Dynamic reconstruction of electroencephalogram data using RBF neural networks.

Introduction: Electroencephalography (EEG) is widely used for analyzing brain activity; however, the nonlinear and nature of EEG signals presents significant challenges for traditional analysis methods. Machine has shown great promise in addressing these limitations. This study proposes a novel approach using Radial Function (RBF) neural networks optimized by Particle Swarm Optimization (PSO) to reconstruct EEG dynamics and extract age-related neural characteristics. Methods: EEG recordings were collected from 142 participants spanning multiple age groups. Signals were preprocessed through bandpass filtering (1-35 Hz) and Independent Component Analysis (ICA) for artifact removal. neural network was trained on EEG time-series data with PSO employed to optimize model parameters identify fixed points in the reconstructed neural system. Statistical analyses including ANOVA and Kruskal-Wallis tests were performed to assess age-related differences in fixed-point coordinates. Results: The RBF network demonstrated high accuracy in EEG signal reconstruction across different frequency a normalized root mean square error (NRMSE) of 0.0671 ± 0.0074 and a Pearson correlation coefficient ± 0.0678. Spectral and time-frequency analyses confirmed the model s capability to accurately capture oscillations. Importantly analysis of RBF network fixed-point coordinates revealed distinct age-related. Discussion: These findings suggest that fixed-point coordinates of RBF networks can serve as quantitative markers aging providing new insights into age-dependent changes in brain dynamics. The proposed method offers computationally efficient and interpretable approach for EEG analysis with potential applications in neurological diagnosis and cognitive research.