This is the authors' abstract. We don't add key points for this paper.
The study focuses on the processing and interpretation of multichannel electroencephalogram (EEG) signals containing high-frequency fluctuations and physiological artifacts of oculographic origin.The paper proposes a hybrid algorithm for removing oculographic artifacts from EEG signals based on a combination of the discrete wavelet transform and independent component analysis and evaluates its effectiveness. This paper proposes the use of a hybrid algorithm for removing oculographic artifacts from electroencephalograms based on a combination of the discrete wavelet transform and independent component analysis and evaluates its effectiveness. The research substantiates the combined use of the discrete wavelet transform and independent component analysis to remove oculographic artifacts from EEG signals. A hybrid algorithm for EEG signal cleaning was proposed and implemented in Python. Experiments were conducted to evaluate the effectiveness of the proposed method compared to baseline methods. It was determined that preliminary wavelet filtering using the Daubechies db4 mother wavelet removes small-scale high-frequency fluctuations from the signal, creating a smoothed basis for the FastICA algorithm. This improves the accuracy of separating the useful signal from artifacts. Artifact detection is performed automatically based on the calculation of the kurtosis value. The proposed hybrid algorithm demonstrated high denoising efficiency compared to the standalone use of wavelet filtering and independent component analysis, achieving a signal-to-noise ratio (SNR) improvement of 10.1 dB. The developed algorithm and software can be utilized in the design of modern medical diagnostic systems and neuromonitoring complexes, as well as in the creation of high-precision brain-computer interfaces.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
KIRILLOVA et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: