Removal of Electrocardiogram Interference from Diaphragmatic Electromyogram Signals using Sliding Singular Spectrum...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Computational Biology

Background:

• Electromyogram (EMG) and Electrocardiogram (ECG) signals are crucial for medical diagnosis but often get mixed during acquisition.
• Common signal acquisition issues include body motion artifacts and electromagnetic interference, complicating signal analysis.
• Traditional filtering methods are insufficient for separating overlapping EMG and ECG spectra.

Purpose of the Study:

• To develop and evaluate a novel signal processing technique for isolating individual EMG and ECG signals from mixed recordings.
• To improve the accuracy and effectiveness of denoising biologically sourced single-channel signals.
• To offer a practical solution for physicians needing isolated physiological data.

Main Methods:

• Utilized a sliding Singular Spectrum Analysis (SSA) algorithm for time series decomposition.
• Applied statistical analysis to identify and separate distinct EMG and ECG signal components.
• Compared the performance of the sliding SSA approach against traditional block-based SSA methods.

Main Results:

• The sliding SSA technique demonstrated superior performance in separating real-world mixed EMG and ECG signals.
• The method successfully decomposed mixed signals into additive components identifiable as either EMG or ECG.
• The proposed approach proved more effective than conventional block-based SSA for single-channel data.

Conclusions:

• Sliding SSA offers a robust and effective method for separating mixed EMG and ECG signals.
• This non-parametric technique provides a straightforward implementation without requiring reference signals or prior training.
• The methodology holds potential for application to other signal types with appropriate statistical adjustments.