PONE-D-19-23638

Efficient Neural Spike Sorting using Data Subdivision and Unification

PLOS ONE

Dear Dr. Bhatti,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

(1) Abstract.

What do you mean by "big data dynamics"? This term is ambiguous. Please change this sentence and the next one, immediately following. Rephrase in order to avoid reference to other algorithms reported in the literature if the Authors do not cite explicitly which ones. The current reference is too general and inappropriate.

How many datasets did you use?

(2) Introduction.

Please, focus the introduction on the problems addressed and thoroughly review the literature and the current state of the art in the field.

1. The review of the literature is not complete, because it missed one key important paper related to this topic, in particular because that paper has introduced for the first time a series of steps that are very close, if not identical, to the steps of data subdivision, clusters formed for each sub-set, unification process by merging neighbor clusters in feature space, thus achieving unified clusters in the end. This paper is the following:

- Aksenova TI, Chibirova OK, Dryga OA, Tetko IV, Benabid AL, Villa AE. An unsupervised automatic method for sorting neuronal spike waveforms in awake and freely moving animals. Methods. 2003; 30(2):178-187. doi: 10.1016/S1046-2023(03)00079-3 : this is the very first paper (2003) to describe unsupervised neural spike sorting based on a fast implementation suitable for real-time application for high-density neural probes.

With respect to application of spike sorting to online experimental procedures, the Authors should also mention:

- Abeles M, Goldstein MH. Multispike train analysis. Proceedings of the IEEE. 1977; 65(5):762-773. doi:10.1109/PROC.1977.10559 : this is a seminal paper (1977) for detecting and identifying the spikes in multispike trains based on signal detection by template matching.

- Wouters J, Kloosterman F, Bertrand A. Towards online spike sorting for high-density neural probes using discriminative template matching with suppression of interfering spikes. J Neural Eng. 2018; 15(5):056005. doi: 10.1088/1741-2552/aace8a : a fast and computationally cheap method for real-time applications.

Consider  recently developed spike sorting algorithms :

Chung, Jason E., Jeremy F. Magland, Alex H. Barnett, Vanessa M. Tolosa, Angela C. Tooker, Kye Y. Lee, Kedar G. Shah, Sarah H. Felix, Loren M. Frank, and Leslie F. Greengard. "A fully automated approach to spike sorting." Neuron 95, no. 6 (2017): 1381-1394.

A more satisfactory review of the literature should also include:

- Zamani M, Demosthenous A. (2014) Feature extraction using extrema sampling of discrete derivatives for spike sorting in implantable upper-limb neural prostheses. IEEE Trans Neural Syst Rehabil Eng. 2014 Jul;22(4):716-726. doi: 10.1109/TNSRE.2014.2309678.

(3) Materials and Methods.

The Authors mention several times the problem of noisy recordings, but they do not examine which types of noise --and/or artifacts-- are present and the methods to face this problem that have been described in the recent literature.

A better way to compare the methods presented by the Authors in their Table 2 and Table 3 could have been to add several known levels of noise to the same benchmarked data set and see how performances and accuracies allow to discriminate the most robust algorithms.

To this end, the Authors should consider these papers:

- Choi JH, Jung HK, Kim T. (2006) A new action potential detector using the MTEO and its effects on spike sorting systems at low signal-to-noise ratios. IEEE Trans Biomed Eng. 2006 Apr;53(4):738-46. doi: 10.1109/TBME.2006.870239

- Paralikar KJ, Rao CR, Clement RS. (2009) New approaches to eliminating common-noise artifacts in recordings from intracortical microelectrode arrays: inter-electrode correlation and virtual referencing. J Neurosci Methods. 2009 Jun 30;181(1):27-35. doi: 10.1016/j.jneumeth.2009.04.014.

- Pillow JW1, Shlens J, Chichilnisky EJ, Simoncelli EP. (2013) A model-based spike sorting algorithm for removing correlation artifacts in multi-neuron recordings. PLoS One. 2013 May 3;8(5):e62123. doi: 10.1371/journal.pone.0062123.

