Sampling rate and time alignment in the estimation of QRS variability

Studies sampling frequency effects in time alignment and averaging in the analysis of high-resolution ECGs (HRECG). The resolution of time alignment is limited by noise and the sampling interval. We theoretically model the sampling rate effect, both in the estimated ensemble-averaged (EA) signal and in the estimated variability with the ensemble deviation (ED) signal. We predict that for an SNR of 30 dB, the sampling frequency should be around 4 kHz to get a misalignment effect of the order of 10% of the noise effect, and then consider misalignment to be negligible. The results were quantified in terms of ED signal power and the power of the error signal. Experimental results corroborate the theoretical predictions and lead us to propose 4 kHz as the desired sampling rate for HRECG variability studies, rather than the 1-kHz rate that is routinely used in many places.