Review of 'Identifying dyslexia in school pupils from eye movement and demographic data using artificial intelligence'

Reviewer: Dinesh Kalla

Publication date of review: 2024-03-19

Dinesh Kalla4

Identifying dyslexia in school pupils from eye movement and demographic data using artificial intelligenceCrossref ScienceOpen

this paper provides a significant contribution to the field of dyslexia research by leveraging AI

Average rating:	    Rated 4 of 5.
Level of importance:	    Rated 4 of 5.
Level of validity:	    Rated 4 of 5.
Level of completeness:	    Rated 4 of 5.
Level of comprehensibility:	    Rated 4 of 5.
Competing interests:	None

Reviewed article

Record: found
Abstract: found
Article: found

Is Open Access

Identifying dyslexia in school pupils from eye movement and demographic data using artificial intelligence

Soroosh Shalileh, Dmitry Ignatov, Anastasiya Lopukhina … (2023)

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.

0 comments Cited 1 times     Rated 4 of 5. – based on 1 reviews

(Latest)

Review information

ScienceOpen disciplines: Uncategorized

Review text

The paper "Identifying dyslexia in school pupils from eye movement and demographic data using artificial intelligence," presents a comprehensive study on the application of AI to detect dyslexia in school pupils based on eye movement data and demographic information. Here's a summary and evaluation of the paper:

Summary of Key Points:

Objectives: The study aimed to address previous shortcomings in dyslexia detection methods by:
- Introducing a new, comprehensive dataset combining eye-movement data with demographic information, including non-verbal intelligence measures.
- Proposing a robust AI-based model for dyslexia detection in primary school pupils.
- Analyzing the importance of different features in identifying dyslexia using the best-performing AI model.
Methodology: The research collected and annotated a significant dataset of eye movements during reading, combined with demographic data. This dataset is notable for its size, making it the largest of its kind globally and containing three class labels rather than the binary classification used in previous datasets. The study explored both classification and regression models for dyslexia prediction, employing Bayesian optimization for hyperparameter tuning across 12 classification and eight regression approaches.
Results: The findings indicated that a combination of multiple data sources (eye movement and demographic information) led to more reliable dyslexia detection solutions. The study highlighted the importance of demographic features, especially IQ, gender, and age, in conjunction with eye movement data, for effective dyslexia identification. Multi-layer perceptron, random forest, gradient boosting, and k-nearest neighbor models showed the most promising results.
Contributions: The paper contributes a large and detailed dataset for dyslexia detection, a comprehensive evaluation of various AI methods for this purpose, and insights into the importance of different features in diagnosing dyslexia. Additionally, it underscores the potential of combining eye-tracking data with demographic information for enhancing the accuracy of dyslexia detection.

Evaluation and Recommendations:

Strengths: The study's holistic approach, combining a large dataset with a wide range of AI models, is commendable. The comprehensive hyperparameter tuning process and the detailed analysis of feature importance add significant value to the research.
Areas for Improvement: While the study is robust, exploring deep learning models in more detail could uncover further insights into dyslexia detection. Additionally, validating the models on an independent dataset from a different demographic or linguistic background could enhance the generalizability of the findings.
Future Directions: Future research could focus on integrating more diverse data types, such as neuroimaging or genetic information, to further improve model accuracy. Developing an accessible tool or application based on the research findings for educators and psychologists could significantly impact early dyslexia detection and intervention strategies.

In summary, this paper provides a significant contribution to the field of dyslexia research by leveraging AI for effective detection. Its methodological rigor, combined with a comprehensive dataset, paves the way for future advancements in the accurate and early identification of dyslexia.