- Takekawa T, Ota K, Murayama M, Fukai T. (2014) Spike detection from noisy neural data in linear-probe recordings. Eur J Neurosci. 2014 Jun;39(11):1943-50. doi: 10.1111/ejn.12614:  an older reference to Takekawa is provided but it should be replaced by this one .

The Authors discuss Spike sorting accuracy (Subsection 3.5) but false alarm ratio is also an extremely important feature to be considered (and discussed in several papers cited above) for the evaluation of the quality of neural spike sorting.

(4) Results.

The Authors should provide the MATLAB codes, with the description of the MATLAB version and environment, of their algorithms.  They compare many methods developed elsewhere and it is of paramount importance to assess that the Authors' implementation follows exactly the algorithms cited in the literature.

A test against a surrogate data set could also be informative for the readers to be convinced of their superior efficiency in the spike sorting procedure claimed by the Authors.

-Optimal length: describe how relevant it is to have the 'optimal length'.

Please, substantiate: 'OL parameter is dependent on the algorithm type rather than on the data dynamics.' The spiking rates may vary by 2 orders of magnitude, so you may end up with clusters that simply don't have enough spikes?

Clarify labeling in Figure 4.

Unification of subclusters:

Describe in detail how you account for differing variances in different dimensions (i.e. principal components).   Explain what 'the standard distribution and normal distribution curves' are.

In general, describe how this technique is applied to the data. Do you apply it to sequential segments, blocks of segments or pairwise across the recording?

Performance evaluation:

Why do you choose two examples where both the conventional and your method do not work for showing performance improvement?

Figure 6: Why those spline fits? Suggests that the different methods are related, please, explain.

Table 3: Numbers suggest a very high accuracy, and no error estimate is given. How did you achieve such a high precision? K-means for example is known to give very different results in different runs. Are these averages over multiple runs? And does the K-means example involve multiple runs to obtain stable clusters? Which of these algorithms converge to the same result every time they are run? Could part of your accuracy improvement be due to running K-means more often, effectively averaging results?

Figure 7:

Lines/symbols are overlapping to an extent that this figure becomes uninformative. Maybe separate plots or cluster centroids for different segments? Please, provide a plot showing the temporal stationarity of firing rates (for different segments).

Please, clarify description of the algorithm concerning temporal speedup. What is the advantage of independent clustering? How does your method compare to a density based approach?

clustering accuracy:

The measure you are using puts a higher weight on large clusters with a lot of spikes. In many datasets, these are multiunit clusters that are hard to separate. It would be nice to have some measure of temporal stationarity.

We would appreciate receiving your revised manuscript by Nov 29 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Gennady Cymbalyuk, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1) Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2)  Thank you for stating the following in the Acknowledgments Section of your manuscript:

[The research work is fully supported by Neural and Cognitive Systems Lab at Institute

for Intelligent Systems Research and Innovation, Deakin University.]

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

 [The author(s) received no specific funding for this work.]

Please include the updated Funding Statement in your cover letter. We will change the online submission form on your behalf.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: N/A

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors address a relevant problem in spike sorting, namely how to deal with datasets from recordings that become increasingly long due to technological advances in recording techniques.

However, the presentation of their results and statistical analyses do not allow me to make any judgements about the validity of their approach. In fact, I believe that the amount of changes that would be necessary for a revised version of this manuscript would effectively amount to a resubmission of the manuscript.

Specifically, the introduction consists of a rather broad discussion about measuring brain activity and its relevance (not immediately related to the manuscript), but almost completely ignores recently developed spike sorting algorithms (e.g.

Chung, Jason E., Jeremy F. Magland, Alex H. Barnett, Vanessa M. Tolosa, Angela C. Tooker, Kye Y. Lee, Kedar G. Shah, Sarah H. Felix, Loren M. Frank, and Leslie F. Greengard. "A fully automated approach to spike sorting." Neuron 95, no. 6 (2017): 1381-1394.

and the sorting algorithms they use for comparison). It would good to have an introduction more specific to the manuscript and especially describing the current state of the art in the field.

