Development of Wearable Wireless Electrocardiogram Detection System using Bluetooth Low Energy

Round 1

Reviewer 1 Report

The paper presents a minimally invasive wearable system for recording of single-lead ECG signal exploiting two modules (master and slave) synchronized via BLE to be placed on wrists. Band and sampling rate achieved by the system are compatible for clinical use.  The paper is well structured and clear for the reader. It describes in details the system architecture, the data processing and how synchronization of data has been addressed. Experimental results and comparisons vs similar systems are provided.      

From the implementation point of view, it could be interesting to add some additional information about the filters implemented in the software.

Author Response

We appreciate the reviewers’ comments suggestions.

Our responses have been added to the attachment, and the main text has been modified based on the valuable comments provided.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper presents a novel method of collecting ECG signals using two wireless wearable modules - a main and a sub-module. The submodule transfers the electric potential to the main module via Bluetooth low energy, and the collected values are simultaneously displayed in the form of a graph. 

Although this paper presents a novel ECG wearable system and has the potential to be accepted, several important points have to be clarified or fixed before I can recommend an acceptance. 

1. The abstract should concisely report the aims and outcomes of the research so that readers exactly know what the paper is about and what is the outcome (results). The abstract should not contain the background of the research (e.g. the explanation of ECG, the explanation of portable devices, etc.). 
2. The introduction should briefly compare the proposed approach to the related and most recent work on the subject (wearable ECG monitors, BLE, ...). The authors should also clearly state the scientific contribution(s) of the proposed work at the end of the introduction.
3. Section 2.1: it is not clear whether the system in the first sentence is the proposed one. There are also several grammar mistakes in the first paragraph (e.g. It also included -> It also includes). 
4. All used components (MCU, BLE module, analog components, etc. ) should be clearly described at the beginning of Section 2.1. - possibly the best place is in Figure 1 and bellow Figure 1. Also, the authors should put a reference to each component used in the system (e.g. reference to datasheets)
5. Page 3, paragraph 2: it is not clear how the number of elements is minimized.
6. Figure 2: the quality of the figure is of low quality. The authors should use dedicated software to produce and export high-quality electrical diagrams and not just capture the screen's portion from PSoC Creator.
7. Page 3, paragraph 4: it is not clear why six-byte packet, what is the format of the packet, what data is transmitted in the packet?
8. Page 3, paragraph 4: the authors claim that they use Low-energy Bluetooth to transmit packets, but it is not clear how the BLE module has been configured to achieve low energy (e.g. connection interval, throughput)? I suggest the authors look at the paper: "Data Transmission Efficiency in Bluetooth Low Energy Versions" Sensors 19, no. 17: 3746. https://doi.org/10.3390/s19173746" 
9. Section 2.3: Please explain why testing the proposed method on only one subject is sufficient?
10. Section 2.3., paragraph 2: the authors mention an ECG simulator, but its usage and role are not explained in the paper. Also, there is a missing reference to the ECG simulator and the reference system.
11. Section 2.4: It is not clear how the 250Hz sampling period leads to distorted results. Please explain. 
12. Figure 4: it is not clear what is presented in Figure 4, synchronization or sampling? Also, it is not clear how the modules are synchronized. What are the labels M1, S1, ... ?
13. Section 2.5. The authors claim that the number of cycles, timer counter and packet size must be carefully adjusted, but they do not show how this is performed in their research and what are the chosen parameters in the proposed system.
14. Section 3.1., Paragraph 1: please add a diagram showing the proposed measurement setup with the reference and the proposed system. 
15. Section 3.1., Paragraph 1: The authors mention some previous studies but failed to reference them. 
16. Section 3.1., Paragraph 1: It is not clear the purpose of two diff amps for Bino modules.
17. Figure 5c: it is not clear whether the results are obtained from the reference modules or the proposed system?
18. Section 3.2.,paragraph 1: row data -> raw data?
19. Section 3.2. what is the purpose of the notch filter? Please add the band suppressed by the notch filter. It is not clear how a monitoring application is saved in the CSV format?
20. Please justify the measurements on only one subject.
21. The authors used the correlation coefficient to compare two discrete signals. In my opinion, this is not enough to draw any conclusions. The authors should also use some other methods to compare two signals, especially as there was only one tested subject. For example, the authors should compare the cumulative distribution functions of two signals (see for example Accurate Indoor Sound Level Measurement on a Low-Power and Low-Cost Wireless Sensor Node. Sensors 2018, 18, 2351. https://doi.org/10.3390/s18072351, page 18) and NRMSE and confidence interval between two signals (see for example: R.Pilipovic, V.Risojevic, P.Bulic. On the Design of an Energy Efficient Digital IIR A-weighting. Sensors 2021, 21(3), page 15)
22. References are incomplete and should contain the most recent related work on ECG measurements with wearable devices. The authors should compare their approach to these work and clearly state what are their contributions. E.g. :

