Assessing the influence of functional connectivity estimation methods on graph metrics in resting-state fMRI

Investigations of brain functional connectivity (FC) using resting state fMRI, have recently taken much interest. Various methods have been suggested for estimation of functional relationship between anatomical brain regions. Among the model based proposed methods, correlation and coherency algorithms are more commonly used. However; the stability of these methods in calculation of brain network's metrics has not been investigated. In this study, using graph theory analysis, we examined the stability of these two algorithms using small worldness indexes and variations of graph characteristics of networks. Our findings suggested that both algorithms resulted in similar properties of brain networks which is confirmed the usability of both methods in functional connectivity studies. However correlation algorithm has lower computational complexity.