-Optimal length: what I missed here is a discussion of how relevant it is to have the 'optimal length'. Can I be off by a factor of 2 and it doesn't really matter?

Also, I'm not sure how the authors come up with this claim: 'OL parameter is dependent on the algorithm type rather than on the data dynamics.' This may be the case in machine learning examples, but here spiking rates may vary by 2 orders of magnitude, so you may end up with clusters that simply don't have enough spikes?

Labelling in Figure 4 is messy, I don't understand what is plotted.

Unification of subclusters:

I don't understand how you account for differing variances in different dimensions (i.e. principal components). And for distances, in 1D, the 95% claim is valid, but here you're talking about volumes. And I'm completely lost about what 'the standard distribution and normal distribution curves' are.

In general, I'm wondering how this technique is applied to the data. Do you apply it to sequential segments, blocks of segments or pairwise across the recording?

Performance evaluation:

Most strikingly, why do you choose two examples where both the conventional and your method do not work for showing performance improvement? Seems not relevant to the reader.

Figure 6: Why those spline fits? Suggests that the different methods are related, but I do not see how.

Table 3: Numbers suggest a very high accuracy, and no error estimate is given. How did you achieve such a high precision? K-means for example is known to give very different results in different runs. Are these averages over multiple runs? And does the K-means example involve multiple runs to obtain stable clusters? Which of these algorithms converge to the same result every time they are run? I am also concerned that part of your accuracy improvement might be due to running K-means more often, effectively averaging results.

Figure 7:

Lines/symbols are overlapping to an extent that this figure becomes uninformative. Maybe separate plots or cluster centroids for different segments? What I am missing here is also a plot showing the temporal stationarity of firing rates (for different segments).

temporal speedup:

If I understood things correctly (and I'm not sure I did), PCA/Wavelet is run on the whole dataset to obtain low dimensional representations of spikes. Then batches of N spikes are clustered. That sounds similar to what Kilosort does, except that batches are used for optimizing clusters rather than clustering them independently. What is the advantage of independent clustering? Mountainsort on the other hand follows a density based approach, which also seems to scale quite well with recording size. How does your method compare to a density based approach?

clustering accuracy:

The measure you are using puts a higher weight on large clusters with a lot of spikes. In many datasets, these are multiunit clusters that are hard to separate. Also, It would be nice to have some measure of temporal stationarity.

Abstract:

6 or 3 datasets?

Reviewer #2: GENERAL COMMENTS

---------------------------------

The Authors present an interesting manuscript about an efficient method to apply spike sorting on large data sets -- in the order of several hundreds of multiple spike trains recorded simultaneously. This topic is central for any project aimed at real-time decoding of brain activity, in particular for brain-machine interfaces. The paper is well written and reads easily. The main principles and methods are clearly presented and the figures are well done. The recommendation is to accept the paper, but there are few suggested corrections to introduce and the paper may be accepted only after appropriate amendments are introduced.

SPECIFIC COMMENTS

---------------------------------

(1) Abstract.

The Authors use the expression "big data dynamics". What does it mean? This sounds a bit weird because it may assume so many different meanings. Please change this sentence and the next one, immediately following. Rephrase in order to avoid reference to other algorithms reported in the literature if the Authors do not cite explicitely which ones. The current reference is too general and inappropriate.

(2) Introduction.

1. The review of the literature is not complete, because it missed one key important paper related to this topic, in particular because that paper has introduced for the first time a series of steps that are very close, if not identical, to the steps of data subdivision, clusters formed for each sub-set, unification process by merging neighbor clusters in feature space, thus achieving unified clusters in the end. This paper is the following:

- Aksenova TI, Chibirova OK, Dryga OA, Tetko IV, Benabid AL, Villa AE. An unsupervised automatic method for sorting neuronal spike waveforms in awake and freely moving animals. Methods. 2003; 30(2):178-187. doi: 10.1016/S1046-2023(03)00079-3 : this is the very first paper (2003) to describe unsupervised neural spike sorting based on a fast implementation suitable for real-time application for high-density neural probes.

