Efficient Computation of Functional Brain Networks: toward Real-Time Functional Connectivity.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Network Science

Background:

• Functional connectivity is crucial for understanding brain information processing, balancing segregation and integration.
• Existing computational methods for connectivity indices and graph-theory measures can be slow, limiting analysis scope.
• There is a need for efficient tools to analyze complex brain networks.

Purpose of the Study:

• To present a suite of open-source tools enhancing the computational efficiency of brain connectivity measures.
• To significantly speed up the calculation of phase synchronization, information theory, and graph-theory metrics.
• To enable large-scale, real-time network analysis of brain function.

Main Methods:

• Implementation of optimized algorithms for Phase Locking Value (PLV), Phase Lag Index (PLI), Imaginary part of the Coherency (ImC), and weighted Phase Lag Index (wPLI).
• Development of efficient computation for Mutual Information and Generalized Synchronization indices.
• Integration of fast implementations for graph-theory measures including Strength, Shortest Path Length, Clustering Coefficient, and Betweenness Centrality.

Main Results:

• Achieved significant computational speed-ups (up to thousands of times faster) for various connectivity and network measures compared to existing implementations.
• Demonstrated substantial increases in computational efficiency for Phase Synchronization measures (PLV, PLI, ImC, wPLI), Mutual Information, and Generalized Synchronization.
• Successfully implemented highly efficient graph-theory network measures.

Conclusions:

• The developed open-source tools dramatically improve the speed of brain connectivity analysis.
• These advancements make complex network analysis of brain function feasible within practical time constraints.
• The tools facilitate whole-head, real-time network analysis, advancing neuroscience research.