A comparative study of virtual reality based cognitive training games through frameworks

Abstract

This thesis presents a comparative analysis of virtual reality-based cognitive training games through the lens of serious game design frameworks. The primary aim is to evaluate the extent to which these commercially available VR cognitive training (VRCT) products align with theoretical design principles and cognitive development objectives. The study focuses on three selected games (NeuroBoostVR, Optics Trainer, and REAKT Performance Trainer) each examined using a distinct serious game framework: the Framework for Adaptive Serious Games, the Framework Based on Research Community Insights, the Cognitive Game Design Analysis, and the Cognitive Training Game Framework. These frameworks allow for a multi-dimensional comparison across layers such as difficulty structure, feedback systems, narrative integration, user interaction, and motivational elements. The methodology adopted is a qualitative comparative content analysis. Each game was played and directly experienced, with attention paid to task structures, feedback types, UI elements, and interaction models. Game features were systematically recorded and mapped onto structured analysis tables derived from the selected frameworks. This process allowed for consistent cross-game comparisons while also highlighting shared and divergent design features. Findings suggest that while all three games aim to enhance cognitive performance, they do so through different strategies and user profiles. NeuroBoostVR targets older adults and individuals with neurological conditions with simplified task structures. Optics Trainer focuses on visual-cognitive skills through eye-based mini-games. REAKT Performance Trainer integrates attention, multitasking, and sensorimotor coordination, primarily for athletes. Although the games share a common purpose, their design decisions, particularly in feedback systems, challenge progression, interaction mechanisms, and narrative integration, differ in sophistication and theoretical alignment. Feedback systems were mostly auditory and visual, with limited use of haptic feedback. REAKT offered more robust performance analytics, such as percentile scores and session graphs, while others relied on basic or static feedback. Difficulty structures also varied; adaptive or user-based scaling was uncommon, and many games offered fixed difficulty levels without transition guidance. This contrasts with the frameworks, which emphasize dynamic difficulty for sustained engagement and cognitive challenge. Narrative elements were minimal or indirect in most games. Only REAKT incorporated environmental storytelling and motivational voiceovers to enhance immersion. Interaction mechanics also varied in complexity: while NeuroBoostVR and Optics Trainer primarily utilized basic VR controls like teleporting or object manipulation, REAKT included more advanced mechanics involving motion tracking and attention-based tasks. Motivational design elements showed significant divergence. NeuroBoostVR offered basic point-collection, whereas Optics Trainer used personal score tracking to motivate players. REAKT integrated competitive features, weekly goals, and rewards systems to strengthen user engagement. In conclusion, the analysis reveals that current VRCT games exhibit varied levels of design maturity across critical dimensions. While some components align with theoretical expectations, others fall short, especially in adaptive design and integrated learning paths. To address this gap, this thesis introduces the Virtual Reality-Based Cognitive Training Games Framework, a model specifically developed to integrate cognitive training objectives with the affordances of VR technology. The proposed framework contributes to academic literature by offering a structured, multidimensional approach for evaluating VRCT games, and it supports practical applications by guiding developers in aligning design elements with cognitive goals. In this way, VR-based cognitive training games can be designed and evaluated more effectively, both scientifically and practically.