ANALYSIS AND CHARACTERIZATION OF VISUAL EVOKED POTENTIAL SIGNALS USING MODE DECOMPOSITIONS

This thesis proposes a framework for the denoising and feature extraction of Visual Evoked Potential (VEP) signals to improve the characterization of neurological responses. The research addresses the non-stationary and low-signal-to-noise ratio (SNR) nature of EEG data by implementing advanced Mode Decomposition techniques. These methods were utilized to decompose raw VEP signals into intrinsic components, followed by the application of Multiscale Fluctuation-Based Dispersion Entropy to quantify signal complexity and identify underlying patterns. The study demonstrates that the combination of mode decomposition and entropy-based measures provides a more sensitive set of biomarkers for identifying neurophysiological irregularities compared to traditional Fourier-based methods. The results contribute to the development of more accurate automated tools for clinical neurological diagnosis.