Neuroadaptive Digital Assessment in Mathematics: A Parametric Approach with STACK and AI-Powered Feedback

Abstract

The intersection of neuroscience, artificial intelligence (AI), and education is driving a transformative shift in how assessment is conceptualized and delivered. This study presents a neuroadaptive assessment framework for middle school mathematics using the STACK (system for teaching and assessment using a computer algebra kernel) plugin in Moodle. By leveraging symbolic computation, parameterised item generation, and AI-compatible feedback systems, the approach enables scalable, personalised, and cognitively aligned testing environments. Through a quasi-experimental design involving two 8th-grade cohorts, we evaluate the cognitive and logistical benefits of parametric digital testing. Findings reveal enhanced conceptual understanding, increased student engagement, and reduced teacher workload. Additionally, the model aligns with principles of neuroplasticity, adaptive learning, and reinforcement feedback, establishing a neurodidactic foundation for future AI integration in education. This approach is consistent with theoretical principles of neuroplasticity and adaptive learning (Zull, 2002) and provides a scalable pathway toward AI-enhanced assessment.