Principal Component Analysis of Electroencephalography Measurements for Dyslexia

This research aimed to inquire about the Principal Component Analysis (PCA) application to the data of Electroencephalogram (EEG) to distinguish between dyslexic and non-dyslexic persons. A total of 200 children were included as participants, with 100 diagnosed with dyslexia and identified utilizing the Test of Integrated Language & Literacy Skills (TILLS) and 100 controls. The data of EEG were collected by EMOTIV's EPOC-X headsets over 3-4 months, causing a high-dimensional dataset of 70 features. PCA was used to efficiently decrease the complicated EEG data dimensionality, disclosing crucial patterns and attributes related to dyslexia. The analysis disclosed that Cluster 0 (dyslexic) highlighted notably greater values in 'A_P8', showing higher activity of the right hemisphere, on the other hand, Cluster 1 (non-dyslexic) had dominance of the typical left hemisphere. The findings of this study highlighted the PCA's effectiveness in enhancing dyslexia classification accuracy and underscored its strength for refining diagnostic accuracy and aiding targeted clinical interventions. This originality of the research lies in its approach, using PCA to reveal underlying neurophysiological markers of dyslexia, providing a novel strategy for enhancing diagnostic accuracy and contributing to more targeted clinical tactics.