Fig 5. The community structure of the SCN for 4 different samples (A,B,C,D) as resolved by our method (top) and by the Rubinov and Sporns [14] approach.

The panels represents the partitions detected, where each community is marked with a different colour. We can see that the Rubinov Sporns method shares a good resemblance to the clear core periphery structure that is detected by our method. In contrast to the random matrix approach, the signed Leuven consistently detects a third module, in a different location at each time.