Review of EEG-based pattern classification frameworks for dyslexia.

Area of Science:

• Neuroscience
• Cognitive Science
• Biomedical Engineering

Background:

• Dyslexia is a common learning disability affecting reading and writing skills.
• It often remains undiagnosed due to the absence of other outward symptoms.
• Electroencephalography (EEG) is being explored as a tool to detect unique brain activity patterns associated with dyslexia.

Purpose of the Study:

• To critically evaluate existing EEG-based pattern classification frameworks for dyslexia.
• To identify the strengths and weaknesses of current methodologies.
• To propose optimizations for enhancing future research in this area.

Main Methods:

• Literature review focusing on data collection, pre-processing, analysis, and classification in EEG studies for dyslexia.
• Analysis of various input features and classification algorithms.
• Identification of effective techniques for signal processing and pattern recognition.

Main Results:

• Incorporating reading and writing tasks during EEG data collection can improve framework performance compared to simple tasks.
• Artefact subspace reconstruction is effective in minimizing noise from body movements in EEG signals.
• Support vector machine (SVM) emerged as a highly promising classifier for EEG-based dyslexia detection.

Conclusions:

• EEG offers a viable method for identifying dyslexia through distinct brain patterns.
• Optimized data collection strategies and advanced signal processing techniques are crucial for improving diagnostic accuracy.
• Further research should focus on refining these methods, particularly utilizing SVM for classification.