Abstract

Background: Power spectral analysis of the occipital cortex is essential for characterizing brain activity during attentional and relaxed states. Objectives: This study aims to develop a predictive model capable of distinguishing between eyes-closed (EC) and eyes-open (EO) states using only two electrodes (O1 and O2), through analysis of power spectral density (PSD) and an interhemispheric asymmetry index. Method: EEG recordings from 33 seventh-and eighth-grade students were processed using the Fast Fourier Transform (FFT) and analyzed with a logistic regression model employing a Cauchit link function. Results: The model yielded an AUC of 84.2%, with satisfactory precision and sensitivity. While the asymmetry index alone was not highly predictive, it significantly improved performance when combined with frequency-band features. Conclusions: This minimal EEG setup demonstrates reliable performance in distinguishing ocular states in non-clinical environments. The approach suggests potential applications in educational and field contexts, emphasizing the value of low-cost EEG solutions in cognitive monitoring.