Influence of Photoplethysmogram Signal Quality on Pulse Arrival Time during Polysomnography
Round 1
Reviewer 1 Report
The present paper focuses on the influence of photoplethysmogram signal quality on pulse arrival time during polysomnography. The text itself is well readable (including the pseudocode) and contains only a few typos (e.g., numbering between line 95 and 96). The advantages of the proposed approach are, among others, the better computational efficiency of the algorithm, good temporal resolution, and robustness in the context of PPG recording quality assessment. It would be useful to list the disadvantages of the approach in the discussion. Some of them could emerge from the questions and comments below:
1. For what reason are the parts of the signal annotated as central apnea and periodic movement of the legs not evaluated?
2. The T wave search algorithm initially uses a range of Ri + 0.06 (s):Ri+1. If no local maximum is found, the range is then extended to Ri:Ri+1. Explain whether a shorter time (<0.06 (s)) between the peak of the R and T waves is physiologically possible. Describe why half the length of the QRS complex was chosen for the selection of the 0.06 (s) constant.
3. Describe the procedure for determining the thresholds θ1 and θ2 and the constant k.
4. Describe why a central signal was used to evaluate the DC component.
5. What test was used to assess the normality of the data? A nonparametric test (paired Wilcoxon signed rank test) combined with Cohen's d value is used in the evaluation. Explain the use of Cohen's d value in case of non-normal distribution of data.
I consider the possible wrong choice of statistical evaluation using Cohen's d value as the biggest error of the article. In my opinion, if we assume a non-normal distribution, the application of the median is not an option since Cohen's d uses the standard deviation, which is a parametric estimate of the distribution. For this reason, clarification or checking by a specialist in statistics is necessary, although there is no reason to assume that this will change the conclusions of the publication.
Author Response
Please see the attachment.
Additional information:
The revised version with noticible changes (red/blue colors) is Sensors2022_Rinkevicius_revised.pdf file.
The manuscript version after revision (without red/blue colors) is Sensors2022_Rinkevicius_after_rev.pdf file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
1. The authors proposed a novel algorithm to improve the PAT estimation in PPG and ECG signals. The proposal is very interesting and meaningful for evaluating the sleep state.
2. The process flow is explained clearly and the data analysis is shown in detail.
3. It is recommended to supplement the effect on the estimation accuracy by the averaging order M.
Author Response
Please see the attachment.
Additional information:
The revised version with noticible changes (red/blue colors) is Sensors2022_Rinkevicius_revised.pdf file.
The manuscript version after revision (without red/blue colors) is Sensors2022_Rinkevicius_after_rev.pdf file.
Author Response File: Author Response.pdf
Reviewer 3 Report
The aim of this paper is to develop a PPG signal quality assessment algorithm for robust PAT estimation, and investigate the influence of signal quality on PAT.
This paper is logically organized and clearly structured. However, several points should be enhanced.
Line 336, after post-processing and eliminating PAT measurements derived from low quality PPG signals deviations in PAT decreased. Why?
The orders of averaging and median filtering for PAT sequences were selected M = 5 and n = 15. Why?
What’s the accuracy of identifying R and T waves in ECG signals during various stages?
How to exclude low quality PPG segments from PSG analyses? How to judge low quality?
It will be better to compare the algorithm of this paper to other algorithms.
Author Response
Please see the attachment.
Additional information:
The revised version with noticible changes (red/blue colors) is Sensors2022_Rinkevicius_revised.pdf file.
The manuscript version after revision (without red/blue colors) is Sensors2022_Rinkevicius_after_rev.pdf file.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The reviewer's questions and comments were handled to his satisfaction.
Reviewer 3 Report
The revised paper has been improved a lot.