New technology for early discovery of reading deficits
| Reference number | |
| Coordinator | Karolinska Institutet - Institutionen för klinisk neurovetenskap |
| Funding from Vinnova | SEK 2 186 601 |
| Project duration | October 2014 - September 2016 |
| Status | Completed |
Purpose and goal
The earlier reading deficits are discovered, the more effective supportive interventions become. However, there are no fast and accurate methods available for screening today. In this project, we have developed a new method for screening that is based on analysis of eye movements during reading. Together with two municipalities, Järfälla and Trosa, we have evaluated the method in primary schools over two years. The aim was to create a screening method that could offer a high accuracy and also be practical to use in the schools.
Expected results and effects
The method we have developed is based on a unique longitudinal research material where children with and without reading difficulties were monitored from second grade in to adulthood. I third grade, eye movement recordings were made, and by analysing these using machine learning we were able to create models that could predict reading difficulties with a high accuracy. In this project, we have validated the method on a large group of children, and can see that the method predicts reading difficulties with an accuracy around 90 % already from first grade.
Planned approach and implementation
A prerequisite to be able to evaluate the method was access to schools, as we wanted to evaluate the method where it was intended to be used. A collaboration with Järfälla and Trosa municipality made this possible and over two years we have recorded eye movements during reading on 3 400 children in grade 1-3. At the same time we did recordings, we also tested their reading ability using standardized tests. About 800 children were tested twice a year apart. Out analysis shows that our method can reliably find those children that are in need of support.