Peer-reviewed article
RQA-based identification of emotions from electrocargiogram signals for emotion regulation in children with autism spectrum disorder.
S Jerritta, N Sindhu ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This is a technical paper about using electrocardiogram signals and machine learning algorithms to identify emotions in children with autism spectrum disorder. The study developed a method to recognize emotional states from heart rate data in 25 children with ASD and 25 typically developing children. While the classification accuracy was high, the paper focuses on the technical algorithm development rather than testing an intervention to improve child well-being.
Key points
- ECG-based emotion recognition achieved 97.9% accuracy for positive emotions and 87.9% for negative emotions in children with ASD
- Recurrent Quantification Analysis (RQA) algorithms were used to extract features from heart rate data
- Study compared children with ASD using personalized emotion elicitation protocols to typically developing children using generic protocols
- The research suggests RQA-based methods warrant further exploration for emotional state identification
Limitations
This is a technical study on algorithm development and does not evaluate whether identifying emotions through ECG signals actually improves emotion regulation, behaviour, or well-being in children with ASD. The sample size is small (25 per group) and there is no information about whether the tool was implemented or tested in practice.
Abstract
Children with Autism Spectrum Disorder (ASD) have difficulties in expressing and regulating their emotions resulting in meltdowns and outbursts that make it difficult for parents, medical practitioners and caretakers. This research aims to recognize the unexpressed positive and negative emotional states of children with ASD using electrocardiogram (ECG) signals. Emotional ECG data is obtained from 25 children with ASD using a personalized emotion elicitation protocol, catered to the emotional need of the child. Emotional data was also obtained from 25 typically developed children using a generic protocol. The ECG data was pre-processed and features were extracted using Recurrent Quantification Analysis (RQA) algorithms. The influence of the various features on emotion recognition is analysed. Classification results indicate 97.9% and 87.9% respectively in identifying the positive and negative emotional states in children with ASD. The better correlation of RQA based ECG features with emotions for children with ASD paves way for RQA and similar nonlinear methods to be explored further for better identification of emotional states.