HRV-based analysis of physical activity effects in e-sports players

Heart rate variability can be used as an indicator of physiological changes that occur in the body for various reasons. In this thesis study, it was tried to understand the effects of physical activity by analyzing the changes in heart rate variability caused by virtual reality-based exercise games before and after e-sports games. In addition, with the help of the obtained data, the classifier performances of different groups were measured and the features that can most effectively reveal the distinction between the groups and the classifier models that can detect this distinction were tried to be found. For this purpose, four different groups were formed between e-sports sessions, where one of two different virtual reality exercise games (exergame) was played, both were played in a hybrid way, and neither was played. Then the heart rate variability data collected from the participants were analyzed. In the analysis of heart rate variability data, RR average, RMSSD and SDNN were used from the time domain features. In the frequency domain, LF/HF feature was used. In addition, from the Poincare technique parameters ellipse area and SD1/SD2 values were used. Besides these, stress indexes of participants were calculated. At the end of the study, it was found that playing virtual reality exercise games can change the autonomic nervous system balance in the direction of parasympathetic activity and improve participants stress indexes. The features that reveal the effects of virtual reality exercise games the most are RMSSD and SD1/SD2 features that reached the highest success rates in the classifiers. Naive bayes and simple gaussian SVM classifiers were among the successful classifiers in e-sports session data.