Automated diagnosis of arrhythmia using combination of CNN and LSTM techniques with variable length heart beats

Area of Science:

• Cardiology
• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Arrhythmia, a disorder of cardiac conduction, presents as irregular heartbeats.
• Electrocardiographic (ECG) signals reflect abnormalities in the heart's conduction system.
• Visual assessment of ECGs for arrhythmia diagnosis is often challenging and time-consuming due to low signal amplitudes.

Purpose of the Study:

• To develop and evaluate an automated system for the accurate and efficient diagnosis of common arrhythmias using ECG signals.
• To address the limitations of manual ECG interpretation by implementing a robust diagnostic tool.

Main Methods:

• A hybrid model combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) was developed for ECG analysis.
• The system was trained and validated on variable-length ECG segments from the MIT-BIT arrhythmia database.
• A ten-fold cross-validation strategy was employed to assess classification performance.

Main Results:

• The proposed CNN-LSTM system achieved high classification performance on variable-length ECG data.
• Achieved an overall accuracy of 98.10%, sensitivity of 97.50%, and specificity of 98.70%.
• Demonstrated effectiveness in diagnosing normal sinus rhythm, left bundle branch block (LBBB), right bundle branch block (RBBB), atrial premature beats (APB), and premature ventricular contraction (PVC).

Conclusions:

• The automated CNN-LSTM system offers a promising solution for expediting arrhythmia diagnosis in clinical settings.
• The model's high accuracy and ability to handle variable-length ECG data can aid clinicians in routine screening.
• This approach has the potential to improve diagnostic accuracy and efficiency for common cardiac arrhythmias.