A Wearable Electrocardiogram Telemonitoring System for Atrial Fibrillation Detection
<p>Illustration of the wearable ECG telemonitoring system.</p> "> Figure 2
<p>Flow chart of the proposed ECG monitoring System.</p> "> Figure 3
<p>Block diagram of the hardware of the ECG patch.</p> "> Figure 4
<p>(<b>a</b>) The hardware circuit board with lithium-ion battery. (<b>b</b>) Front view of the ECG patch device with disposable ECG electrodes. (<b>c</b>) The three electrodes of the patch and four charge contacts of the patch and the charging base. (RLD = right leg drive). (<b>d</b>) The ECG patch was charging on the charging base.</p> "> Figure 5
<p>The main user interface of the Android APP.</p> "> Figure 6
<p>Flow chart of the ECG classification method proposed in this work.</p> "> Figure 7
<p>The main web page of the web application.</p> "> Figure 8
<p>Flow chart of the web application of ECG diagnosis.</p> "> Figure 9
<p>Illustration of the importance of the 31 features. (The top 17 features were selected to train the final classifier).</p> "> Figure 10
<p>MCC scores according to different number of top-importance features. MCC = Matthews correlation coefficient.</p> "> Figure 11
<p>The ECG monitoring system in operation. (<b>a</b>) The ECG patch was connected to the FLUKE MPS450. (<b>b</b>,<b>c</b>) The Android APP displayed the normal (<b>b</b>) and abnormal (<b>c</b>) ECG waveforms in real time. (<b>d</b>) The doctor’s web browser displayed the 30 s ECG waveforms. (<b>e</b>) The historical data module of the APP showed the doctor’s diagnosis.</p> "> Figure 11 Cont.
<p>The ECG monitoring system in operation. (<b>a</b>) The ECG patch was connected to the FLUKE MPS450. (<b>b</b>,<b>c</b>) The Android APP displayed the normal (<b>b</b>) and abnormal (<b>c</b>) ECG waveforms in real time. (<b>d</b>) The doctor’s web browser displayed the 30 s ECG waveforms. (<b>e</b>) The historical data module of the APP showed the doctor’s diagnosis.</p> ">
Abstract
:1. Introduction
2. Materials and Methods
2.1. Hardware Design of ECG Patches
2.2. Android APP Software Design
2.3. ECG Classification Algorithm
2.3.1. Feature Extraction
2.3.2. ECG Classification
2.3.3. ECG Dataset
2.4. Cloud Software Design
3. Results
3.1. Algorithm Perforcemence
3.2. System Verification
4. Discussion
- (1)
- At present, the system was only tested with open-access ECG databases. Clinical studies may be conducted in future work.
- (2)
- An ECG quality assessment method may be developed and implemented in the smartphone. The quality assessment method could automatically estimate the signal quality in real time, and the user would be alerted when the signal quality is too low.
- (3)
- Currently, when a 30 s ECG segment is detected as AF by the proposed algorithm, all the ECG signals will be transferred to the doctor for diagnosis. However, the algorithm may have false-negative detections of AF. Some improvements may be considered in future work, including (a) improving the performance of the AF detector, especially reducing false-negative detections; (b) an index for AF detection can be provided for doctors so that the doctors can adjust the false negative and false positive rates by selecting different values of the index; (c) statistical analysis can be conducted on the clinical data collected by the proposed system for estimations of false positive and false negative rates, with causes for misdetection analyzed.
