We describe techniques to sonify rhythmic activity of epileptic seizures as measured by human EEG. Event-based mapping of parameters is found to be informative in terms of auto- and cross-correlations of the multivariate data. For the study, a group of patients with childhood absence seizures are selected. We find consistent intra-patient conservation of the rhythmic pattern as well as inter-patient variations, especially in terms of cross-correlations. The sound synthesis is suitable for online sonification. Thus, the application of the proposed sonification in clinical monitoring is possible.