Individual Deviation-Based Functional Hypergraph for Identifying Subtypes of Autism Spectrum Disorder.

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Autism spectrum disorder (ASD) is highly heterogeneous, complicating diagnosis and treatment.
• Previous subtype identification methods using pairwise similarities may underestimate relationship complexity.
• Incorporating typical development (TD) data can enhance diagnostic insights.

Purpose of the Study:

• To develop a novel method for identifying ASD subtypes by analyzing inter-individual relationships.
• To leverage functional MRI data and a hypergraph framework to capture complex relationships.
• To improve the understanding and diagnosis of ASD heterogeneity.

Main Methods:

• Constructed an individual deviation-based hypergraph (ID-Hypergraph) using an elastic net model on functional MRI data.
• Applied a novel community detection clustering algorithm to the ID-Hypergraph for subtype identification.
• Validated findings on discovery and replication datasets from the Autism Brain Imaging Data Exchange (ABIDE).

Main Results:

• Identified four reproducible ASD subtypes with consistent ALFF patterns across datasets.
• Observed significant variations in communication domains among the identified subtypes.
• Achieved over 80% accuracy in classifying individuals into these ASD subtypes.

Conclusions:

• The ID-hypergraph approach effectively identifies reproducible ASD subtypes.
• This method enhances the elucidation of ASD heterogeneity and aids in subtype diagnosis.
• The findings suggest potential for improved, personalized ASD diagnostic strategies.