Modeling User Experience in Electronic Entertainment Using Psychophysiological Measurements

Analyses of user experience in electronic entertainment industry currently rely on self-reporting methods, such as surveys, ratings, focus group interviews, etc. We argue that self-reporting alone carries inherent problems - mainly the subject bias and interpretation difficulties - and therefore should not be used as a sole metric. To deal with this problem, we propose a possibility of creating a model of consumer experience based on psychophysiological measurements and describe how such model can be trained using machine learning methods. Models trained exclusively on real-time data produced by autonomic nervous system and involuntary physiological responses is not susceptible to subjective bias, misinterpretation and imprecision caused by the delay between the experience and the interview. This paper proposes a potentially promising direction for future research and presents an introductory analysis of available biological data sources, their relevance to user experience modeling and technical prerequisites for their collection. Multiple psychophysiological measurements (such as heart rate, electrodermal activity or respiratory activity) should be used in combination with self-reporting methods to prepare training sets for machine learning models. During our initial experiments, we collected time-series heart rate data for two computer games - Hearthstone and Dota 2. This preliminary analysis suggests the existence of a correlation between psychophysiological measurements and in-game events. Actual ready-to-use user experience models are out of the scope of this paper.