The cluster index revisited: detection of mesolevel dynamical structures

We present an approach to detect mesolevel dynamical structures (MDSs), which are emerging patterns composed of elements that behave in a coherent way and loosely interact with the rest of the system. We extend the notion of cluster index to deal with discrete nonlinear dynamical systems and we propose a procedure to apply this measure to identify significant MDSs. We apply this method to several dynamical systems, both biological as well as artificial ones, and we show that it is possible to detect important functional relations among set of nodes that can not be detected by usual correlation-based approaches nor motifs analysis.