Multiresolution decomposition of the signal-averaged ECG of postinfarction patients with and without bundle branch block

Wavelet transformation based on multiresolution decomposition using the method proposed by Mallat (1989) was applied to the signal-averaged ECG (SAECG) of postinfarction patients with left (76), right (79) and no (195) bundle branch block (BBB) in order to discriminate and characterize spectrally these groups. The relative energy of the QRS complex was calculated in seven frequency bands ranging from 0 to 250 Hz. Left BBB patients had a significant energy reduction (p<0.0001) in the frequency range from 3.9 to 250 Hz compared with no BBB patients and from 15.6 to 250 Hz compared with right BBB patients. Right BBB patients displayed a significant energy decrease (p<0.0001) in the frequency range from 3.9 to 31.2 Hz compared with no BBB patients. The frequency band from 15.6 to 31.2 Hz showed significantly different energies among the three studied groups. Multiresolution decomposition of SAECG allowed a discrimination between right, left and no BBB as well as their spectral characterization