5. Conclusions
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
References
- Chugh, S.S.; Havmoeller, R.; Narayanan, K.; Singh, D.; Rienstra, M.; Benjamin, E.J.; Gillum, R.F.; Kim, Y.H.; McAnulty, J.H.; Zheng, Z.J.; et al. Worldwide epidemiology of atrial fibrillation a global burden of disease 2010 study. Circulation 2014, 129, 837–847. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Thrall, G.; Lane, D.; Carroll, D.; Lip, G.Y.H. Quality of life in patients with atrial fibrillation: A systematic review. Am. J. Med. 2006, 119, 448.e1–448.e19. [Google Scholar] [CrossRef] [PubMed]
- Reinier, K.; Marijon, E.; Uy-Evanado, A.; Teodorescu, C.; Narayanan, K.; Chugh, H.; Gunson, K.; Jui, J.; Chugh, S.S. The association between atrial fibrillation and sudden cardiac death: The relevance of heart failure. JACC Heart Fail. 2014, 2, 221–227. [Google Scholar] [CrossRef] [PubMed]
- Kirchhof, P.; Benussi, S.; Kotecha, D.; Ahlsson, A.; Atar, D.; Casadei, B.; Castella, M.; Diener, H.C.; Heidbuchel, H.; Hendriks, J.; et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur. Heart J. 2016, 37, 2893–2962. [Google Scholar] [CrossRef] [Green Version]
- Lau, J.K.; Lowres, N.; Neubeck, L.; Brieger, D.B.; Sy, R.W.; Galloway, C.D.; Albert, D.E.; Freedman, S.B. iPhone ECG application for community screening to detect silent atrial fibrillation: A novel technology to prevent stroke. Int. J. Cardiol. 2013, 165, 193–194. [Google Scholar] [CrossRef]
- Halcox, J.P.J.; Wareham, K.; Cardew, A.; Gilmore, M.; Barry, J.P.; Phillips, C.; Gravenor, M.B. Assessment of remote heart rhythm sampling using the AliveCor heart monitor to screen for atrial fibrillation: The REHEARSE-AF study. Circulation 2017, 136, 1784–1794. [Google Scholar] [CrossRef]
- Turakhia, M.P.; Ullal, A.J.; Hoang, D.D.; Than, C.T.; Miller, J.D.; Friday, K.J.; Perez, M.V.; Freeman, J.V.; Wang, P.J.; Heidenreich, P.A. Feasibility of extended ambulatory electrocardiogram monitoring to identify silent atrial fibrillation in high-risk patients: The screening study for undiagnosed atrial fibrillation (STUDY-AF). Clin. Cardiol. 2015, 38, 285–292. [Google Scholar] [CrossRef]
- Steinhubl, S.R.; Waalen, J.; Edwards, A.M.; Ariniello, L.M.; Mehta, R.R.; Ebner, G.S.; Carter, C.; Baca-Motes, K.; Felicione, E.; Sarich, T.; et al. Effect of a home-based wearable continuous ECG monitoring patch on detection of undiagnosed atrial fibrillation: The mSToPS randomized clinical trial. J. Am. Med. Assoc. 2018, 320, 146–155. [Google Scholar] [CrossRef] [Green Version]
- Fukuma, N.; Hasumi, E.; Fujiu, K.; Waki, K.; Toyooka, T.; Komuro, I.; Ohe, K. Feasibility of a T-shirt-type wearable electrocardiography monitor for detection of covert atrial fibrillation in young healthy adults. Sci. Rep. 2019, 9, 11768. [Google Scholar] [CrossRef] [Green Version]
- Yang, Z.; Zhou, Q.H.; Lei, L.; Zheng, K.; Xiang, W. An IoT-cloud based wearable ECG monitoring system for smart healthcare. J. Med. Syst. 2016, 40, 286. [Google Scholar] [CrossRef]
- Hussein, A.F.; Kumar, N.A.; Burbano-Fernandez, M.; Ramirez-Gonzalez, G.; Abdulhay, E.; De Albuquerque, V.H.C. An automated remote cloud-based heart rate variability monitoring system. IEEE Access 2018, 6, 77055–77064. [Google Scholar] [CrossRef]
- Venkatesan, C.; Karthigaikumar, P.; Satheeskumaran, S. Mobile cloud computing for ECG telemonitoring and real-time coronary heart disease risk detection. Biomed. Signal Process. Control 2018, 44, 138–145. [Google Scholar] [CrossRef]
- Slocum, J.; Sahakian, A.; Swiryn, S. Diagnosis of atrial fibrillation from surface electrocardiograms based on computer-detected atrial activity. J. Electrocardiol. 1992, 25, 1–8. [Google Scholar] [CrossRef]
- Portet, F. P wave detector with PP rhythm tracking: Evaluation in different arrhythmia contexts. Physiol. Meas. 2008, 29, 141–155. [Google Scholar] [CrossRef]
