Kolmogorov complexity of finite sequences and recognition of different preictal EEG patterns

The problem of an adequate quantitative interpretation of epileptic EEG recordings is of great importance in the understanding, recognition and treatment of epilepsy. In recent years, much effort has been made to develop computerized methods which can characterize different interictal, ictal and postictal stages. The main issue of whether there exist a preictal phenomenon is unresolved. In this paper, we address this issue making use of the most basic representation of data complexity, namely the algorithmic information content. In general this measure, also known as Kolmogorov complexity, represents the compressibility of the data strings. It can also be used to describe properties (linear and nonlinear) of the underlying dynamical system. We analyze Kolmogorov complexity and related characteristics of intracranial EEG recordings containing preictal, ictal and postictal segments