Detecting atrial fibrillation by deep convolutional neural networks

• Published in: Computers in biology and medicine Vol. 93; pp. 84 - 92
• Main Authors: Xia, Yong, Wulan, Naren, Wang, Kuanquan, Zhang, Henggui
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01-02-2018; Elsevier Limited
• Subjects: Algorithms; Arrhythmia; Artificial neural networks; Atrial fibrillation; Atrial Fibrillation - diagnosis; Atrial Fibrillation - physiopathology; Cardiac arrhythmia; Classification; Deep convolutional neural networks; Echocardiography; EKG; Electrocardiography; Electronic Data Processing; Entropy; Female; Fibrillation; Fourier transforms; Health risks; Heart diseases; Humans; Machine learning; Male; Morbidity; Neural networks; Neural Networks (Computer); Risk factors; Segments; Sensitivity; Short-term Fourier transform; Stationary wavelet transform; Two dimensional analysis; Voice recognition; Wavelet analysis; Wavelet transforms; United States > US; Deep convolutional neural networks; Stationary wavelet transform; Atrial fibrillation; Short-term Fourier transform
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2017.12.007

Atrial fibrillation (AF) is the most common cardiac arrhythmia. The incidence of AF increases with age, causing high risks of stroke and increased morbidity and mortality. Efficient and accurate diagnosis of AF based on the ECG is valuable in clinical settings and remains challenging. In this paper, we proposed a novel method with high reliability and accuracy for AF detection via deep learning.
The short-term Fourier transform (STFT) and stationary wavelet transform (SWT) were used to analyze ECG segments to obtain two-dimensional (2-D) matrix input suitable for deep convolutional neural networks. Then, two different deep convolutional neural network models corresponding to STFT output and SWT output were developed. Our new method did not require detection of P or R peaks, nor feature designs for classification, in contrast to existing algorithms. Finally, the performances of the two models were evaluated and compared with those of existing algorithms.
Our proposed method demonstrated favorable performances on ECG segments as short as 5 s. The deep convolutional neural network using input generated by STFT, presented a sensitivity of 98.34%, specificity of 98.24% and accuracy of 98.29%. For the deep convolutional neural network using input generated by SWT, a sensitivity of 98.79%, specificity of 97.87% and accuracy of 98.63% was achieved.
The proposed method using deep convolutional neural networks shows high sensitivity, specificity and accuracy, and, therefore, is a valuable tool for AF detection.
•A new application of deep convolutional neural network to the detection of atrial fibrillation was explored.•The short-term Fourier transform and stationary wavelet transform were used separately for signal conversion.•Good classification performance was obtained for ECG segments as short as 5 s.•Performance was not dependent on the detection of P or R peaks and the extraction of hand-crafted features.