A Whitening Filtering Approach for the Detection of Abnormal Intra-QRS Potentials in Signal-Averaged Electrocardiograms

The detection of extremely low-amplitude and high-frequency abnormal intra-QRS potentials (AIQP) in signal-averaged electrocardiograms is a potential technique to improve the diagnostic performance of the risk of ventricular arrhythmias. Instead of using the modeling residuals of the previous studies, this study develops a whitening filtering approach to whiten the input QRS wave and use the optimized filter coefficients to detect the presence of AIQP due to the high-frequency characteristics of AIQP that may increase the higher order coefficients. To demonstrate the performance of the proposed method for detecting the AIQP, a color noise with a frequency band ranged from 40 Hz to 250 Hz is embedded into the normal QRS wave to simulate AIQP, and the detection performance of AIQP is tested under four signal-to-noise ratios (SNRs) (AIQP vs. QRS) including -50 dB, -46 dB, -40 dB, and -34 dB. The study results demonstrate that the presence of AIQP can increase the amplitude of the higher order coefficients of the whitening filter. The accuracy of AIQP detection can at least reach 75%, 83%, 92%, and 96% under SNRs of -50 dB, -46 dB, -40 dB, and -34 dB, respectively, in the X, Y and Z leads.