Understanding the Effect of Sound of a Horror Audio-Visual Stimulus on R-R Interval Signal Using Recurrence and Empirical Mode Decomposition analyses

Abstract

The present study delineates the analysis of the cardiac physiology after providing an audio visual stimulus, with and without sound. In the study, the RR Interval signals (RRISs) were extracted from the electrocardiogram (ECG) signals. The RRISs were analyzed using recurrence analysis and the important features were predicted. In a separate study, the RRISs were decomposed using empirical mode decomposition (EMD) into Intrinsic Mode Functions (IMFs) and the statistical features of the IMFs were calculated. The important features were estimated by t-test, Boosted Tree (BT), Classification And Regression Tree (CART) and Random Forest (RF) based classifiers. These predictors were used for the Artificial Neural Network (ANN) based classification. The classification efficiency suggested that the features obtained from the recurrence analysis were far superior in estimating the cardiac physiology as compared to the features obtained from the IMFs after EMD decomposition.