Statistical Analysis of Emotional Response through Physiological Signals

Emotion plays an essential role in human interaction. This paper reports a new technique for accessing the effects of human emotional responses from physiological signals with the help of statistical analysis. For emotional elicitation, a two-dimensional valence-arousal model is used for emotion elicitation. Four types of Physiological signals, namely, Heart Rate (HR), Skin Temperatures (ST), Respiration Rate (RR) and Galvanic Skin Responses (GSR) were employed for emotional analysis. A healthy group of volunteers were exhibited a few sets of images received from International Affective Picture System (IAPS) having four different states of emotion. The experiment is non-invasive. This paper aims to identify the emotional effects expressed in human physiological signals with the help of pictorial stimuli. We conducted a series of Analysis of Variance (ANOVA) tests using the following physiological signals, namely, HR, ST, GSR, and ST, to understand the emotional effect in each of the signals. We have observed significant differences between different emotional classes in terms of four bio-medical signals like HR, GSR, RR, and ST. Significant interactions among the subjects and the stimuli are observed. The ordering among the mean values of each of the physiological signals for various emotional classes has been established. It is observed that mean response values corresponding to physiological signal ST show a complete ordering, whereas, for the signal RR, they are homogeneous.