Quantification of ECG late potentials by wavelet transformation

Area of Science:

• Cardiology
• Biomedical Signal Processing
• Medical Diagnostics

Background:

• Late potentials in electrocardiogram (ECG) recordings are crucial indicators of cardiac electrical instability.
• Sustained ventricular tachycardia (VT) is a life-threatening arrhythmia associated with abnormal ventricular electrical activity.
• Traditional signal processing methods may have limitations in analyzing non-stationary ECG signals.

Purpose of the Study:

• To investigate the efficacy of wavelet transformation for analyzing late potentials in ECG recordings.
• To compare the performance of wavelet transformation with Fast Fourier Transform (FFT) spectrograms for signal analysis.
• To develop a quantitative method for discriminating between patients with VT and healthy subjects using ECG data.

Main Methods:

• ECG recordings from patients with sustained VT and healthy controls were preprocessed.
• Wavelet transformation was applied to analyze the non-stationary ECG signals.
• Energy distribution plots and scalograms were generated from wavelet transformed signals.
• Classification was performed based on energy in specific time-frequency regions.

Main Results:

• Wavelet transformation demonstrated superior detection accuracy and frequency resolution compared to FFT spectrograms for artificial test signals.
• Quantitative discrimination between VT patients and healthy subjects was achieved using wavelet-transformed ECG signals.
• The energy within the 100-300 Hz frequency band during the terminal QRS complex segment provided the best classification results.

Conclusions:

• Wavelet transformation is a powerful tool for analyzing non-stationary ECG signals, particularly for identifying late potentials.
• This method offers improved accuracy in distinguishing between pathological and normal cardiac electrical activity.
• The findings suggest a potential for wavelet-based ECG analysis in the clinical diagnosis of ventricular tachycardia.