Beat-to-beat noise removal in non-invasive His-bundle electrocardiogram

Area of Science:

• Biomedical Engineering
• Cardiovascular Electrophysiology
• Signal Processing

Background:

• His-bundle electrocardiogram (ECG) micropotentials are crucial for diagnosing cardiac conduction abnormalities.
• Traditional signal averaging limits beat-by-beat analysis due to noise.
• Existing methods struggle with real-time removal of power-line and electromyographic (EMG) interference.

Purpose of the Study:

• To develop a method for surface beat-to-beat His-bundle potential recovery without signal shape distortion.
• To address the limitations of serial-signal averaging for His-bundle ECG analysis.
• To enable accurate detection of low-amplitude His potentials in noisy ECG signals.

Main Methods:

• A novel approach combining consecutive power-line interference subtraction and wavelet-domain EMG noise suppression.
• Modification of wavelet-domain filtering to meet stringent criteria for low-amplitude His potential detection.
• Beat-to-beat wavelet-domain Wiener filtering utilizing a pilot signal estimate from approximately 20 averaged heartbeats.

Main Results:

• Achieved over 4.5 dB improvement in signal-to-noise ratio (SNR) along the P-Q segment.
• Reduced mean absolute error by over 20% along the P-Q segment.
• Demonstrated applicability for ECG signals with moderate EMG noise and initial SNR ≥ 15 dB.

Conclusions:

• The developed method enables accurate beat-to-beat His-bundle potential recovery from surface ECG.
• This technique overcomes limitations of traditional signal averaging for noise removal.
• It offers a significant advancement for diagnosing cardiac conditions using high-fidelity His-bundle ECG signals.