Quantitative features in determining the cardiorespiratory synchronism from heartbeat and respiratory signals captured by wearable sensors

Nonlinear interactions between the respiratory and the cardiac systems determine the cardiorespiratory synchronism that has been recently inspired as a beneficial method to categorize the health of a person's nervous system. However, features that may quantify the synchronism have not been explored adequately. The objective of this thesis work is to quantify the synchronization due to different audiovisual emotional elicitations. Sets of images and sounds selected from the international affective picture system and international affective digital sounds are established to stimulate five different emotions namely negative low arousal, positive low arousal, negative high arousal, positive high arousal and neutral low arousal. The sets are presented to twenty heathy volunteers in a pre-defined order and the heartbeat and respiration signals from the volunteers are recorded simultaneously by using an electronic circuit designed to house two wearable sensors. The signals are transferred to a personal computer and processed. Poincare analyses are performed to extract a total of nine features quantifying the heartbeat and the respiration cycle variability for each stimulus. Multinomial logistic regression models utilizing the features are developed and statistical tests are performed to determine the performance of each feature and each model.