Bispectral analysis of human electroencephalogram (EEG) signals during various sleep stages

Interactions among neural signals in different frequency bands have become a focus of strong interest in biomedical signal processing. Bispectral analysis a type of higher order spectral analysis, provides with the ability to investigate such nonlinear interactions. Detection of sleep signal in EEG was commonly performed inefficiently by doctor´s eye inspection. The parametric estimation AR model has poor performance when the signal to noise ratio is small. M.R.Raghuveer and others show that the performance of estimation can be improved if the higher order spectra (Polyspectra) are used. But not only conventional bispectrum methods but also parametric bispectrum methods are difficult to implementation due to complication of calculation such as matrix multiplication and inversion. In this paper, bispectral based sleep signal analysis, which has some genuine advantages in implementation is developed by using the principle of generated auto bispectrum and cross bispectrum proposed by Xue Wang.