Related Experiment Videos
Atrial activity enhancement by Wiener filtering using an artificial neural network
C Vásquez1, A Hernández, F Mora
1Grupo de Bioingeniería y Biofísica Aplicada, Universidad Simón Bolívar, Caracas, Venezuela.
IEEE Transactions on Bio-Medical Engineering
|August 14, 2001
Summary
A new method using a dynamic time delay neural network (TDNN) effectively cancels ventricular activity in ECG signals. This technique improves P-wave and atrial fibrillation detection, even with noisy data.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Computational Neuroscience
Background:
- Ventricular activity in electrocardiogram (ECG) signals can obscure important atrial activity, hindering the detection of arrhythmias like atrial fibrillation.
- Existing methods for ventricular activity cancellation often rely on QRS complex detection, which can be unreliable in noisy conditions or with varying QRS morphologies.
Purpose of the Study:
- To introduce and evaluate a novel technique for canceling ventricular activity in ECG signals.
- To assess the performance of the proposed method for P-wave detection and atrial fibrillation analysis.
- To compare the novel approach against existing adaptive cancellation schemes.
Main Methods:
- A dynamic time delay neural network (TDNN) was employed to estimate a time-varying, nonlinear transfer function between two ECG leads.
- An Elman TDNN configuration with nine input samples, 20 hidden neurons (sigmoidal tangential activation), and one linear output neuron yielded optimal results.
- The method was designed to operate without a preceding QRS detection stage.
Main Results:
- The novel TDNN-based method demonstrated superior performance compared to a previously published adaptive cancellation scheme.
- Quantitative evaluation using the MIT-BIH arrhythmia database confirmed the effectiveness of the proposed approach.
- The technique proved robust against noisy ECG episodes and variations in QRS complex morphology.
Conclusions:
- The proposed dynamic time delay neural network technique offers an effective and robust solution for ventricular activity cancellation in ECG.
- This method enhances the detection of P-waves and atrial fibrillation, outperforming existing adaptive cancellation techniques.
- The TDNN approach's independence from QRS detection and its resilience to noise and morphological variations make it a valuable tool in cardiac signal analysis.