Actionable companion · Accessibility & inclusion

Learning with agency: Identifying dyslexia in school pupils from eye movement and demographic data using artificial intelligence

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims or recommendations attributed to the source authors. No independent peer review of this guide is asserted.

Read original paper Search paper agent CC BY 4.0

Evidence and context

Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Primary school. Verify the participant population and task before transferring the findings.

Original authors’ abstract, reproduced under the linked source license:

This paper represents our research results in the pursuit of the following objectives: (i) to introduce a novel multi-sources data set to tackle the shortcomings of the previous data sets, (ii) to propose a robust artificial intelligence-based solution to identify dyslexia in primary school pupils, (iii) to investigate our psycholinguistic knowledge by studying the importance of the features in identifying dyslexia by our best AI model. In order to achieve the first objective, we collected and annotated a new set of eye-movement-during-reading data. Furthermore, we collected demographic data, including the measure of non-verbal intelligence, to form our three data sources. Our data set is the largest eye-movement data set globally. Unlike the previously introduced binary-class data sets, it contains (A) three class labels and (B) reading speed. Concerning the second objective, we formulated the task of dyslexia prediction as regression and classification problems and scrutinized the performance of 12 classifications and eight regressions approaches. We exploited the Bayesian optimization method to fine-tune the hyperparameters of the models: and reported the average and the standard deviation of our evaluation metrics in a stratified ten-fold cross-validation. Our studies showed that multi-layer perceptron, random forest, gradient boosting, and k-nearest neighbor form the group having the most acceptable results. Moreover, we showed that although separately using each data source did not lead to accurate results, their combination led to a reliable solution. We also determined the importance of the features of our best classifier: our findings showed that the IQ, gender, and age are the top three important features; we also showed that fixation along the y-axis is more important than other fixation data. Dyslexia detection, eye fixation, eye movement, demographic, classification, regression, artificial intelligence.

What kind of evidence is this?

Supporting AI learning science · Design requires full-text appraisal

Stage: Primary school. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.

Study methods · source passage

Previous eye-tracking studies of reading Russian-speaking children with and without dyslexia are few in number and focused on comparing fixation durations, progressive saccades, and regressions in these two groups of participants ([25–27]). Their findings were consistent with the results of other studies in alphabetical languages. Namely, all three studies agreed that children with dyslexia produced longer fixations and were more sensitive to word length and frequency compared to typically-developing readers. Also, Parshina et al. (2022) applied the ScanPath method to investigate which global reading processes children in grades 1 through 5 with and without dyslexia adopted to read entire sentences. The authors identified five reading processes and concluded that children with dyslexia relied on the same processes that their typically developing peers but with a 3-year reading delay. Importantly, no previous studies of reading in Russian have ever aimed to classify readers with and without dyslexia based on their eye movements.

Limitations and uncertainty · source passage

Last but not least, in each comparison, the proximity of the top three results and the fact that each algorithm has been tuned separately and then trained and evaluated on ten different disjoint train-test splits can be considered an acceptable evaluation policy to examine the generalizability power of the algorithms. In other words, the likelihood of the occurrence of overfitting (under-fitting) for two or more algorithms over ten different data splits is less than the case in which one simply applies only one algorithm.

Design the system before the pilot

Use this study to inform a learning mechanism or implementation question. It is not direct evidence that autonomous agents improve learning. Keep the proposed agent behavior separate from what the source actually tested.

Co-design accommodations with learners and specialists. Check access and error burden across groups without inferring subgroup effectiveness from small or unreported samples.

A 90-day district pilot

Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
Days 16–30 · Design. Pilot one teacher-led unit with short guided turns. The tutor asks for an explanation and offers a graduated hint before any worked solution. Use concrete examples and an equivalent task without AI. Name an accountable owner and set the baseline, success threshold and budget before implementation.
Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
Days 61–90 · Decide. Use an age-appropriate independent post-task and a delayed parallel task. Score explanation quality and transfer, not the number of agent messages or completion rate alone. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.

The board decision brief

Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.

Attribution and reuse

Soroosh Shalileh; Dmitry Ignatov; Anastasiya Lopukhina; Olga Dragoy. Identifying dyslexia in school pupils from eye movement and demographic data using artificial intelligence. PLOS ONE (2023-11-22). https://doi.org/10.1371/journal.pone.0292047

Companion prepared 2026-10-30. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.