An efficient method for ectopic beats cancellation based on radial basis function

Area of Science:

• Cardiology
• Biomedical Engineering
• Artificial Intelligence

Background:

• Surface Electrocardiogram (ECG) is a primary noninvasive tool for diagnosing heart conditions.
• Ectopic beats, common in both healthy individuals and patients, introduce significant errors in ECG analysis.
• Accurate ECG interpretation requires effective methods for ectopic beat cancellation.

Purpose of the Study:

• To present a novel method for electrocardiogram ectopic beat cancellation using Radial Basis Function Neural Networks (RBFNN).
• To develop a customizable ECG beat classifier through a trainable neural network ensemble for improved ECG processing.
• To enhance individualized healthcare by improving the accuracy of ECG analysis.

Main Methods:

• Utilized the MIT-BIH arrhythmia database to obtain six types of heartbeats: Normal Beats (NB), Premature Ventricular Contractions (PVC), Left Bundle Branch Blocks (LBBB), Right Bundle Branch Blocks (RBBB), Paced Beats (PB), and Ectopic Beats (EB).
• Extracted four morphological features from each heartbeat after preprocessing.
• Applied a Radial Basis Function Neural Network (RBFNN) based ensemble approach for beat classification and ectopic beat cancellation.

Main Results:

• The RBFNN-based method achieved an average ectopic beat reduction (EBR) of 7.23 ± 2.18.
• Traditional methods achieved a best-case EBR of 4.05 ± 2.13.
• The RBFNN method demonstrated highly accurate ectopic beat reduction with minimal distortion of the QRST complex.

Conclusions:

• Radial Basis Function Neural Networks offer a superior approach for ectopic beat cancellation in ECG signals.
• The proposed method significantly enhances the accuracy of ECG analysis by reducing ectopic beats.
• This technique holds promise for improving diagnostic precision and enabling more personalized cardiac healthcare.