Multichannel spectral estimation of ECGs

Multichannel spectral estimation and coherence analysis offers a new approach to the study of electrocardiograms in order to help improve our technique of frequency band adaptive filtering. A comparison between the frequency spectra of inputs enables us to determine an optimal configuration for the filter bank. Such a comparison must be considered in the context of multichannel spectral estimation and coherence analysis. A parametric modelling approach has been used to estimate the multichannel power spectral densities for the coherence function as this offers a higher frequency resolution and does not suffer from bias present in FFT-based methods. The Vieira-Morf multichannel spectral estimation algorithm was found to offer the best resolution without introducing cross terms in the spectral estimates.