Emotion-Driven Gameplay Adaptation: Keeping Players in Flow

Maintaining player engagement and sustaining the game flow requires a difficulty level that adapts dynamically to the player's emotional responses, whether pleasant (e.g., joy, engagement) or unpleasant (e.g., anger, frustration). This paper explores a non-intrusive-sensing approach that integrates facial expression recognition into a minimalist arcade game, demonstrating how real-time adaptation can be achieved with a contact-free setup. The adaptation logic combines two channels: rapid corrections driven by emotional signals, and gradual adjustments based on gameplay performance trends, modulating key parameters such as speed, gravity, and responsiveness. Preliminary user observations suggested that mapping unpleasant emotions to reduce difficulty produces relief, while pleasant emotions increase challenge, maintaining alignment with flow theory and preserving playability. These findings show that emotion-driven adaptation can sustain flow in a gaming environment, providing a foundation for future multimodal and large-scale studies.