Methods of time variant bispectral analysis to investigate transient quadratic phase coupling in biomedical signals

In biosignal analysis it is often interesting to ivestigate coupling between one or more signals. Especially quadratic phase coupling plays an important role and can be analysed and quantified using the bispectral analysis. In biological systems such coupling is often bound on temporal events / processes and therefore appeas only transient. This on the other hand sets special requirements for the analysis algorithm. As a result methods for time variant bispectral analysis (bispectrum, bicoherence, phase bicoherence) in this project are developed based on adaptive recursive cumulant estimation III. order, on the adaptive recursive Fourier transformation, on the Gabor expansion and on the Wigner distribution with recursive kernel and used for the analysis of quadratic phase coupling in the EEG during burst suppression (see figure) and for heart rate components of newborns.