🤖 AI Summary
To address the limitation of conventional attention assessments for children with neurodevelopmental disorders (NDDs)—which rely solely on game scores and fail to capture authentic learning processes—this paper proposes a multimodal adaptive assessment framework. The framework integrates three complementary modalities: eye-tracking data (spatial attention), interaction timing sequences (sustained engagement), and in-game behavioral logs (task performance), fused via a dynamically weighted scoring model. It incorporates progressive difficulty adjustment and an adaptive multiplier mechanism to mitigate score inflation for high performers while enhancing sensitivity to improvement among low-performing individuals. Crucially, it establishes, for the first time, an explicit mapping between oculomotor features and educational behaviors. Validation using MAE, RMSE, and bivariate correlation analysis demonstrates significantly improved modeling accuracy of the attention–learning relationship; assessment outputs meet quality thresholds for real-world educational deployment, providing interpretable and actionable quantitative evidence for personalized intervention.
📝 Abstract
This paper introduces an innovative adaptive scoring framework for children with Neurodevelopmental Disorders (NDD) that is attributed to the integration of multiple metrics, such as spatial attention patterns, temporal engagement, and game performance data, to create a comprehensive assessment of learning that goes beyond traditional game scoring. The framework employs a progressive difficulty adaptation method, which focuses on specific stimuli for each level and adjusts weights dynamically to accommodate increasing cognitive load and learning complexity. Additionally, it includes capabilities for temporal analysis, such as detecting engagement periods, providing rewards for sustained attention, and implementing an adaptive multiplier framework based on performance levels. To avoid over-rewarding high performers while maximizing improvement potential for students who are struggling, the designed framework features an adaptive temporal impact framework that adjusts performance scales accordingly. We also established a multi-metric validation framework using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Pearson correlation, and Spearman correlation, along with defined quality thresholds for assessing deployment readiness in educational settings. This research bridges the gap between technical eye-tracking metrics and educational insights by explicitly mapping attention patterns to learning behaviors, enabling actionable pedagogical interventions.