MiS Preprint Repository

The presented work introduces shortly a novel segmentation method and a modeling approach for multivariate quasi-stationary data. The combination of both parts allows the extraction of low-dimensional models from multivariate data. The segmentation method is applied both to event-related potentials and fields and early auditory evoked potentials. Additionally, the early auditory wave $Pa$ is modeled by a two-dimensional dynamical system. The segmentation method detects ERP- and ERF-components and early auditory waves objectively, which illustrates the independence of the segmentation method from the number of segments. Additionally, we find a common topology of wave $Pa$, which indicates a common underlying attractor in the brain