A single-trial toolbox for advanced sleep polysomnographic preprocessing.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Polysomnographic (PSG) studies monitor sleep using multiple biological signals.
• Effective signal preprocessing is crucial for accurate feature extraction and classification in sleep analysis.
• Existing methods may not sufficiently address noise and artifacts in multi-channel PSG recordings.

Purpose of the Study:

• To develop and evaluate a novel preprocessing toolbox for multi-channel PSG recordings.
• To enhance the efficiency of feature extraction and classification through improved signal quality.
• To assess the toolbox's performance in noise and artifact rejection.

Main Methods:

• The proposed toolbox integrates segmentation, filtering, denoising, whitening, and artifact removal.
• Multi-channel PSG data was preprocessed using single-trial and multi-stage approaches.
• Clinical experiments were conducted to evaluate the toolbox's efficacy.

Main Results:

• The preprocessing toolbox demonstrated superior performance in rejecting artifacts and noise.
• Quantitative and qualitative metrics confirmed the effectiveness of the proposed methods.
• Enhanced signal quality was achieved through single-trial and multi-stage preprocessing.

Conclusions:

• The developed PSG preprocessing toolbox significantly improves sleep signal analysis.
• The methods offer enhanced noise and artifact rejection, leading to more reliable sleep monitoring.
• This work contributes to more profitable feature extraction and classification in sleep studies.