How do players from different cultures actually feel and react while playing the same game—and how do those emotions show up in the body?
In the article “Multimodal Analysis of Emotions in Gaming: Understanding Cultural Influences,” Gursesli and colleagues combine AI-based Facial Emotion Recognition (FER) with Heart Rate Variability (HRV) to track both visible expressions and autonomic nervous system responses of 109 players from Italy, Japan, and Korea while they played two casual games (Snake and Matching Pairs).
Three core insights emerge from their findings:
- Facial expressions during gameplay are not universal “read-outs” of emotion; they are strongly shaped by cultural display norms. Italian players show more positive expressions (especially happiness) and fewer negative ones, Korean players show more frequent negative expressions, and Japanese players tend to restrain expressions of fear.
- Emotions are dynamic, not static. Using a Markov-inspired operator on continuous FER outputs, the study models how players move between emotional states over time. Emotional transitions (for example, how quickly players reach anger, surprise, or happiness) depend on both culture and game demands: a fast, reactive game like Snake produces different temporal emotional patterns than a more mnemonic game like Matching Pairs.
- Physiological signals add unique information. HRV-based indices—both canonical and more complex nonlinear features—are strong predictors of performance and vary across game types. For Snake, models combining FER-derived dynamic features and HRV achieve very high predictive power for performance, suggesting that integrating behavioral and autonomic signals can capture cognitive engagement and stress regulation during play.
For anyone using games beyond entertainment—training, education, assessment, or therapeutic contexts—this work shows that what is visible on the face is only part of the story, and that culture and physiology systematically shape how players experience, regulate, and perform in game-based tasks.
###########################
ACTIONABLE TAKEAWAYS:
- Treat facial expressions as culturally moderated signals, not universal emotion read-outs: design and interpret in-game emotion tracking with local display rules in mind (for example, restrained fear in Japan, more overt negativity in Korea, more overt positivity in Italy).
- When using games for training or assessment, consider temporal emotion dynamics (how quickly players move into or out of specific emotional states), not just average levels, especially in fast, high-load tasks.
- Combine behavioral measures (performance, in-game events) with physiological indices like HRV to gain more reliable insight into engagement, stress, and cognitive load during gameplay.
- Match game mechanics to the emotional–physiological profile that is needed: reactive, time-pressured games (like Snake) elicit different emotional trajectories and autonomic patterns than memory/attention-heavy games (like Matching Pairs).
- In competitive or performance-focused contexts (for example, eSports, serious games), include emotional regulation and autonomic recovery in training, since HRV-related markers are closely tied to maintaining performance under cognitive pressure.
###########################
Citation:Gursesli, M. C., Tarchi, P., Calà, F., Frassineti, L., Guazzini, A., Duradoni, M., Park, K., Thawonmas, R., You, X., & Lanata, A. (2026). Multimodal Analysis of Emotions in Gaming: Understanding Cultural Influences. IEEE Transactions on Affective Computing. Advance online publication. https://doi.org/10.1109/TAFFC.2026.3660362