With respect to application of spike sorting to online experimental procedures, the Authors should also mention:

- Abeles M, Goldstein MH. Multispike train analysis. Proceedings of the IEEE. 1977; 65(5):762-773. doi:10.1109/PROC.1977.10559 : this is a seminal paper (1977) for detecting and identifying the spikes in multispike trains based on signal detection by template matching.

- Wouters J, Kloosterman F, Bertrand A. Towards online spike sorting for high-density neural probes using discriminative template matching with suppression of interfering spikes. J Neural Eng. 2018; 15(5):056005. doi: 10.1088/1741-2552/aace8a : a fast and computationally cheap method for real-time applications.

A more satisfactory review of the literature should also include:

- Zamani M, Demosthenous A. (2014) Feature extraction using extrema sampling of discrete derivatives for spike sorting in implantable upper-limb neural prostheses. IEEE Trans Neural Syst Rehabil Eng. 2014 Jul;22(4):716-726. doi: 10.1109/TNSRE.2014.2309678.

(3) Materials and Methods.

The Authors mention several times the problem of noisy recordings, but they do not examine which types of noise --and/or artifacts-- are present and the methods to face this problem that have been described in the recent literature.

A better way to compare the methods presented by the Authors in their Table 2 and Table 3 could have been to add several known levels of noise to the same benchmarked data set and see how performances and accuracies allow to discriminate the most robust algorithms.

To this end, the Authors should consider these papers:

- Choi JH, Jung HK, Kim T. (2006) A new action potential detector using the MTEO and its effects on spike sorting systems at low signal-to-noise ratios. IEEE Trans Biomed Eng. 2006 Apr;53(4):738-46. doi: 10.1109/TBME.2006.870239

- Paralikar KJ, Rao CR, Clement RS. (2009) New approaches to eliminating common-noise artifacts in recordings from intracortical microelectrode arrays: inter-electrode correlation and virtual referencing. J Neurosci Methods. 2009 Jun 30;181(1):27-35. doi: 10.1016/j.jneumeth.2009.04.014.

- Pillow JW1, Shlens J, Chichilnisky EJ, Simoncelli EP. (2013) A model-based spike sorting algorithm for removing correlation artifacts in multi-neuron recordings. PLoS One. 2013 May 3;8(5):e62123. doi: 10.1371/journal.pone.0062123.

- Takekawa T, Ota K, Murayama M, Fukai T. (2014) Spike detection from noisy neural data in linear-probe recordings. Eur J Neurosci. 2014 Jun;39(11):1943-50. doi: 10.1111/ejn.12614: an older reference to Takekawa is provided but it should be replaced by this one .

The Authors discuss Spike sorting accuracy (Subsection 3.5) but false alarm ratio is also an extremely important feature to be considered (and discussed in several papers cited above) for the evaluation of the quality of neural spike sorting.

(4) Results.

The Authors should provide the MATLAB codes, with the description of the MATLAB version and environment, of their algorithms. They compare many methods developed elsewhere and it is of paramount importance to assess that the Authors' implementation follows exactly the algorithms cited in the literature.

A test against a surrogate data set could also be informative for the readers to be convinced of their superior efficiency in the spike sorting procedure claimed by the Authors.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

PONE-D-19-23638R1

Efficient Neural Spike Sorting using Data Subdivision and Unification

PLOS ONE

Dear Dr. Bhatti,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please, seriously revise the manuscript to to clarify the concerns described  below and  fix  typos.

--Figure 7: Lines/symbols are overlapping to an extent that this figure becomes uninformative.

