A Novel Wavelet Transform-Homogeneity Model for Sudden Cardiac Death Prediction Using ECG Signals

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Sudden cardiac death (SCD) remains a leading cause of mortality worldwide.
• Accurate and timely prediction of SCD is crucial for preventative interventions.
• Existing prediction methods often require extensive computational resources and complex feature extraction.

Purpose of the Study:

• To introduce a novel, computationally efficient methodology for predicting SCD using electrocardiogram (ECG) signals.
• To evaluate the accuracy and predictive time window of the proposed SCD prediction method.
• To demonstrate the advantages of the new approach over existing techniques.

Main Methods:

• Utilizing wavelet packet transform (WPT) for ECG signal processing.
• Employing the homogeneity index (HI), a nonlinear time series measurement, as a key predictive feature.
• Implementing an Enhanced Probabilistic Neural Network (EPNN) for classification.

Main Results:

• The proposed methodology achieved a high prediction accuracy of 95.8% for SCD events.
• The method successfully predicted SCD risk up to 20 minutes prior to onset.
• The approach demonstrated superior performance compared to previous methods using fewer features and direct ECG signal analysis.

Conclusions:

• The developed methodology offers a highly accurate and computationally efficient approach for real-time SCD prediction.
• Direct ECG signal analysis with HI and EPNN provides a significant advancement in early SCD risk detection.
• This method holds promise for improving patient outcomes by enabling timely preventative measures against SCD.