Exploring extensions of neurotransmitter-based emotion models

Abstract

Advancements in artificial intelligence have significantly enhanced the gaming experience, enabling more engaging and adaptive interactions between players and digital characters. A key aspect of this progress is the ability of non-player characters (NPCs) to display more lifelike realistic emotional responses that simulate the fluid and unpredictable nature of human emotions. This work presents a novel emotion model integrating Lövheim’s Cube of Emotions with Plutchik’s Wheel of Emotions, combining the dynamic aspects of the former with the detailed structure of the latter. The model was expanded from a 22-emotion, 21-point mapping to a more detailed version with 24 emotions across 52 points, allowing for better emotional differentiation. Two algorithms were upgraded and tested: an extended cube of emotions using the Euclidean distance, and the same cube incorporating fuzzy logic. Both methods showed significantly better results than their previous versions, with the Euclidean being the best overall. That indicates a more precise mapping of emotions. However, it can only return one emotion at a time. While the Fuzzy Logic method allows for more than one emotional response at the same time, associating neurotransmitters and emotions within fuzzy rules was quite complex.

Keywords

• Emotion Modeling
• Lövheim Cube
• Plutchik’s Wheel
• Fuzzy Logic
• Neuroscience-inspired AI
• Dynamic Emotional Responses