The correlation between increased network randomization and increased TLE duration is found to be most pronounced when using a high threshold for the number of edges (k = 3). The choice of a threshold is somewhat arbitrary, but also provides information about the connections that change over time. We speculate that in this study, more random networks in patients with a longer epilepsy duration reflect a randomization of the strongest local connections. This randomization results in a less optimal local network configuration, which might reflect a higher vulnerability for seizures.
