Spatiotemporal signal analysis: recognition of interacting modes

We present a new algorithm to analyse spatiotemporal signals emerging from nonlinear systems which are close to instability. The algorithm aims at a recognition of spatial modes, both order parameters and enslaved modes, and equations describing the dynamics of the amplitudes by order parameter equations and center manifolds. We demonstrate the method by a simulated example from the field of hydrodynamics (Benard instability) and discuss results of the method applied to spatiotemporal EEG-data of epileptic seizures