Quantitative and qualitative approaches to understanding and facilitating high-level cognitive perceptual-motor skill expertise using competitive esports

Abstract

The study of complex cognitive-perceptual-motor skill acquisition remains a critical area within cognitive science, yet significant gaps persist between quantitative and qualitative research methodologies. This dissertation bridges those gaps by integrating data-driven experimental approaches with experiential, process-level accounts of expertise development. Competitive eSports serves as an ideal paradigm for studying high-level skill acquisition, providing a naturalistic environment with rich, repeatable performance metrics. Across a series of studies, latent cognitive markers distinguishing expert and novice players were examined using approaches including principal component analysis and linear discriminant analysis. Results suggest a limited role for broad latent abilities alone (e.g., isolated reaction speed or attentional measures) and greater importance for task-specific adaptation, strategic control, and higher-order training structure. To complement these findings, the dissertation includes an extended autoethnographic analysis of self-directed expertise development, introducing practical constructs such as “mental stack” and illustrating how learning dynamics unfold over time in real training contexts. Building from this mixed-methods foundation, later chapters develop and implement an information theoretic computational model of skill, formalizing expertise as bounded-optimal coupling between representation and action under explicit information costs. A two-stage optimization framework is used to map capacity trade-offs, and model outputs are linked to reaction-time interpretation through an information-to-time bridge. The dissertation also presents a novel fighting-game-like task environment and an initial reinforcement learning baseline to provide a concrete testbed for model-grounded experimentation. Together, these contributions offer a unified framework for expertise research that connects quantitative and qualitative insights, and supports both theoretical advancement and practical training design in complex performance domains.