A Novel EOG-Based Approach for the Assessment and Intervention of Pediatric Dyslexia

To address the scarcity of objective diagnostic criteria and intelligent intervention methods for childhood dyslexia, this study designed and implemented an integrated assessment and intervention system based on electrooculography (EOG) signals. The assessment module utilizes low-cost, portable hardware to capture EOG signals in real-time during reading. Following preprocessing, key eye-movement features such as amplitude, slope, and acceleration are extracted and fed into a Random Forest classifier, which identifies abnormal eye movements with an accuracy of 95.38%, establishing a robust quantitative diagnostic basis. For intervention, the system presents reading materials in customizable formats. EOG signals generated during left-to-right eye traversals are processed to compute real-time reading speed. This metric drives an adaptive algorithm that dynamically adjusts the text presentation rate, enabling personalized training. To evaluate comprehension, graphical cognitive tasks are administered after each passage, requiring children to perform a cognitive judgment by selecting the image that best corresponds to the narrative. Collectively, this system offers a novel paradigm for the objective quantification and intelligent rehabilitation of pediatric dyslexia.