The Impact of Gravitational Stress on Cardiac Dynamics Using Entropy

Until now, the gravitational stress effect on the time domain and frequency heart parameters has been well-documented. However, cardiac signal dynamics have not been studied adequately under the influence of postural changes. In addition, the effect of body positions on the bio-signals has been investigated only from the aspect of feature extraction and the classification problem has not been considered. Among the physiological signals, the heart rate (HR) becomes an emerging modality that captured the attention of many researchers due to its noninvasive recording and its ability to assess autonomic tone modulation. This study attempted to classify cardiac dynamics concerning postural changes by evaluating different entropy algorithms. In this study, the ECG signals of 10 participants (five women and five men with a mean age of 28.7±1.2 years) were designated from the database available at Physionet. First, the RR-intervals of electrocardiograms complying with the Pan-Tomkins procedure were estimated. Second, several entropy measures, including Shannon, log energy, sample, differential, Tsallis, Renyi, and approximate entropy, were calculated while participants were in supine rest, in two rapid head-up tilts, two stand-ups, and two slow head-up tilts. Then, we applied the support vector machine to classify different postures using one group vs. all other remaining groups (OVA) and one body posture vs. the resting supine position (BVR) in a k-fold cross-validation scheme. Empirical results showed that using the entropy measure in a BVR scheme leads to higher of accuracy rates up to 100%. This framework opens an avenue of research for different gravitational stress-based conditions in a broad range of applications like disease management, sports, and astronautics.