Landreani, F.; Faini, A.; Martin-Yebra, A.; Morri, M.; Parati, G.; Caiani, E.G. Assessment of Ultra-Short Heart Variability Indices Derived by Smartphone Accelerometers for Stress Detection. Sensors 2019, 19, 3729. https://doi.org/10.3390/s19173729

Passler, S.; Müller, N.; Senner, V. In-Ear Pulse Rate Measurement: A Valid Alternative to Heart Rate Derived from Electrocardiography? Sensors 2019, 19, 3641. https://doi.org/10.3390/s19173641

Comparison of Continuous ECG Monitoring by Wearable Patch Device and Conventional Telemonitoring Device. J Korean Med Sci. 2020 Nov 16; 35(44): e363.

Hwang J, Kim J, Choi KJ, Cho MS, Nam GB, Kim YH. Assessing accuracy of wrist-worn wearable devices in measurement of paroxysmal supraventricular tachycardia heart rate. Korean Circ J. 2019;49(5):437–445

Yenikomshian M, Jarvis J, Patton C, Yee C, Mortimer R, Birnbaum H, et al. Cardiac arrhythmia detection outcomes among patients monitored with the Zio patch system: a systematic literature review. Curr Med Res Opin. 2019;35(10):1659–1670.

Walsh JA, 3rd, Topol EJ, Steinhubl SR. Novel wireless devices for cardiac monitoring. Circulation. 2014;130(7):573–581.

Lau JK, Lowres N, Neubeck L, Brieger DB, Sy RW, Galloway CD, et al. iPhone ECG application for community screening to detect silent atrial fibrillation: a novel technology to prevent stroke. Int J Cardiol. 2013;165(1):193–194. 

Seshadri DR, Bittel B, Browsky D, Houghtaling P, Drummond CK, Desai M, et al. Accuracy of the Apple watch 4 to measure heart rate in patients with atrial fibrillation. IEEE J Transl Eng Health Med. 2019;8:2700204.

Author Response

We appreciate the reviewers’ comments suggestions.

Our responses have been added to the attachment, and the main text has been modified based on the valuable comments provided.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper describes the development and implementation of a wireless ECG monitoring system with BLE interface. This kind of system are not novel and there are some commercial implementations of wereable ECG monitoring systems with wireless conectivity. I am not sure which are the contributions (when compared to other systems) claimed by the authors. In order to be able to substract the measurement of the left hand module and the right hand module, to calculate the potential between them, a commont reference is needed, so, either both modules have to be connected to each other or they have to be connected to a common point (for example  the right leg). If both modules don't have a common reference, the potential difference from left hand to right hand cannot be correctly computed. This is not properly explained in the paper. If the authors are assuming that they can obtain the potential difference withour a common reference, the assumption is completely wrong. If there is a common reference, then which is the contribution of the proposed system when compared to others? I would encourage the authors to explain this point better.

The general structure of the paper is a bit confusing and some parts (like the comparison of the "proposed system" with the "reference system" not particularly well described. A better discussion of why the reference system is the "bino-kit" (no reference provided), and of the conducted tests, is necessary. This section has to be thoroughly improved.

 

 

Author Response

We appreciate the reviewers’ comments suggestions.

Our responses have been added to the attachment, and the main text has been modified based on the valuable comments provided.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

In my opinion, the authors are still neglecting the error caused by the fact that the modules attached to each hand do not share a common reference.

