Software for Rapid Detection of Dyslexia Developed in Russia
HSE scientists have developed a software tool for assessing the presence and degree of dyslexia in school students based on their gender, age, school grade, and eye-tracking data. The application is expected to be introduced into clinical practice in 2024. The underlying studies were conducted by specialists in machine learning and neurolinguistics at the HSE AI Research Centre.
The software, which employs a machine learning model for diagnosing reading disorders in children based on eye-tracking data, is intended for use by psychologists, speech therapists, and medical practitioners.
The application—Dyslector—will be accessible across multiple platforms, including desktop computers running Windows and macOS, as well as mobile devices operating on Android and iOS. While the participant reads sentences from a laptop screen or mobile device, their eye movements are captured using an eye-tracker. The software user inputs the participant's gender, school grade, and age, along with the time and coordinates of gaze fixation, and the software generates information regarding the risk of dyslexia or confirming its presence.
The software does not require the involvement of a data scientist and makes it possible, within a very short time, to detect reading disorders in children and confirm the presence of dyslexia. Upon entering the necessary data, the user clicks the 'Assess Degree of Dyslexia' button, and the application displays the model's prediction: normal, risk of dyslexia, or dyslexia.
Director of the Centre for Language and Brain, Head of the project 'Diagnostic and assistive speech technologies based on artificial intelligence' at the HSE AI Research Centre
'The machine learning methods employed in this project rely on extensive and unique bodies of data. The software makes it possible to detect dyslexia risk in children based on eye movements and to assess the oculomotor parameters that could contribute to reading difficulties. A unique aspect of this tool is its ability to detect reading disorders or dyslexia within a very brief period, unlike traditional neuropsychological or speech therapy assessments, which typically require significantly more time and the involvement of a specialist.'
The AI Research Centre was created at HSE University as part of the 'Artificial Intelligence’ federal project within the 'Digital Economy' national programme.