Maybe separate plots or cluster centroids for different segments? Please, provide a plot

showing the temporal stationarity of firing rates (for different segments).

--10) Temporal speedup: Please, clarify description of the algorithm concerning temporal speedup.

What is the advantage of independent clustering? How does your method compare to a

density based approach?

--11) Clustering accuracy: The measure you are using puts a higher weight on large clusters with a

lot of spikes. In many datasets, these are multiunit clusters that are hard to separate. It would

be nice to have some measure of temporal stationarity.

Other comments:

Please, clarify what readers are supposed to see in the example Figs 7+8 Both the conventional and the proposed method seem to produce identical results, but the sequence of plotting the different lines has changed. The bottom graphs are identical. Are these placeholder figures?

-'The surrounding region between −2SD to 2SD, containing about 95 percent of the

cluster data...' This statement is still wrong.

-What are 'Quirogo datasets'?

We would appreciate receiving your revised manuscript by Feb 06 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Gennady Cymbalyuk, Ph.D.

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I believe the authors still need more time to polish their manuscript. There are a lot of typos, and a few of my previous comments have not been addressed, specifically:

--Figure 7: Lines/symbols are overlapping to an extent that this figure becomes uninformative.

Maybe separate plots or cluster centroids for different segments? Please, provide a plot

showing the temporal stationarity of firing rates (for different segments).

--10) Temporal speedup: Please, clarify description of the algorithm concerning temporal speedup.

What is the advantage of independent clustering? How does your method compare to a

density based approach?

--11) Clustering accuracy: The measure you are using puts a higher weight on large clusters with a

lot of spikes. In many datasets, these are multiunit clusters that are hard to separate. It would

be nice to have some measure of temporal stationarity.

Other comments:

I'm still not sure what I'm supposed to see in the example Figs 7+8 Both the conventional and the proposed method seem to produce identical results, but the sequence of plotting the different lines has changed. The bottom graphs are identical. Are these placeholder figures?

-'The surrounding region between −2SD to 2SD, containing about 95 percent of the

cluster data...' This statement is still wrong.

-What are 'Quirogo datasets'?

Reviewer #2: The manuscript can be processed for publication as is. All comments have been addressed adequately

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

PONE-D-19-23638R2

Efficient Neural Spike Sorting using Data Subdivision and Unification

PLOS ONE

Dear Dr. Bhatti,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Jun 21 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Alexandros Iosifidis

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Please address the comments of Reviewer 1.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Apart from a methological issue pointed out below (which needs to be discussed), a few missing details in the Methods and some awkward sentences and typos (I probably missed some and would encourage the authors to do another round of proofreading), this manuscript is now in a good shape.

Main points:

-please fix typos and grammatical errors (see below for a list of suggestions).

-I think I'm still missing some crucial information about the analysis. First I thought, that the performance improvement was somewhat related to nonstationarities in the data and you have shown (great, thanks) that this is clearly not the case. Another thing that I kept pointing out in my reviews and is still somewhat misleading in the presentation of the method is that in a high dimensional multivariate gaussian distribution, the probability for a datapoint to be within a 2 sigma radius from the center is not 95% but rather dependent on the number of dimensions, i.e. at most (95%)^d (for L1 norm), where d is the number of dimensions (or PCA components/ features). I haven't really found the number of dimensions you used in the paper (and you really do need to report it, it is a crucial number), but there is one figure suggesting the use of 10 features/dimensions. This seems high to me (and you may want to discuss such a parameter choice in the Discussion), what would have expected from other work would be 3-4 features. In any case, in the 10 feature case, your 2 sigma radius then accounts for at most 60% of the datapoints, so there are a lot of points outside your cluster boundaries. Does that explain why those widely used algorithms are working so poorly? If so, that's fine, but you want to discuss it in the Discussion section. It is also not clear to me how different dimensions are handled and you should elaborate a bit on that in the Methods. Is each dimension scaled such that variances match? If that is the case you're downweighting the first principal component and effectively explaining noisy, low variance features? Or am I missing something more subtle? You're reporting a performance improvement and I still don't see any reason why this should happen and especially why it would happen so consistently, given that all these algorithms have been used very successfully for years. I'm totally fine with the speed improvement and follow the argument that this should happen. But a general classification performance improvement is very hard to believe, so you need to at least report the specific circumstances under which it happens, i.e. the number of features/ dimensions and make clear that you're potentially inflating tiny differences in principal components with small variances (unless you corrected for that in some way, in that case it should be reported). Ideally, you should have some idea about a mechanism for the the performance improvement and discuss it in the Discussion (is it some kind of regularization effect that would be beneficial for noisy data?).

