Comparison of Different Methods to Define Dynamic Brain Connectivity Analysis.

This study compares methods for analyzing the dynamic functional connectivity of the brain using the imaginary component of the complex Pearson correlation coefficient as the index of functional connectivity. The most commonly used method of analysis using a constant sliding window with predefined narrow and wide window widths was compared to methods that use adaptive window widths for analysis, produced with the relative intersection of confidence intervals algorithm and single-scale time-dependent algorithm. The comparison of methods was done on synthetic signals by calculating the energy estimation error. Additionally, an example of dynamic functional connectivity estimation is provided using real signals.