Synchroextracting wavelet transform-based enhancement of weak EEG signals for improved signal detection

This research investigates the utilization of advanced time-frequency analysis (TFA) methodologies for processing intricate EEG signals. Recognizing the constraints of traditional techniques like Wavelet Transform (WT) and Synchroextracting Transform (SET) in handling non-stationary EEG signals—particularly their limitations in achieving optimal time-frequency clarity and energy distribution—a novel approach termed Second-Order Synchroextracting Wavelet Transform (SEWT2) is presented. SEWT2 integrates sophisticated higher-order frequency estimation and reallocation strategies, markedly improving its ability to capture dynamic EEG signal features and enhance the quality of time-frequency representation. To further address challenges posed by high-noise environments in EEG data, an enhanced variant, Improved SEWT2 (ISEWT2), is developed, specifically tailored for robust signal analysis. Rigorous evaluations through numerical simulations on amplitude-modulated and frequency-modulated signals demonstrate that ISEWT2 delivers superior performance, paving the way for innovative applications in neurological diagnostics and interventions.