Bivariate spectral analysis of nonstationary cardiovascular variability series through Cross Wigner Distribution

Wigner Ville distribution has many of the desirable features for the spectral analysis of cardiovascular series during the transients of autonomic tests, but it is not able to get insight into the mutual relationships between the series, making power measurements from the spectra often unreliable. This work proposes the Cross Wigner Distribution (CWD) as a new timevariant bivariate spectral analysis method. The CWD is compared with an approach based on adaptive filtering in order to evaluate its reliability in detecting correlated components between cardiovascular series.