Heart rate variability during exercise: A comparison of artefact correction methods

Abstract

There is a need for standard practice in the collection and processing of RR interval data recorded using heart rate monitors (HRM) in research. This paper assessed the validity of (a) RR intervals and heart rate variability (HRV) data obtained using a HRM during incremental exercise, and (b) artefact correction methods. Eighteen participants completed an active orthostatic test and incremental running V̇O2MAX test, while simultaneous recordings using the V800 and an electrocardiogram were made. Artefacts were corrected by deletion; degree zero, linear, cubic and spline interpolation; and using Kubios HRV software. Agreement was assessed using percentage bias and effect size (ES), intra-class correlation coefficients (ICC) and Bland–Altman limits of agreement (LoA). The number of artefacts increased relative to the intensity of the exercise, to a peak of 4.46% during 80-100% V̇O2MAX. Correction of RR intervals was necessary with unacceptably increased percentage bias, LoA and ES and reduced ICC in all but supine and standing recordings. All correction methods resulted in data with reduced percentage bias and ES for resting and <60% V̇O2MAX exercise recordings. However, at >60% V̇O2MAX, even when correction methods were applied, large amounts of variation were present in RMSSD, LF:HF ratio, SD1 and SampEn. Linear interpolation produced corrected RR intervals with the lowest bias and ES. However, caution should be given to HRV parameters at high exercise intensities, as large amounts of variation were still present. Recommendations for minimising recording artefacts are discussed, along with guidelines for their identification, correction and reporting.