Spatial Dynamics and Functional Connectivity in EEG: Insights from Lexical Processing

Abstract

Lexical processing is a core cognitive function involving the integration of semantic and symbolic information across distributed brain networks. In this study, we investigate the spatial dynamics and functional connectivity underlying lexical categorization using electroencephalography (EEG) data from a silent reading task. Participants were presented with words from two categories--social and numeric--while EEG signals were recorded and analyzed. Frequency band power and inter-electrode coherence were extracted as features to train Random Forest classifiers for word category prediction. The spatial dynamics model, based on band power from selected electrodes, achieved 100% classification accuracy and identified region- and frequency-specific contributions, notably in the theta and alpha bands across parietal and occipital regions. A complementary functional connectivity model achieved 85% accuracy, highlighting the role of inter-regional coherence in lexical differentiation. These findings demonstrate the potential of combining traditional EEG analysis with machine learning for decoding semantic categories, providing foundational insights for future brain-computer interface applications.