- Huang, C.; Ye, S.M.; Chen, H.; Li, D.L.; He, F.T.; Tu, Y.W. A novel method for detection of the transition between atrial fibrillation and sinus rhythm. IEEE Trans. Biomed. Eng. 2011, 58, 1113–1119. [Google Scholar] [CrossRef]
- Lee, J.; Nam, Y.; McManus, D.D.; Chon, K.H. Time-varying coherence function for atrial fibrillation detection. IEEE Trans. Biomed. Eng. 2013, 60, 2783–2793. [Google Scholar]
- Sarkar, S.; Ritscher, D.; Mehra, R. A detector for a chronic implantable atrial tachyarrhythmia monitor. IEEE Trans. Biomed. Eng. 2008, 55, 1219–1224. [Google Scholar] [CrossRef]
- Tateno, K.; Glass, L. Automatic detection of atrial fibrillation using the coefficient of variation and density histograms of RR and ΔRR intervals. Med. Biol. Eng. Comput. 2001, 39, 664–671. [Google Scholar] [CrossRef]
- Zhou, X.L.; Ding, H.X.; Ung, B.; Pickwell-MacPherson, E.; Zhang, Y.T. Automatic online detection of atrial fibrillation based on symbolic dynamics and Shannon entropy. Biomed. Eng. Online 2014, 13, 18. [Google Scholar] [CrossRef] [Green Version]
- Clifford, G.D.; Liu, C.; Moody, B.; Lehman, L.H.; Silva, I.; Li, Q.; Johnson, A.E.; Mark, R.G. AF classification from a short single lead ECG recording: The PhysioNet/computing in cardiology challenge 2017. In Proceedings of the 2017 Computing in Cardiology, Rennes, France, 24–27 September 2017; pp. 1–4. [Google Scholar]
- Datta, S.; Puri, C.; Mukherjee, A.; Banerjee, R.; Choudhury, A.D.; Singh, R.; Ukil, A.; Bandyopadhyay, S.; Pal, A.; Khandelwal, S. Identifying normal, AF and other abnormal ECG rhythms using a cascaded binary classifier. In Proceedings of the 2017 Computing in Cardiology, Rennes, France, 24–27 September 2017; pp. 1–4. [Google Scholar]
- Hong, S.; Wu, M.; Zhou, Y.; Wang, Q.; Shang, J.; Li, H.; Xie, J. ENCASE: An ENsemble ClASsifiEr for ECG classification using expert features and deep neural networks. In Proceedings of the 2017 Computing in Cardiology, Rennes, France, 24–27 September 2017; pp. 1–4. [Google Scholar]
- Teijeiro, T.; García, C.A.; Castro, D.; Félix, P. Arrhythmia classification from the abductive interpretation of short single-lead ECG records. In Proceedings of the 2017 Computing in Cardiology, Rennes, France, 24–27 September 2017; pp. 1–4. [Google Scholar]
- Xiong, Z.; Stiles, M.K.; Zhao, J. Robust ECG signal classification for detection of atrial fibrillation using a novel neural network. In Proceedings of the 2017 Computing in Cardiology, Rennes, France, 24–27 September 2017; pp. 1–4. [Google Scholar]
- Zabihi, M.; Rad, A.B.; Katsaggelos, A.K.; Kiranyaz, S.; Narkilahti, S.; Gabbouj, M. Detection of atrial fibrillation in ECG hand-held devices using a random forest classifier. In Proceedings of the 2017 Computing in Cardiology, Rennes, France, 24–27 September 2017; pp. 1–4. [Google Scholar]
- Zihlmann, M.; Perekrestenko, D.; Tschannen, M. Convolutional recurrent neural networks for electrocardiogram classification. In Proceedings of the 2017 Computing in Cardiology, Rennes, France, 24–27 September 2017; pp. 1–4. [Google Scholar]
- Shao, M.; Bin, G.; Wu, S.; Bin, G.; Huang, J.; Zhou, Z. Detection of atrial fibrillation from ECG recordings using decision tree ensemble with multi-level features. Physiol. Meas. 2018, 39, 094008. [Google Scholar] [CrossRef]
- Yuan, L.; Yuan, Y.; Zhou, Z.; Bai, Y.; Wu, S. A fetal ECG monitoring system based on the Android smartphone. Sensors 2019, 19, 446. [Google Scholar] [CrossRef] [Green Version]
- Dorogush, A.V.; Ershov, V.; Gulin, A. CatBoost: Gradient boosting with categorical features support. arXiv 2018, arXiv:1810.11363. [Google Scholar]
- Dash, S.; Chon, K.H.; Lu, S.; Raeder, E.A. Automatic real time detection of atrial fibrillation. Ann. Biomed. Eng. 2009, 37, 1701–1709. [Google Scholar] [CrossRef]
