Comparison of Simson-method and Daubechies wavelet method in the detection of late fields

Late field detection for cardiac biomagnetic signals is difficult. One reason is the low signal to noise ratio. The Simson algorithm is a standard procedure to detect and quantify late fields in the terminal portion of the QRS complex using a bidirectional Butterworth filter (p=4, 40-250 Hz). The signal output includes frequencies ranging from 30 to 250 Hz. In our datasets it was difficult to detect late fields with frequency rage above 100 Hz. With the discrete wavelet transformation (DWT) using Daubechies wavelets we divided the signal in different frequency intervals with sufficient time localization. For these signal outputs we calculated the parameter RMS₄₀ QRS_DUR and LAS_DUR and compared it with the parameter extracted using Butterworth filtered signal. The parameters of the Simson-method and of the wavelet output with frequency 50±30 Hz correlate linearly. This correlation vanished for wavelet outputs with higher frequency content