According to the author's response "ECGs can be conducted by measuring the potential difference in the system. The proposed system has a set virtual ground in REF OUT, as shown in Fig. 3 of the manuscript. Therefore, the potential of the signal measured by these electrodes could be checked, and the measured value for each module is shown in 5 (c). It is possible to conduct the ECG using one electrode because it uses a virtual ground; 1.3–1.6 V is set as a virtual ground, and an ECG waveform is output by grasping the biopotential measured in the master module and the slave module.". However, since the two modules are battery powered and not connected to each other, the "virtual ground" of each module is decoupled from the "virtual ground" of the other module, so they are not the same "virtual ground". If the "virtual ground" of the two modules were interconnected, or they were connected to the patient's right leg, this would be true, but as they are floating there is no way to assure the proposed system is measuring the potential difference between the two hands.

In fact, the proposed system measures the voltage variations in the left and right hand through a decoupling capacitor (C_1 in fig. 2 schematic), and then substracts these signals. Of course, probably this could be an approximation to the ECG I derivation, and it could be useful for certain purposes (such as measuring the heart-rate), but I believe there could be some differences or distortions bewteen the obtained signal and the actual ECG-I signal that could render the obtained signal invalid for diagnosis, and such differences/distortions are not properly analysed or discussed in the paper.  

Another important source of error/distortions is the (lack of) synchronization between the left hand and right hand sampling and digitizing processes, as the authors themselves acknowledge in subsections 2.4 and 2.5. However, the explanation provided by the authors in these sections about how the required synchronization is achieved (if it is achieved at all), or about how the errors of a lack of syncrhonization are minimized,  is a bit obscure and even self-contradictory (from section 2.4, it is not even clear which is the sampling rate, 250, 500 or 750 Hz?). A better explanation and an error analysis should be mandatory also in these sections.

 

Author Response

We appreciate the reviewers’ comments suggestions.

Our responses have been added to the attachment, and the main text has been modified based on the valuable comments provided.

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

The authors have not made any change in the manuscript to address my previous comments.

In them ("point 2") I was not asking about the effect of the sampling frequency, but by the effect of synchronization errors between both modules. I also asked about the following sentence "If the 250 Hz sampling period is altered and both modules exhibit an asynchronous delay, the output of the ECG module may yield severely distorted results. Therefore, we designed the monitoring modules with a sampling period hat was 500 Hz higher than the above-mentioned rate, and synchronization issues were not considered." This sentence makes unclear which sampling rate is being set and needs clarification (I believe the sampling rate it's 250Hz, but then what is "500Hz higher than the above mentioned rate"). In any case, the main problem I was asking about was the synchronization, not the effect of the sampling rate. 

Of course if there exists important differences in the sampling period of the two modules, a great deal of distortion would happen from the begining, as the authors themselves explain. However, if a crystal-oscillator clock is used, great differences in the sampling rate are not expected to happen. But, even if this is the case, there is still a need of syncrhonization for the two modules for (at least) two reasons:

• They have to start sampling virtually at the same time. If they don't, there is an error derived from both signals not being acquired exacly at the same moment. 
• Even if the the sampling clock of the two modules are near the nominal frequency, there would probably be a drift over time, so the error of not being synchornized will become more and more significant as time goes by.

So my questions in "point 2" were:

• Please analyse/discuss in the manuscript the effect of these synchronizations errors.
• Please explain if these problems are solved/minimized in your system and how.

Regarding Author's answer to the "point 1" of my previous comments, I know that "The concept of the ECG system proposed in this study is that each module is connected wirelessly to eliminate the inconvenience of wires when users measure signals", but I think this has a cost in signal integrity that the authors do not discuss nor acknowledge in the manuscript. As for the explanation about the virtual ground in authors' reply "Since the power consumption between each module is almost the same, it can be said that the voltage of the virtual ground is the same", I am afraid I have to beg your pardon but I don't get it. Could you elaborate a bit more on this?

Author Response

We appreciate the reviewers’ comments and the editor’s suggestions. Our responses are given below, and the main text has been modified based on the valuable comments provided.

Please check the attached file.

Author Response File: Author Response.pdf

Round 4

Reviewer 3 Report

The authors have included in the manuscript the required information about synchronization.

I have still some concerns about the signal integrity of the system, but since the authors have provided some empirical evidence of the proposed system working properly in several scenarios, and none of the other reviewers seem to have similar concerns, I think the paper can be published now.