Improvement in high-resolution ECG analysis by interpolation before time alignment

In this work the authors study the use of interpolation prior to time alignment and averaging in the analysis of high resolution ECGs. The resolution of time alignment is limited by the sampling interval and interpolation makes the alignment error reduced. Five alignment methods were used in combination with white noise and 50/60 Hz interference. The results were quantified in terms of power of the deviation signal, the deviation of the misalignment error and the energy of the error signal. A considerable decrease of the deviation signal power was noted when the sampling rate was increased from 1 kHz to 4 kHz. Modest improvements were only obtained when the sampling rate was further increased, making 4 kHz suitable sampling rate for high frequency studies. In actual and simulated ECG records it was found that the filtered QRS duration index increased only 4 ms whereas RMS40 and LAS40 remained unchanged.