- Goldberger, A.L.; Amaral, L.A.N.; Glass, L.; Hausdorff, J.M.; Ivanov, P.C.; Mark, R.G.; Mietus, J.E.; Moody, G.B.; Peng, C.K.; Stanley, H.E. PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation 2000, 101, e215–e220. [Google Scholar] [CrossRef] [Green Version]
- Hamilton, P. Open source ECG analysis. In Proceedings of the 2002 Computers in Cardiology, Memphis, TN, USA, 22–25 September 2002; pp. 101–104. [Google Scholar]
- Tsipouras, M.G.; Fotiadis, D.I.; Sideris, D. An arrhythmia classification system based on the RR-interval signal. Artif. Intell. Med. 2005, 33, 237–250. [Google Scholar] [CrossRef]
- Boughorbel, S.; Jarray, F.; El-Anbari, M. Optimal classifier for imbalanced data using Matthews Correlation Coefficient metric. PLoS ONE 2017, 12, e0177678. [Google Scholar] [CrossRef]
- Xia, H.N.; Garcia, G.A.; Bains, J.; Wortham, D.C.; Zhao, X.P. Matrix of regularity for improving the quality of ECGs. Physiol. Meas. 2012, 33, 1535–1548. [Google Scholar] [CrossRef]
- Jiang, K.; Huang, C.; Ye, S.M.; Chen, H. High accuracy in automatic detection of atrial fibrillation for Holter monitoring. J. Zhejiang Univ. Sci. B 2012, 13, 751–756. [Google Scholar] [CrossRef]
- Asgari, S.; Mehrnia, A.; Moussavi, M. Automatic detection of atrial fibrillation using stationary wavelet transform and support vector machine. Comput. Biol. Med. 2015, 60, 132–142. [Google Scholar] [CrossRef]
Module | Description |
---|---|
User module | Register user information to the cloud and log in the APP with registered mobile number and password |
BLE module | Connect the ECG patch via BLE and receive ECG data continuously |
Display and storage module | Display ECG waveforms and store the raw ECG data to the internal storage in real time |
Historical data module | Display historical ECG waveforms and the related doctor’s diagnosis |
Cloud module | Transmit 30 s ECG data to the cloud and receive the doctor’s diagnosis |
Group/Feature | Description | |
---|---|---|
AF features | ||
AFEvidence [17] | AFEvidence feature was calculated based on the two-dimensional histogram built on the Lorenz distribution of dRR intervals from the whole ECG recording. | |
Shannon entropy [30] | Shannon entropy feature was computed on the histogram of the dRR intervals. | |
Kolmogorov–Smirnov test [15] | Kolmogorov–Smirnov test feature was obtained by evaluating the difference between the distribution of the ECG recording and the reference distribution for AF. | |
Morphology features | Ten kinds of Morphology features were extracted by using two open source libraries, the ECGPUWAVE [31] and the OSEA [32], respectively, including (1) QRS duration, (2) PR interval, (3) QT interval, (4) QS interval, (5) ST amplitude, (6) P amplitude, (7) Q amplitude, (8) R amplitude, (9) S amplitude, and (10) T amplitude. | |
RR interval features | ||
Median RR interval | Median RR interval feature was the median value of all RR intervals extracted from the entire ECG recording. | |
Index of arrhythmia [33] | Index of arrhythmia feature was the number of abnormal beats in an ECG recording. The abnormal beats were determined by four knowledge-based conditions. These conditions were judged according to three continuous RR intervals and their mean value. | |
Features for Noisy class | ||
Similarity index of QRS | Similarity index of QRS was the mean value of correlation coefficients calculated between every two QRS waveforms from the whole ECG recording. | |
Signal quality index | Signal quality index represented the ratio of high signal quality beats in an ECG record. A high signal quality beat was decided by evaluating the amplitudes of the isoelectric level. | |
Q-R smoothness index(QRsi) | QRsi feature represented the smoothness of the segment from the average beat. The segment was the increased QRS amplitude lasting from QRS onset to R-peak. QRsi value was defined as the peak numbers computed on the difference values of samples. |