Specifically, do report the number of features/PCA components used. Do make clear whether the standard deviation was estimated for each component separately, thus enhancing the effect of small components, or whether (and how) you accounted for differences in the variances of features/PCA components. Ideally, specify a typical variation between variances of the features/PCA components (e.g. ratio between largest and smallest) and mention whether the results were sensitive to the number of PCA components.

A thorough analysis of the effect of dimensionality and scaling is certainly beyond the scope of this article, but I'm sure you made observations what happens if you change these parameters. You shall discuss them in the Discussion, and maybe even speculate about a mechanism or a scenario that tends to give performance improvements.

-Figure 7 has errorbars now, so please mention briefly how you obtained them/what they reflect. Further, numbers reported suggest a huge precision in comparison to these errorbars. Please round them, and wherever refered to in the text, add the uncertainty in brackets (e.g. 53+-6 %). You may leave the uncertainty in the table for clarity as it is already shown in Figure 7.

Other remarks

Figure 8+9: markers and labels don't match.

ln65: Brain consists

ln105: automatically estimate

ln119: presented data analysis issues due to progressive technological advancements of neural recordings

ln126: Although they have proposed an

efficient method for spike sorting, it still lacks the speed researchers require

ln130:The larger is the size the slower is the speed and large is the computational time required by spike sorting algorithms. -- rephrase or simply leave out (what, other than the obvious, are you trying to say?)

ln133: They reported, (?)

ln136: These second and third order operations prove the non-linear behaviour of spectral clustering.

ln138: To motivate our analysis,

ln141: The dependency of speed and computational time on data

size in spike-sorting has made it very difficult

ln142: identify the total number of

ln144: breast cancer cell data

ln149: Despite these challenges, ...

ln151: However, limited work has considered enhancing computational

ln153: The proposed algorithm pre-processes data to

ln154: time and to enhance speed and efficiency of a wide range

ln156: by parallel computing approaches to further

ln159: The novelty of the proposed mechanism

ln162:The first step involves subdivision of data into data-subsets of optimal length.

ln164:The second step involves clustering spikes in data-subsets

using conventional spike sorting algorithms.

ln165: The last step involves unification

ln166: clusters are then used to label

ln170:of conventional algorithms but rather performs additional data

ln171: the proposed mechanism very versatile and

ln175: uses a density based

The second step involves clustering

The last step involves unification

ln180: overall time of the spike sorting process.

ln193:The total number N of optimal subdivisions is estimated

ln195, 199: ,where L is the

ln223: of the algorithm depends on the length

ln227: (O L ) forms a direct

ln228: and an inverse relationship

the X-axis

the Y-axis

The computational time is the processing time after a movmean

filter (20 datapoints length) filtered the unwanted ripples in the plot and returned smooth curves. (representing computational time (why twice?))

The average value over ten repetitive analyses

robustness of the measure

ln241:'It is observed that, the

variations in data dimensionality does not have any effect on estimating the bounded

region. Whatever is the dimensionality of data, when the ED is calculated, the result is

always a single entry in one-dimensional space. For all EDs The standard deviation

(SD) is calculated using [66] and normal distribution curves are formed based on [67].' --Not even wrong. If you have a multivariate Gaussian distribution, the density distribution as a function of the radius is not Gaussian. The square of the radius (equals the sum of squares of Gaussian distributed random variables and) follows a Chi-squared distribution (check Wikipedia?) and you can imagine (take the cumulative distribution and rescale the x-axis) what follows for the distribution of the radius itself.

