Algorithm for screening students for dyslexia through reading. Data collected with the use of eye-tracking technology in the Greek language

Abstract

In this study, the reading process was evaluated with the use of eye-tracking technology. Emphasis was placed on the specific learning disability of reading and writing (developmental dyslexia), which leads to difficulties in reading accuracy and fluency, among other things. While developmental dyslexia is not an oculomotor problem, the difficulties in decoding and fluency are depicted in fixations and saccadic movements during reading. Using data recorded using eye-tracking and the analysis of the reading path through eye movements, a reading evaluation tool was developed, distinguishing students with and without reading difficulties. In this study, 137 native Greek students participated, aged 8.1 to 15.6 years, of whom 62 were girls and 75 were boys. All participants read two texts of increasing difficulty in the Greek language and were evaluated while reading silently. The results of the study revealed that the success rate in student distinction was 90%. In particular, 63 of the 68 typical readers (93% specificity) and 60 of the 69 readers with developmental dyslexia (87% sensitivity) were evaluated and correctly classified by the model, under circular validation. Taking into account how important is early intervention for solving problems in language skills like reading, the tool of this study records eye movements during reading (i.e. it records the reading path which is unique for each reader), quantifies the data and identifies the readers with poor reading skills, in realation to their age. The use of this tool is to detect students with reading difficulties and help specialists create customized treatment plans (complementary with other tools) and track the progress of such treatment plans.