Correcting the Activity-Specific Component of Heart Rate Variability Using Dynamic Body Acceleration Under Free-Moving Conditions

Abstract

Heart rate variability (HRV) analysis is a widely used technique to assess sympatho-vagal regulation in response to various internal or external stressors. However, HRV measurements under free-moving conditions are highly susceptible to subjects’ physical activity levels because physical activity alters energy metabolism, which inevitably modulates the cardiorespiratory system and thereby changes the sympatho-vagal balance, regardless of stressors. Thus, researchers must simultaneously quantify the effect of physical activity on HRV to reliably assess sympatho-vagal balance under free-moving conditions. In the present study, dynamic body acceleration (DBA), which was developed in the field of animal ecology as a quantitative proxy for activity-specific energy expenditure, was used as a factor to correct for physical activity when evaluating HRV in freely moving subjects. Body acceleration and heart inter-beat intervals were simultaneously measured in cattle and sheep, and the vectorial DBA and HRV parameters were evaluated at 5-min intervals. Next, the effects of DBA on the HRV parameters were statistically analyzed. The heart rate (HR) and most of the HRV parameters were affected by DBA in both animal species, and the inclusion of the effect of DBA in the HRV analysis greatly influenced the frequency domain and nonlinear HRV parameters. By removing the effect of physical activity quantified using DBA, we could fairly compare the stress levels of animals with different physical activity levels under different management conditions. Moreover, we analyzed and compared the HRV parameters before and after correcting for the mean HR, with and without inclusion of DBA. The results were somewhat unexpected, as the effect of DBA was a highly significant source of HRV also in parameters corrected for mean HR. In conclusion, the inclusion of DBA as a physical activity index is a simple and useful method for correcting the activity-specific component of HRV under free-moving conditions.

Effects of Heat Stress on Heart Rate Variability in Free-Moving Sheep and Goats Assessed With Correction for Physical Activity

“…However, it is noticed that the correlation between THI and HRV was significant but considerably weak, which might be due to the existence of thermoneutral range for tested animals during the experimental periods. In addition, despite the fact that the animals were housed in an animal shelter and their physical activity levels were low, VeDBA was significantly correlated with all HRV parameters (Table 4), which was in accordance with our previous study (27). Furthermore, the absolute values of the correlation coefficients between VeDBA and the HRV parameters were mostly higher than those between the THI and the HRV parameters.…”

“…where Y ijkl is the value of the mean HR and HRV parameters (SDNN, RMSSD, HF, LF/HF, L max , and %DET) per 5 min, µ is the overall mean, catTHI i (i = 1-5) is the fixed effect of the THI category, Species j (j = 1 or 2) is the fixed effect of species, Animal k(j) (k = 1-6) is the random effect of individual animals nested within species, and e ijkl is the error. Furthermore, based on the method by Oishi et al (27), the above statistical model was transformed into the following model in order to include the effect of quantified physical activity on mean HR and HRV parameters:…”

“…One of the possible methods to consider the effect is to delete the sections of the heart inter-beat interval data with excessive movement before the HRV analysis (48-50), but there have been no general ways for separating the effect of changes in quantified physical activity levels on HRV from the influence of other regulatory processes (24). Therefore, in this study, the vectorial dynamic body acceleration (VeDBA), which can be a proxy for the activity-specific energy requirements of animals (25,51,52), was used to remove the confounding effect of physical activity on HRV based on the method by Oishi et al (27). To calculate VeDBA, triaxial acceleration values obtained by the acceleration loggers were used in the following equation:…”

“…In fact, HRV is greatly affected by the behavior of animals, and some studies have focused on the effects of behavior on HRV, such as body posture (9), step behavior (17), and behavioral types (26). When the effects of psychophysiological changes on HRV of animals have to be evaluated, particularly under freemoving conditions, the quantified physical activity level of the target animals should also be considered as a key element that influences HRV (27).…”

An external telemetry system for recording resting heart rate variability and heart rate in free-ranging large wild mammals

“…Here, we have described and assessed a robust, externally mounted heart rate monitor for use in wild mammals, that delivers millisecond precision measures for IBI’s, allowing computation of key physiological parameters, in particular, measures of resting HRV when coupled with information on individual activity [ 38 , 39 , 58 ]. Our system allows for collection of IBI data on many individuals concurrently over prolonged time-periods.…”