Figure 7: Continues positive trend is observed ???

Errorbars represent...

ln344: To cater for schocastic ??? variations of some of the algorithms

ln335: over 10 repetitons.

Reviewer #2: The latest revision and revised/new figures made the manuscript even more clear. The manuscript can be published as is.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

PONE-D-19-23638R3

Efficient Neural Spike Sorting using Data Subdivision and Unification

PLOS ONE

Dear Dr. Bhatti,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Sep 17 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Alexandros Iosifidis

Academic Editor

PLOS ONE

Journal Requirements:

Additional Editor Comments (if provided):

Please address the comments of Reviewer 1, by making sure that all definitions in the paper are precise.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Lines 237-253: Please talk to a statistician (or someone who knows English and statistics) and reframe (and please put references from peer reviewed publications). This is about making your paper understandable to a reader, and I'm not asking for a layman explanation here (in fact, I feel that you're trying to explain a lot of things you don't need to explain, I'm asking for correctness. The standard deviation of a random variable is a well known and defined quantity and your equation does not reflect the standard deviation of euclidean distances.

The euclidean distances are strictly positive numbers, but in Figure 6, you're suggesting that they are Gaussian distributed, and therefore negative values are possible. So I really don't understand what you are doing here. And I would also be very interested whether you scale your principal components in some way, to match their variances, or whether the first principal components have larger weights.

The main issue that I raised in the last revision was that when you're working in a 10 dimensional space, things are a little more complicated. For example, if you have a standard normal distribution in 10 dimensions, then the euclidean distances (ED) follow a Chi-square distribution with 10 degrees of freedom (see https://en.wikipedia.org/wiki/Chi-square_distribution).

other comments:

ln. 288 kolmogorov-Smirnov (KS) test --> Kolmogorov-Smirnov (KS) test

ln. 296 Please make clear what you mean by this sentence: 'It is

observed that 10 PCA features ensures the cumulative explained variance of over 85%

up to 95%, in case of the data sets employed in this study.' Please reframe this sentence.

The matlab function-- Please capitalize MATLAB.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PONE-D-19-23638R4

Efficient Neural Spike Sorting using Data Subdivision and Unification

PLOS ONE

Dear Dr. Bhatti,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Jan 24 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Alexandros Iosifidis

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Please address the comments provided by the Reviewer on the current version of the paper, and submit a point-to-point response letter with the revised paper.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: There are a few final edits I'd suggest in the modified part of the manuscript to improve readability, but I think the manuscript is in good shape now, and the methods are presented clearly.

lines 275-286 and Figure 7: I'd suggest to remove this paragraph and Figure 7 since

- I'm not sure how the explanation of a z-score helps in understanding the method.

- you are not plotting normal distributions in any of the Figures anymore, you defined the standard deviation in (7), so Equation (9) is not necessary (you're not using any other property of the normal distribution other than its standard deviation, and you cannot assume that the distribution of z scores is normal).

line 287: A short motivational sentence about what you're planning to do with the z score values would be great, e.g. 'We wanted to determine outliers for each spike cluster. To this aim, we considered two scenarios where the z scores distributions of a given cluster were either consistent with a normal distribution or skewed. There are numerous...'

if you rather want to keep these lines:

line 287: data distribution --> z-score distribution

line 284: Euclidean

line 281, 283: 'used to plot', 'is plotted': I don't find these plots anywhere, so please reformulate.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Efficient Neural Spike Sorting using Data Subdivision and Unification

PONE-D-19-23638R5

Dear Dr. Bhatti,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Alexandros Iosifidis

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Authors addressed all Reviewers' comments. Congratulations for the acceptance of your paper.

Reviewers' comments: