Study of epileptic rat's EEG using bispectrum analysis

In order to obtain a sensitive parameter to discriminate the different stages of epilepsy, we studied pilocarpine-induced epileptic rat's ECoG and EHG by bispectrum analysis method based on the assumption that EEG is nonGaussian and nonlinear signal. In this paper, we proposed a model of EEG signals according to the parameter model stimulated by nonGaussian white noise to estimate the bispectrum of EEG. The results showed that the bispectrum analysis is sensitive to the epileptic and nonepileptic EEG. From these results, the quantified parameters presenting the features of epileptic EEG can be found, which could be new evidences to clinical monitoring and predicting of seizure.