Consensus between networks based on streamline fiber tracking and probabilistic fiber tractography

Exploring structural brain network from diffusion MRI has been established as a useful method for understanding the topological organization of brain structure in health and disease. However, there is still no wildly accepted methods for constructing structural network. Definition of structural network edge is one of the most important issues in network construction. To generate network edges, streamline fiber tracking and probabilistic fiber tractography are wildly used. In this study, we explored the effect of networks based on these two different methods and investigated whether these two kinds of networks were mutually replaceable. To fulfill this, we compared four network properties (clustering coefficient, global efficiency, local efficiency, and shortest path length) between blind and sighted subjects. We found a high agreement in these properties between the two kinds of networks across a wide range of corresponding edge thresholds. Our findings demonstrated the utility of analyzing networks based on both these two methods to identify consensus results.