Neuro-Adaptive Pedagogy and the Cognitive Elasticity Model: Regulating Effort and Engagement in AI-Mediated Learning Environments[Pédagogie neuro-adaptative et modèle d’élasticité cognitive : réguler l’effort et l’engagement dans les environnements d’apprentissage médiés par l’IA]

Adaptive quizzes are increasingly used to individualize practice in digital learning, but IS research still under-specifies how AI-driven adaptivity regulates student effort during task performance. This paper addresses that gap by proposing the Cognitive Elasticity Model, which explains learning engagement as depending not only on challenge-skill balance or cognitive load alone, but on the learner's capacity to remain within a productive effort zone as task demands shift. Grounded in Flow Theory and Cognitive Load Theory, the model positions adaptive quizzing as a system-level mechanism for regulating effort over time rather than only improving assessment efficiency. The study examines this claim through a three-condition experiment with 180 students: a Moodle adaptive quiz using Bayesian Knowledge Tracing, a non-adaptive Moodle quiz with identical items and timing, and an equivalent paper-based quiz. Cognitive load and flow are assessed primarily through phase-level self-report measures. These are complemented by EEG-derived indicators, including the Cognitive Load Score (CLS) for workload regulation and theta/alpha markers associated with flow-related engagement, while physiological strain is assessed through repeated ECG-based mean heart rate snapshots interpreted conservatively as indicators of stress-related arousal rather than direct readouts of internal states.