Intermittency-driven complexity in the brain: towards a general-purpose event detection algorithm

In this work we first discuss a well-known theoretical framework for the analysis and modeling of self-organized structures in complex systems. These self-organized states are metastable and rapid transition events mark the passages between self-organization and background or between two different self-organized states. Thus, our approach focuses on characterizing and modeling the complex system as a intermittent point process describing the sequence of transition events. Complexity is usually associated with the emergence of a renewal point process with power-law distributed inter-event times, hence the term fractal intermittency. This point process drives the self-organizing behavior of the complex system, a condition denoted here as intermittency-driven complexity. In order to find the underlying intermittent birth-death process of selforganization, we introduce and discuss a preliminary version of an algorithm for the detection of transition events in human electroencephalograms. As the sequence of transition events is known, the complexity of the intermittent point process can be investigated by applying an algorithm for the scaling analysis of diffusion processes driven by the intermittent process itself. The method is briefly illustrated by discussing some preliminary analyses carried out on real electroencephalograms.