Analysis of video games design styles based on neuro-affective computational model

Abstract

Visual elements in video games contribute to a huge factor in design decision of the production. This research explores the science behind the psychology of design styles in video games via quantitative analysis. Previous works studying design styles and their influence over the human mind lack quantitative analysis. Their approach of using subjective experiments do not accurately provide a clear picture of the design-psyche relationship. Current guidelines for designers to decide design style is vague, relying on traditional formula of using shapes as symbols for psychological influences. In particular, affective responses towards particular Design Styles were never understood. Thus, an analysis of Design Styles using Electroencephalogram (EEG) to study its effect on the human emotions is proposed. Classifying EEG data via machine learning allows accurate numerical analysis of Design Styles and their affectivity. The findings enable the understanding of the cause and effect between design and emotions. It can help designers exploit their design potential and the research framework will be a significant contribution in the field of computer science. Subjects play 2 video games of distinct design styles – abstract and realistic - following a strict research protocol for EEG experiments. EEG data sets collected include brain’s cognitive functions, stimulated emotional responses, resting state, game playing sessions, and post-experiment state. These numerical measurements are classified using machine learning by applying a known computational model that was constructed from stimulated emotional responses. IAPS serves as the instrument for emotional stimulation. Computer models of valence and arousal are generated through PSD feature extraction while MLP algorithm is applied for the machine learning classification process. Tables and spreadsheets are then laid out for statistical analysis to correlate design and the human psyche. Results from these analysis show evidences of designs styles influences over the subjects’ brain activity. Indeed, specific cognitive functions such as memory, literacy, and reasoning are present within the game playing session, indicating a sub-conscious activity from the game that was not purposely designed to stimulate them. While the emotional state never changed between the two different games, the intensity and mood certainly show some distinctions. It seems that arousal intensity are more responsive towards abstract than realistic design. Positive valency also showed its association with abstract design. On the other hand, realism seems to be associated with stable and sustained mood. Deviation of arousal responses reveals that ‘appeal’ of the design styles depend on the person, not an influence resulted from the design styles. It points out that in the end, design styles may influence certain specific cognitive activity, but they are not substitutes for the designer’s artistic ability to make good design.