Decoding Dyslexia from Electroencephalogram Signals Using a Graph-Based Neural Network
AI, Computer Science and Robotics Technology · 05.05.2026 · 5
DOI: 10.5772/acrt.20250147 ↗Kurz gefasst
Fachartikel in der Zeitschrift AI, Computer Science and Robotics Technology, online erschienen am 05.05.2026. Verfasst von Madhuchhanda Basak und 2 weiteren Autor:innen. Die Arbeit nennt 17 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
A graph neural network architecture based on GraphSAGE is presented in this work for the classification of electroencephalogram (EEG) data in the context of dyslexia detection. The proposed framework employs data preprocessing, class balancing, and k -nearest neighbor graph construction using cosine similarity to capture inter-sample relationships. By aggregating neighborhood information, the GraphSAGE model learns discriminative representations that enable accurate differentiation between dyslexic and non-dyslexic subjects. Subject-independent evaluation was performed using leave-one-subject-out cross-validation to ensure clinical generalization across unseen individuals. Experimental results demonstrate strong performance with high accuracy and an Area Under the Receiver Operating Characteristic Curve of 0.97, indicating effective generalization. Visualization through graph mappings and stands for t-distributed Stochastic Neighbor Embedding confirms clear class separability and meaningful structural learning. The proposed scalable and interpretable framework offers promising potential for EEG-based dyslexia classification, cognitive monitoring, and early neurological disorder diagnosis.
Zitieren
APA
Basak, M., Maiti, D., & Chaudhury, D. (2026). Decoding Dyslexia from Electroencephalogram Signals Using a Graph-Based Neural Network. AI, Computer Science and Robotics Technology, 5. https://doi.org/10.5772/acrt.20250147
Vancouver
Basak M, Maiti D, Chaudhury D. Decoding Dyslexia from Electroencephalogram Signals Using a Graph-Based Neural Network. AI, Computer Science and Robotics Technology. 2026;5. doi:10.5772/acrt.20250147
BibTeX
@article{Basak2026,
author = {Basak, Madhuchhanda and Maiti, Diptadip and Chaudhury, Debasis},
title = {Decoding Dyslexia from Electroencephalogram Signals Using a Graph-Based Neural Network},
journal = {AI, Computer Science and Robotics Technology},
year = {2026},
volume = {5},
doi = {10.5772/acrt.20250147}
}
EUVIMED