Identifying Relationships in Functional and Structural Connectome Data Using a Hypergraph Learning Method.

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Systems Neuroscience

Background:

• The brain connectome offers insights into neuronal network organization at regional and whole-brain levels.
• Advancements in analyzing structural and functional connectomes exist, but integrating them for complex relationship discovery remains a challenge.
• Combining structural and functional connectome data is crucial for understanding health and disease, including neurological disorders.

Purpose of the Study:

• To propose a novel connectome feature selection technique integrating structural and functional data.
• To identify complex relationships between fiber density and signal synchronization.
• To enhance diagnostic accuracy for neurological conditions like autism.

Main Methods:

• Utilized hypergraphs for connectome feature selection, combining structural and functional connectome data.
• Employed Support Vector Machine (SVM) classifiers for subject identification.
• Used publicly available connectome data from the UMCD database.

Main Results:

• The proposed method achieved 88% accuracy in identifying autism subjects.
• SVM classifiers trained with selected features demonstrated high diagnostic performance.
• The integrated approach effectively captured connectivity relationships.

Conclusions:

• The novel hypergraph-based feature selection technique successfully integrates structural and functional connectome data.
• This combined approach shows promise for improving outcome forecasting in autism.
• The method may have broader applications in understanding brain health and disease.