Variations of HRV analysis in different approaches

Abstract

The analysis of normal-to-normal (NN) intervals acquired from a continuous electrocardiogram (ECG) record is a standard method to evaluate the variations in heart rate. For the advantage of responding to the pumping action of the heart, photoplethysmography (PPG) has also been used extensively in the analysis of heart rate variability (HRV). However, there is little literature available on the variation between the HRV analysis derived from ECG and PPG. In the research described here, experiments of recording short-term (les 5 minutes) ECG and PPG signals simultaneously from healthy subjects (male, N=10) under control were carried out to investigate the possibility of such variation. Automatic computer analysis is provided for the analysis of correlation coefficient and the LF/HF ratio by autoregressive (AR) spectral analysis for evenly resampled sequences. The identical results are highly expected. However, the correlation coefficient between RRI (R-R interval from ECG) and PPI (peak-to-peak interval from PPG) is 0.86plusmn0.15, which should be unity for perfectly matched patterns. In additions, the relative LF/HF ratios are 2.49plusmn1.13 (for ECG) and 2.73plusmn0.82 (for PPG) respectively. Though there is no statistical difference, the worst likelihood ratio (LR) reaches the deviation of 19.04%. From the experimental results, it can be appreciated that there is indeed variation for HRV analysis in two different approaches even for healthy subjects under well-controlled conditions. For abnormal subjects in clinical applications, such variation may be expected to become more apparent. Though the variation is minor, it is suggested to obey the standard measure of HRV proposed by Task Force for consistent conclusions.