Claim Missing Document
Check
Articles

Found 1 Documents
Search
Journal : International Journal of Robotics and Control Systems

Recent Advances in Artificial Intelligence for Dyslexia Detection: A Systematic Review Pamungkas, Yuri; Rangkuti, Rahmah Yasinta; Karim, Abdul; Sangsawang, Thosporn
International Journal of Robotics and Control Systems Vol 5, No 3 (2025)
Publisher : Association for Scientific Computing Electronics and Engineering (ASCEE)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31763/ijrcs.v5i3.2057

Abstract

The prevalence of dyslexia, a common neurodevelopmental learning disorder, poses ongoing challenges for early detection and intervention. With the advancement of artificial intelligence (AI) technologies in the fields of healthcare and education, AI has emerged as a promising tool for supporting dyslexia screening and diagnosis. This systematic review aimed to identify recent developments in AI applications for dyslexia detection, focusing on the methods used, types of algorithms, datasets, and their performance outcomes. A comprehensive literature search was conducted in 2025 across databases including ScienceDirect, IEEE Xplore, and PubMed using a combination of relevant MeSH terms. The article selection process followed the PRISMA guidelines, resulting in the inclusion of 31 eligible studies. Data were extracted on AI approaches, algorithm types, dataset characteristics, and key performance metrics. The results revealed that machine learning (ML) was the most widely applied method (58.06%), followed by multi-method (22.58%), deep learning (16.13%), and large language models (3.23%). Among the ML algorithms, Random Forest and Decision Tree were the most commonly used due to their robustness and performance on structured datasets. In the deep learning category, CNN were the most frequently used models, especially for image-based and sequential input data. The datasets varied widely, including digital cognitive tasks, EEG, MRI, handwriting, and eye-tracking data, with several studies employing multimodal combinations. Ensemble and hybrid models demonstrated superior performance, with some achieving accuracy rates exceeding 98%. This review highlights that AI, particularly ML and multimodal ensemble methods, holds strong potential for improving the accuracy, scalability, and accessibility of dyslexia detection. Future research should prioritize large-scale, multimodal datasets, interpretable models, and adaptive learning systems to enhance real-world implementation.