A systematic framework for functional connectivity measures.

Area of Science:

• Neuroscience
• Computational Biology
• Network Science

Background:

• Characterizing functional connectivity is crucial for understanding brain networks across modalities like electrophysiology and fMRI.
• Existing functional connectivity measures produce varied results, necessitating an evaluation of their applicability.
• A systematic approach is needed to assess the performance and limitations of different connectivity analysis techniques.

Purpose of the Study:

• To establish a comprehensive framework for evaluating the performance of diverse functional connectivity measures.
• To benchmark a wide array of methods using simulated datasets from various generative models.
• To provide guidance on selecting appropriate functional connectivity measures based on data characteristics.

Main Methods:

• Benchmarking 42 functional connectivity methods using 10,000 simulated datasets from 5 generative models.
• Optimizing method parameters (e.g., window size, model order) using ROC analysis.
• Assessing method performance across different signal-to-noise ratios and network configurations.

Main Results:

• The study systematically evaluated the performance of 42 functional connectivity methods.
• Parameter optimization was performed using threshold-free ROC analysis.
• Method performance varied significantly based on generative model type, signal-to-noise ratio, and network complexity.

Conclusions:

• A robust framework for evaluating functional connectivity measures has been developed.
• The findings highlight the context-dependent performance of different connectivity methods.
• A provided MATLAB toolbox facilitates further analysis and method comparison.