Beat-to-beat T-wave alternans detection using the Ensemble Empirical Mode Decomposition method

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• T-wave alternans (TWA) signifies a beat-to-beat variation in ECG repolarization.
• Detecting TWA is crucial for assessing cardiac conditions.

Purpose of the Study:

• To evaluate the efficacy of the Ensemble Empirical Mode Decomposition (EEMD) method for beat-to-beat TWA detection.
• To assess TWA in patients with reduced left ventricular ejection fraction.

Main Methods:

• ECG data from 69 healthy subjects and 39 cardiac patients were analyzed.
• Beat-to-beat T-waves were extracted after QT interval determination.
• The EEMD method decomposed T-waves into intrinsic mode functions (IMFs) for feature extraction.
• Four classifiers were used to detect TWA.

Main Results:

• Simulations showed TWA detection accuracy and sensitivity exceeding 80%.
• The Ensemble classifier achieved 74% accuracy and 100% positive predictive value on real Holter ECG data.
• Global classifiers outperformed subject-based classifiers in simulations.

Conclusions:

• The Ensemble EMD method combined with an Ensemble classifier is effective for beat-to-beat TWA detection.
• This approach offers a promising tool for non-invasive cardiac risk stratification.