Hypergraph representation of multilayer brain network enhances autism spectrum disorder detection

Area of Science:

• Neuroscience
• Biomedical Engineering
• Data Science

Background:

• Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition.
• Understanding functional brain network alterations is crucial for ASD diagnosis.
• Conventional network analysis methods may miss higher-order connectivity patterns.

Purpose of the Study:

• To introduce a novel hypergraph-based framework for analyzing functional brain networks in children with ASD.
• To identify unique connectivity signatures associated with ASD.
• To enhance diagnostic accuracy for ASD using advanced network analysis.

Main Methods:

• Utilized resting-state electroencephalography (EEG) data from children with ASD.
• Employed a two-stage analysis: constructing multilayer networks via recurrence quantification analysis, then transforming them into hypergraphs.
• Applied support vector machines (SVM) for classification using derived network features.

Main Results:

• Identified distinctive functional brain connectivity signatures in ASD, particularly in bilateral frontal regions.
• Hypergraph representations revealed complex neural relationships and patterns previously obscured in traditional analyses.
• Hypergraph-derived features achieved 81% classification accuracy (F1-score), significantly outperforming multilayer network features (57%).

Conclusions:

• Hypergraphs offer a more stable and informative approach to identifying biomarkers for ASD.
• The proposed framework provides a powerful analytical tool for studying neurodevelopmental disorders.
• This hypergraph-based method shows significant clinical potential for developing more objective diagnostic tools for ASD.