The balance between integration and segregation drives network dynamics maximizing multistability and metastability

Area of Science:

• Computational Neuroscience
• Network Science
• Systems Neuroscience

Background:

• The brain's structural connectivity shapes its dynamic functional states, a key aspect of cognitive flexibility and stability.
• Understanding the whole-brain structure-dynamics relationship is crucial but remains challenging, with prior work focusing on local mechanisms.

Purpose of the Study:

• To investigate how network integration-segregation balance influences brain dynamics, specifically multistability and metastability.
• To identify key structural network properties that predict neural dynamical complexity.

Main Methods:

• Analyzed a spectrum of network models using structural metrics (modularity, efficiency, small-worldness).
• Simulated neural activity using a neural mass model and analyzed Functional Connectivity Dynamics (FCD).
• Quantified structure-dynamics relationships using Mutual Information (MI).

Main Results:

• Segregated networks sustain dynamic synchronization; small-world networks exhibit the richest dynamics.
• Peak dynamical richness was observed in networks with intermediate small-worldness, characterized by high FCD variance and metastability.
• Network modularity was the strongest predictor of dynamics, enabling transitions between states.

Conclusions:

• A balance between local specialization, global integration, and modularity is critical for fostering dynamic complexity in brain networks.
• These findings enhance the understanding of how structural features govern neural dynamics, supporting cognitive functions.