Dyslexia Diagnostics Based on Eye Movements and Artificial Intelligence Methods: A Review

Abstract

<p style="text-align: justify;">The review considers methods of dyslexia diagnostics based on eye movement data and implemented on the basis of artificial intelligence. A number of studies have shown that eye movements in people with dyslexia may differ from those of people with normal reading abilities. Since 2015, studies have begun to appear in which the eye movements of observers with and without dyslexia were analyzed using various artificial intelligence methods. To date, there are a number of papers using both simple and more complex models (with neural networks and deep learning). This review discusses what accuracy of diagnosis has been achieved by researchers, for which groups of subjects and for which languages the current results have been shown, what types of algorithms have been used, and other practical aspects of conducting such diagnosis. According to the data analyzed, dyslexia diagnostics by eye movements and artificial intelligence methods is very promising and may have a significant impact on early diagnosing of reading problems.</p>