Emotion Recognition in a Multi-Componential Framework: The Role of Physiology

Abstract

Emotions are rich and complex experiences involving various behavioral and physiological responses. While many empirical studies have focused on discrete and dimensional representations of emotions, these representations do not fully reconcile with recent neuroscience studies that increasingly suggest a multi-process mechanism underlying emotional experience. Moreover, the latter view accords with psychological theories that consider emotions as multicomponent phenomena, such as appraisal theories. Although there is no complete consensus on the specific components of emotions and fundamental principles defining their organization, the Component Process Model (CPM) is well established framework describing an emotion as a dynamic process with five major highly interrelated components: cognitive appraisal, expression, motivation, physiology and feeling. Yet, few studies have systematically investigated a range of discrete emotions through this full multi-componential view. In the present study, we therefore elicited various emotions during movie watching and measured their manifestation across these components. Our primary goal was to investigate the relationship between physiological measures and the theoretically defined components of emotions. In addition, we also investigated whether discrete emotions could be predicted from information provided by the multicomponent response patterns, as well as the specific contributions of each component in such predictions. Results suggest that physiological features are interrelated to all other components of emotion, but the least significant predictors for emotion classification. Overall, emotion prediction was significantly higher when classifiers were trained with all five components. The findings therefore support a description of emotion as a dynamic multicomponent process, in which the emergence of a conscious feeling state requires the integration of all the components.