Multiple-region directed functional connectivity based on phase delays.

Area of Science:

• Neuroimaging
• Network Science
• Computational Neuroscience

Background:

• Traditional neural network analysis often assumes linear relationships between brain regions.
• Existing methods may not fully capture the complexity of multi-region interactions.

Purpose of the Study:

• To present a high-order statistical framework for calculating directed functional connectivity.
• To characterize brain networks as linear, nonlinear, or disconnected using wavelet analysis and spectral coherence.
• To determine network temporal hierarchy and directionality via phase delays.

Main Methods:

• Developed a high-order statistical framework using wavelet analysis and spectral coherence.
• Derived mathematical expressions for network characterization in four regions.
• Validated the framework using computer simulations of the Kuramoto model.
• Applied the analysis to resting-state fMRI data from 40 healthy subjects.

Main Results:

• The framework accurately determined network directionality with low sensitivity to noise.
• Ventral visual system networks were predominantly linear.
• Motor system and default mode network (DMN) exhibited nonlinear network compositions.
• Identified distinct temporal hierarchies for the ventral visual, motor, and DMN systems.

Conclusions:

• The high-order statistical framework provides a robust method for analyzing complex brain networks.
• This approach reveals nonlinear interactions within key brain systems like the motor system and DMN.
• The method has broad applicability across various scientific disciplines and neuroimaging data types.