Related Experiment Video
Updated: Oct 13, 2025
05:04
An Electrophysiology Protocol to Measure Reward Anticipation and Processing in Children
Published on: October 4, 2018
7.1K
Region-Wise Brain Response Classification of ASD Children Using EEG and BiLSTM RNN.
Thanga Aarthy Manoharan1, Menaka Radhakrishnan2
1Vellore Institute of Technology, Chennai, TN, India.
Clinical EEG and Neuroscience
|November 18, 2021
Summary
Recurrence quantification analysis (RQA) of electroencephalography (EEG) signals effectively distinguished autism spectrum disorder (ASD) from typically developing (TD) individuals. Specific EEG channels and analysis parameters achieved high classification accuracy, offering a potential biomarker for ASD.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Developmental Psychology
Background:
- Autism spectrum disorder (ASD) is a neurodevelopmental condition marked by sensory modulation impairments, affecting adaptive behavior and intellectual function.
- Understanding neural dynamics in ASD is crucial for developing effective diagnostic and therapeutic strategies.
- Electroencephalography (EEG) provides a window into brain activity, but advanced analysis is needed to capture complex neural patterns in ASD.
Purpose of the Study:
- To investigate neural dynamics differences in children with ASD compared to typically developing (TD) children using EEG.
- To explore the efficacy of non-linear EEG signal analysis, specifically Recurrence Quantification Analysis (RQA), for identifying biomarkers in ASD.
- To develop and validate a classification model for distinguishing ASD from TD individuals based on EEG-derived features.
Main Methods:
- Employed electroencephalography (EEG) to record brain activity in 20 children with ASD and 20 TD children.
- Utilized Recurrence Quantification Analysis (RQA) with a cosine distance metric and various parameter settings to analyze nonlinear EEG dynamics.
- Developed a BiLSTM (bi-long short-term memory) network to classify ASD and TD groups using RQA-derived features, testing channel-wise accuracy.
Main Results:
- The study identified specific RQA parameters and EEG channels that effectively differentiate between ASD and TD groups.
- The combination of T3 and T5 EEG channels, using a fixed amount of nearest neighbors (FAN) and the cosine distance metric, yielded the highest classification accuracy.
- A classification accuracy of 91.86% was achieved, demonstrating the potential of this RQA-based approach for ASD identification.
Conclusions:
- Non-linear analysis of EEG signals using RQA is a promising method for understanding neural differences in ASD.
- The identified optimal RQA parameters and channels (T3, T5) show potential as reliable biomarkers for discriminating ASD from TD individuals.
- This approach offers a data-driven method for potentially aiding in the early detection and characterization of autism spectrum disorder.
