Application of Connected Iterative Scan in Biological Neural Network.

The current study applied connected iterative scan (CIS), a graph-theoretical-based approach that was previously developed to detect clusters among social networks, in detecting the functional clusters among the biological neural network of human brain extracted from resting-state functional MRI data. Unlike traditional clustering or modularity methods, CIS allows for overlapping among the detected clusters. CIS was tested on a simulation dataset as well as a biological dataset. CIS was able to detect the overlaps designed in the simulation dataset; CIS was also able to detect the designed overlap in the simulation dataset, and it also detected the overlap between the two networks in the biological dataset, which provided new knowledge about the architecture of the biological neural network. In conclusion, CIS is shown to be useful for revealing the delicate structures among the functional clusters in the biological neural network.