Spatiotemporal Eye-Tracking Feature Set for Improved Recognition of Dyslexic Reading Patterns in Children.

Area of Science:

• Neuroscience
• Developmental Psychology
• Computer Science

Background:

• Dyslexia significantly impacts academic performance and is prevalent.
• Objective tools for dyslexia detection, monitoring, and treatment are a high priority.

Purpose of the Study:

• To detect and analyze dyslexic tendencies in Serbian children using eye-tracking measures.
• To introduce novel features for dyslexia recognition and explore color influences.

Main Methods:

• Eye-tracking data collected from 30 children (15 dyslexic, 15 non-dyslexic) reading texts on varied color backgrounds.
• Analysis of nine conventional and five novel eye-tracking features using machine learning algorithms.
• Leave-one-out cross-validation for performance evaluation.

Main Results:

• Achieved a highest accuracy of 94% for dyslexia recognition using all features.
• Identified features representing fixation gaze complexity as most important for detection.
• Statistical analysis revealed high inter-subject variability in color influence on dyslexic tendencies.

Conclusions:

• The study introduces novel, separable features for dyslexia detection in Serbian children.
• Proposed features can serve as objective biomarkers for diagnosing and tracking dyslexia.
• Eye-tracking analysis offers a promising avenue for dyslexia assessment, especially in shallow orthographic systems.