e-learning for students with ASD. An emotion recognition approach based on facial recognition, an emotion regulation model, and a mathematics e-learning platform, were developed to realize the e-learning model. Two e-learning conditions: timed contest and increased difficulty of learning, were created for gathering information by observing two indexes: mathematical learning performance and negative emotional behaviors in each condition. An experiment in a mathematical e-learning context was performed to evaluate the performance of e-learning and emotion regulation effectiveness. The results of the emotion recognition classifier reached a 93.34% average recognition rate, and the participants of this experiment displayed a statistically significant decrease in targeted negative behaviors from baseline to intervention (p = .000) and significant improvements in mathematics learning performance (p = .005); however, responses to emotion regulation interventions varied among the participants. Implications for research and practice are discussed.

Keywords

• autism spectrum disorder
• emotion regulation
• emotion recognition
• e-learning
• intelligent systems
• mathematics learning