A Computationally Efficient Hybrid Approach for Electrocardiogram-Based Arrhythmia Prediction

Area of Science:

• Cardiology
• Artificial Intelligence
• Biomedical Engineering

Background:

• Cardiovascular diseases, particularly arrhythmias, are a significant global cause of mortality.
• Early detection and diagnosis of arrhythmias are crucial for effective patient management.
• Automated systems are needed to assist in the timely identification of cardiac irregularities.

Purpose of the Study:

• To develop and validate a deep learning model for accurate arrhythmia detection using electrocardiogram (ECG) signals.
• To classify five distinct types of heartbeats: Normal (N), Left Bundle Branch Block (L), Right Bundle Branch Block (R), Atrial Premature Beat (A), and Premature Ventricular Contraction (V).
• To assess the model's performance against existing state-of-the-art methods.

Main Methods:

• Utilized Lead I ECG signals from multiple large-scale databases (MIT-BIH Arrhythmia, Supraventricular, INCART 12-lead, Sudden Cardiac Death Holter).
• Preprocessed data by segmenting into 180-sample windows, applying Min-Max normalization, and balancing classes using Synthetic Minority Over-sampling Technique (SMOTE).
• Employed a hybrid deep learning architecture combining 1D Convolutional Neural Networks (CNNs) for feature extraction and transformer layers for temporal pattern analysis, optimized with Adam and regularization techniques.

Main Results:

• Achieved an exceptional accuracy, precision, and F1-score of 99.99% across all five arrhythmia classes.
• Demonstrated superior performance compared to the TN4 model and other leading arrhythmia detection models.
• Highlighted the robustness of features extracted by the CNNs and the hybrid deep architecture.

Conclusions:

• The developed deep learning model demonstrates high efficacy and potential for real-time, scalable arrhythmia detection.
• This AI-driven approach can significantly contribute to advancing personalized digital healthcare solutions for cardiovascular conditions.
• The model's performance suggests a promising tool for early diagnosis and management of arrhythmias, potentially reducing mortality rates.