Integrating Valence and Arousal Within an Agent-Based Model of Emotion Contagion

Existing agent-based models of emotion contagion that account for the emotional diversity in groups have mostly focussed on the spread of categorical emotions (happy, sad, angry). In practice this raises problems with regard to how the spread of different emotions should interact. Can one be both very happy and very angry at the same time, or shift quickly between these states? And which emotion should determine the behaviour of an agent? The present paper explores an alternative where dimensional emotions spread in a crowd and the emotional state of an agent equals its location in the valence-arousal space, corresponding to a single emotion. We propose an agent-based model that is an extension of the ASCRIBE model. Furthermore, building on recent work that found an attention bias in participants toward emotionally salient stimuli, we examine the effects of attention bias in the context of emotion contagion at the crowd level. We have simulated a crowd in a soccer arena wherein several types of visitors react differently to the same events in the game (goals), with and without attention bias. Our results give a first indication that a dimensional approach to emotion contagion has the potential to solve these challenges without the need to model mood as an extra layer, though further study is required with regard to model validation and the translation from emotional state to behaviour in order to accurately simulate the complexity of real-world crowds that are emotionally diverse