Complex HRV analysis during exercise

The rhythm of the heart beat at rest is affected by various physiological processes like respiration, blood pressure, body temperature regulation and arrhythmia, and also external factors at various time scales. The interbeat intervals fluctuate in a complex manner and show scale-invariance with typically a 1/f-like spectrum for frequencies below 0.1Hz, corresponding to long-range correlations. Interestingly, for subjects with severe heart disease, it has been shown that these long-range correlations are destroyed. Exercise also modifies various physiological processes and hence one can expect that there are corresponding changes in the modulation of interbeat intervals. Real-time detection of these changes could allow for novel non-invasive measurements of bio-markers.

Power spectral methods are affected by non-stationarity such as trends which perturb the signal even more during exercise. A number of novel methods for the analysis of non-stationary time series have been developed over the last decades. Some of them, like detrended fluctuation analysis, allow a robust detrending and detection of scaling exponents for correlations on different long and shorter times scales but do not provide much information about the physiological origin of the correlations. Other time-domain methods can detect characteristics of dynamic modulations of non-stationary time series but are purely empirical and can induce mixing of (physiological) modes. I shall review relevant methods, demonstrate their application to HRV, and discuss their advantages and disadvantages.

The ultimate goal shall be extraction of information from RR time series during exercise that are related to physiological processes which determine the physiological states, like exercise intensity, and certain characteristic thresholds.

Endurance training is characterized by a high total volume of hemodynamic load. As a result, endurance athletes have large cardiac cavity dimensions and extremely low resting heart rate. Often however, resting heart rate shows minor to moderate training effects in the normal population, but submaximal heart rate at a standardized exercise intensity may change more both in relative and absolute terms. Some training-induced changes are only observed at the onset, during or after cessation of exercise. Thus, studying the acute effects of exercise on heart rate variability may provide additional tools for the analysis of training effects. Exercise produces dynamic changes in autonomic activity. The onset of exercise increases sympathetic drive and reduces vagal tone.

The potential fields of application in terms of endurance training include assessment of exercise intensity, monitoring training adaptations and detecting overtraining. At present, applications for monitoring endurance training are mostly based on measurements performed at rest. In my talk I will explore whether some of the limitations related to resting measurements could be overcome by using exercise as a trigger. I will also review potential new approaches that may be useful for monitoring endurance training.