Variational Mode Extraction: A New Efficient Method to Derive Respiratory Signals from ECG

Area of Science:

• Biomedical Engineering
• Signal Processing
• Physiological Monitoring

Background:

• ECG-derived respiration (EDR) is a non-invasive monitoring method.
• Existing EDR extraction techniques like EMD are sensitive to noise.
• Variational Mode Decomposition (VMD) offers noise robustness but incurs high computational costs.

Purpose of the Study:

• To introduce Variational Mode Extraction (VME) for EDR signal extraction.
• To address the limitations of existing methods in terms of noise sensitivity and computational efficiency.
• To propose VME as a more efficient and accurate approach for EDR monitoring.

Main Methods:

• Developed Variational Mode Extraction (VME) based on VMD principles.
• Introduced a new criterion focusing on minimizing residual energy at the mode's center frequency.
• Applied VME to extract EDR signals, considering EDR as a mode with near-zero center frequency.

Main Results:

• VME demonstrated superior accuracy in extracting EDR signals compared to VMD, EMD, and bandpass filtering.
• The proposed VME method exhibited significantly lower computational cost.
• VME achieved a higher convergence rate in EDR signal extraction.

Conclusions:

• VME is a highly effective and computationally efficient method for extracting EDR signals.
• The new method provides a valuable alternative for accurate respiration monitoring using ECG.
• VME overcomes the noise sensitivity and computational burden of previous techniques.