The problem of detecting a possible pre-seizure state in epilepsy from EEG signals, has been addressed by many authors over the past decades but still remains unsolved up to now. Different approaches of time series analysis of brain electrical activity are already providing valuable insights into the complex dynamics of the brain and may lead to the extraction of signal features that are able to identify an impending epileptic seizure with sufficient specificity and reliability. In this contribution models based on Cellular Nonlinear Networks (CNN) are considered to analyze intracranial EEG, taking into account mutual dependencies between neighboring electrodes. Solutions of Reaction-Diffusion CNN (RD-CNN) models are used in order to approximate short segments of EEG-signals. In comparison the behaviour of linear spatio-temporal systems is evaluated.