Parameter | Value |
---|---|
learning_rate | 0.1 |
Iterations | 276 |
early_stopping_rounds | 20 |
depth | 8 |
l2_leaf_reg | 3 |
bagging_temperature | 0.7 |
random_strength | 0.2 |
leaf_estimation_method | “Newton” |
random_seed | (Random integer) |
loss_function | MultiClass |
eval_metric | Matthews correlation coefficient (MCC) [34] |
Database | Training Set | Test Set | Annotation | Total | |||
---|---|---|---|---|---|---|---|
Normal | AF | Other | Noisy | ||||
AFDB-2017 | 6822 | 1706 | 5076 | 758 | 2415 | 279 | 8528 |
MITBIH-AFDB | 22,252 | 5564 | 16,554 | 11,066 | 196 | 0 | 27,816 |
Total | 29,074 | 7270 | 21,630 | 11,824 | 2611 | 279 | 36,344 |
Used by | URL of Web API 1 | Description |
---|---|---|
Android APP | /user/register | Receive user information including mobile number and password, and store them in a database. |
/user/login | Receive mobile number and password from Android APP, determine the correctness and return the authentication result. | |
/user/ecgsegment | Receive 30 s ECG data, analyze it using the ECG classification algorithm, save the data and the analysis results in a database, and send a notification to the web application. | |
/user/ecgdiagnosis | Provide the ECG diagnosis requested by the Android APP. | |
Web application of ECG diagnosis | /doc/login | Receive the doctor’s username and password from the web browser, and determine the validity. |
/doc/userinfo | Transmit the requested user’s personal information to the web browser. | |
/doc/ecgsegment | Transmit the requested 30 s ECG data to the web browser. | |
/doc/ecganalysis | Receive the doctor’s diagnosis, save it in a database and send a notification to the Android APP. |
Classifier | Testing Set | No. of Cases | Acc | F1n | F1a | F1o | F1 |
---|---|---|---|---|---|---|---|
Cross-validation | Training set | 29,074 | 0.96 | 0.98 | 0.98 | 0.79 | 0.92 |
CatBoost model | Test set | 7270 | 0.96 | 0.98 | 0.98 | 0.80 | 0.92 |
Authors | Algorithm | No. of Features | Training Set (AFDB-2017) 1 | Test Set 1 | |||
---|---|---|---|---|---|---|---|
F1n | F1a | F1o | F1 | F1 | |||
Teijeiro et al. [23] | XGBoost and DNN | 79 | 0.94 2 | 0.90 2 | 0.84 2 | 0.89 2 | 0.83 2 |
Datta et al. [21] | DTE | 150 | 0.99 2 | 0.94 2 | 0.98 2 | 0.97 2 | 0.83 2 |
Zabihi et al. [25] | DTE | 491 | 0.98 2 | 0.93 2 | 0.95 2 | 0.95 2 | 0.83 2 |
Hong et al. [22] | DNN and XGBoost | 300 | 0.99 2 | 0.94 2 | 0.98 2 | 0.97 2 | 0.83 2 |
Zihlmann et al. [26] | DNN | - | 0.93 2 | 0.91 2 | 0.83 2 | 0.89 2 | 0.82 2 |
Xiong et al. [24] | DNN | - | 0.93 2 | 0.88 2 | 0.83 2 | 0.88 2 | 0.82 2 |
Our previous work [27] | DTE | 30 | 0.93 2 | 0.88 2 | 0.82 2 | 0.87 2 | 0.82 2 |
This work | CatBoost | 17 | 0.95 | 0.90 | 0.87 | 0.91 | - |
Algorithm | Sensitivity (%) | Specificity (%) | Accuracy (%) |
---|---|---|---|
Slocum et al. [13] | 62.80 | 77.46 | - |
Tateno and Glass [18] | 94.40 | 97.20 | - |
Sarkar et al. [17] | 97.50 | 99.00 | - |
Huang et al. [15] | 96.10 | 98.10 | - |
Lee et al. [16] | 98.20 | 97.70 | - |
Jiang et al. [36] | 98.20 | 97.50 | |
Zhou et al. [19] | 96.89 | 98.25 | 97.67 |
Asgari et al. [37] | 97.00 | 97.10 | 97.10 |
This work | 99.61 | 99.64 | 99.62 |
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Shao, M.; Zhou, Z.; Bin, G.; Bai, Y.; Wu, S. A Wearable Electrocardiogram Telemonitoring System for Atrial Fibrillation Detection. Sensors 2020, 20, 606. https://doi.org/10.3390/s20030606
Shao M, Zhou Z, Bin G, Bai Y, Wu S. A Wearable Electrocardiogram Telemonitoring System for Atrial Fibrillation Detection. Sensors. 2020; 20(3):606. https://doi.org/10.3390/s20030606
Chicago/Turabian StyleShao, Minggang, Zhuhuang Zhou, Guangyu Bin, Yanping Bai, and Shuicai Wu. 2020. "A Wearable Electrocardiogram Telemonitoring System for Atrial Fibrillation Detection" Sensors 20, no. 3: 606. https://doi.org/10.3390/s20030606