Convolutional neural network for classification of eight types of arrhythmia using 2D time\u2013frequency feature map from standard 12-lead electrocardiogram

Da Un Jeong,Ki Moo Lim

doi:10.1038/s41598-021-99975-6

Abstract

Electrocardiograms (ECGs) are widely used for diagnosing cardiac arrhythmia based on the deformation of signal shapes due to changes in various heart diseases. However, these abnormal signs may not be observed in some 12 ECG channels, depending on the location, the heart shape, and the type of cardiac arrhythmia. Therefore, it is necessary to closely and comprehensively observe ECG records acquired from 12 channel electrodes to diagnose cardiac arrhythmias accurately. In this study, we proposed a clustering algorithm that can classify persistent cardiac arrhythmia as well as episodic cardiac arrhythmias using the standard 12-lead ECG records and the 2D CNN model using the time–frequency feature maps to classify the eight types of arrhythmias and normal sinus rhythm. The standard 12-lead ECG records were provided by China Physiological Signal Challenge 2018 and consisted of 6877 patients. The proposed algorithm showed high performance in classifying persistent cardiac arrhythmias; however, its accuracy was somewhat low in classifying episodic arrhythmias. If our proposed model is trained and verified using more clinical data, we believe it can be used as an auxiliary device for diagnosing cardiac arrhythmias.

Highlights

In this study, cardiac arrhythmias in which abnormal signs are continuously observed in the ECG record were defined as ’persistent cardiac arrhythmias’
The highest F1 score was observed for left bundle branch block (LBBB) prediction at 0.89, and the classified F1 scores for atrial fibrillation (AF), right bundle branch block (RBBB), and I-AVB were 0.86, 0.85, and 0.80, respectively
The classification performance of episodic cardiac arrhythmias such as Premature atrial contraction (PAC), premature ventricular contraction (PVC), and segment elevation (STE) was relatively low compared to persistent cardiac arrhythmias

Summary

Introduction

Cardiac arrhythmias in which abnormal signs are continuously observed in the ECG record were defined as ’persistent cardiac arrhythmias’ (the meaning of ’persistent’ in ’persistent cardiac arrhythmias’ is different from that in ’persistent atrial fibrillation’). Unlike persistent cardiac arrhythmias, cardiac arrhythmias that occur intermittently in normal signals are defined as "episodic cardiac arrhythmias." ST-segment elevation (STE) and ST-segment depression (STD), which are mainly caused by myocardial ischemia or infarction, are observed intermittently in the normal sinus rhythm (NSR). They are diagnosed when the height difference between the S-peak and the PR segment, which denotes the slope of the ST-segment of the ECG, is greater than ± 0.05 mV (one small scale of ECG is 1 mm, which means 0.04 s on the x-axis and 0.1 mV on the y-axis). Mostayed et al proposed a bidirectional LSTM network classifier to detect nine arrhythmias, including NSR, and achieved a final F1 score of 74.1515

Methods

Results

Conclusion