Sleep stage classification using single-channel EOG.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Sleep Medicine

Background:

• Accurate sleep stage classification is crucial for diagnosing sleep disorders and conducting sleep research.
• Existing methods often struggle with specific sleep stages, like S1, and rely on complex multi-channel data.
• Electrooculogram (EOG) signals offer a potential avenue for simplified yet effective sleep analysis.

Purpose of the Study:

• To develop and evaluate an automatic sleep stage classification system using single-channel Electrooculogram (EOG) signals.
• To investigate the efficacy of various statistical features extracted in the Discrete Wavelet Transform (DWT) domain for sleep stage discrimination.
• To compare the performance of machine learning classifiers, including RUSBoost, Random Forest, and SVM, for sleep stage scoring.

Main Methods:

• Single-channel EOG signals were analyzed using Discrete Wavelet Transform (DWT).
• Statistical features like Spectral Entropy, Moment-based Measures, Refined Composite Multiscale Dispersion Entropy (RCMDE), and Autoregressive (AR) Model Coefficients were extracted.
• Feature selection was performed using Neighborhood Component Analysis, followed by classification using RUSBoost, RF, and SVM on public sleep databases (Sleep-EDF, Sleep-EDFX, ISRUC-Sleep).

Main Results:

• The proposed EOG-based method demonstrated superior accuracy compared to state-of-the-art EOG techniques.
• Performance was comparable or superior to single-channel EEG-based methods.
• The RUSBoost classifier achieved significantly higher accuracy for S1 sleep stage classification than existing EOG and EEG methods.

Conclusions:

• Automatic sleep stage classification using single-channel EOG signals and DWT-based features is feasible and effective.
• The proposed method offers a promising alternative to traditional multi-channel approaches, particularly for improving S1 stage detection.
• This technique has the potential to enhance the diagnosis and study of sleep disorders.