In this study, we introduced and applied a novel histogram transformation technique to enhance the interpretability and discriminative power of Cross-Frequency Coupling (CFC) maps derived from EEG signals for dyslexia detection. Our approach addresses the challenge of subtle intensity differences in CFC maps, which can hinder the accurate identification of dyslexia-related patterns.
Through visual inspection and quantitative analysis, we demonstrated the effectiveness of the histogram transformation technique in amplifying intensity differences within CFC maps. Specifically, our results show significant improvements in the significance of CFC map pixels, particularly in the Alpha-Beta coupling band, post-transformation. This enhancement in discriminative power was further supported by the reduction in entropy and the identification of texture feature changes through Gray-Level Co-occurrence Matrix (GLCM) analysis.
