Figure 3
FC states and comparison with Static FC. Five recurrent FC patterns, or states, were obtained from clustering the leading eigenvectors of the dFCs of all participants. (A) Each of the five FC patterns is represented by a vector V C , where V C (n) weighs the contribution of each brain area n to that pattern (displayed in cortical space). Elements with the same sign in V C are colored alike to illustrate the network partition captured by V C (see Methods - FC Leading Eigenvector). States are ranked according to their probability of occurrence (PC, in %). (B) V C V C T illustrates the NxN connectivity pattern corresponding to each state. (C) dFC averaged over the time points of each state. (D) The static FC averaged over all subjects (top) correlates strongly (r = 0.839) with the weighted sum of the five connectivity patterns captured by V C V C T (middle). The correlation increases up to r = 0.997 when considering the weighted sum of the dFC averaged over the time points of each state (bottom).
