Exploring Spectral and Nonlinear Characteristics of Neural Dynamics during Cognitive Loading

Understanding the neural dynamics of cognitive tasks is fundamental to neuroscience. This study investigated neural responses during cognitive loading in arithmetic tasks using both linear and nonlinear characterizations of electroencephalography (EEG) signals. We extracted spectral features (alpha, beta, delta, theta, and gamma band powers) and nonlinear features (Higuchi’s Fractal Dimension, sample entropy, and spectral entropy) to analyze neural dynamics under cognitive load. The EEG signals were initially denoised using Discrete Wavelet Transform, followed by feature extraction and channel-wise analysis. The results indicate that cognitive loading significantly shifts power across all frequency bands. Nonlinear features provide deeper insights into the adaptability of the brain to cognitive demands. Importantly, changes in neural dynamics are not uniform across brain regions, with frontal and temporal areas being more involved. The features studied effectively characterized neural changes during cognitive loading, and hold potential for future research in various contexts, including educational settings.