Structure and dynamics analysis of brain functional hypernetworks based on the null models.

Area of Science:

• Neuroscience
• Network Science
• Computational Biology

Background:

• Brain functional hypernetworks model complex interactions for disease diagnosis.
• Limited research exists on the structure and dynamics of these hypernetworks.
• Understanding hypernetwork features is key to elucidating brain function and pathology.

Purpose of the Study:

• To introduce a hypernetwork null model for analyzing feature dependencies in brain functional hypernetworks.
• To investigate the structural and dynamic properties of brain functional hypernetworks.
• To explore how different attributes contribute to the overall network topology and function.

Main Methods:

• Developed an optimized hyper dK-series algorithm to construct null models preserving node and hyperedge attributes.
• Introduced multiple node and hyperedge attributes to both original and null hypernetwork models.
• Calculated similarity and correlation between topological attributes of original and null models to assess feature dependencies.

Main Results:

• Identified varying levels of dependence between different features of interest in brain functional hypernetworks.
• Node degree emerged as the primary dependency attribute across multiple metrics.
• Hyperedge degree and redundancy coefficients showed partial dependency, suggesting unique topological information.

Conclusions:

• Node degree contains redundant information relative to other attributes.
• Hyperedge degree and redundancy coefficients may capture additional topological nuances.
• Redundancy exists between hypernetwork clustering coefficients (HCC²) and (HCC³), necessitating careful consideration in network analysis.