Machine-Learning-Based Elderly Stroke Monitoring System Using Electroencephalography Vital Signals

Round 1

Reviewer 1 Report

Thank you for the opportunity to review this manuscript. The purpose was the investigation of a health monitoring system to predict precursors of stroke diseases in real-time during walking based on EEG data.
I completely see the interest of the topic for the readership of the scope of Applied Sciences. However, the clarity of writing should be clearly improved and I have some lingering comments to improve the manuscript. Especially, the general structure should be clearly re-ordered according to the Author Guidelines of the Journal. After all reading, I still would rather like to see the whole manuscript embedded in a clear Introduction, Methods or Pre-Results and Results Section, and a differentiated discussion of the own findings according to the practical field. This could help to give the whole manuscript a more focused and clear structure.

Major general comments:

The study and the embedded findings have high practical interest and add very valuable content to the field of stroke pre-recognition; however, there are some major revisions required before taken into account for publication.

Native-speaker revision is recommended.

Abstract:

Try to point out more clearly, in terms of writing style, what the purpose, the plain results and deduced conclusions of your study are.

L19-22: This section about stroke definition and consequences emerged being misleading during reading. Place it at the beginning of the Introduction but reduce the content or consider if this is really necessary content for the Abstract. If not, you could leave that out.

Keywords:

Try to avoid to put abbreviations in parentheses in keywords section. If using abbreviations as keywords, set them as single keyword.

Manuscript:

Introduction:

General comments:

Generally, the introduction covers all necessary content deduce the purpose of the study. However, it appears in many ways too detailed and explanatory. Try to reduce the length of your Introduction by focusing on the really relevant theoretical background required to deduce the purpose of your study. Especially in L41-62; 75-97.

Please be more careful in referencing. I miss referencing at various occasions throughout the Introduction. I.e.: L59; L74; L96; L107

Specific Comments:

L99 et seq.: When addressing “other diseases”, you should motivate that before-hand. I suppose your focus should stay to stroke. Furthermore, it would be good to have a brief link from the theoretical background to the deducted purpose in terms of a sum-up of the gathered theoretical findings.

L100 et seq.: There is a missing theoretical link between the purpose and why these EEG data feature analysis is relevant towards stroke. Furthermore, you are anticipating own Methods and Results to detailed for the Introduction section. Clear outlining why the method is appropriate according to your purpose should be elaborated with clear referencing.

L115-120: You have to reduce content of your theoretical part: You cannot merge a literature review and a experimental study into one manuscript. This a too comprehensive and holistic approach. Focus on the purpose of your study.

Section: Related Works

Generally, I do not see any reason for structuring your manuscript in a two-part Introduction. These chapters have to be merged into one Introduction section, covering all the related works necessary for clearly deducing your research questions or purpose of your study.

L122-176: This is a vast amount of information about EEG measurement and studies. You have to clearly reduce this section; short general introduction of EEG measurements and related studies and especially, concentrate on the final section of that part from L163-176. This is the really related content to your study. The other stuff is nice to read but often too far off your purpose.

L177-225: There is significant content in this section. But as mentioned before, here as well, it’s too vast. You have to merge this with the sections before, to give one Introduction section, which covers all necessary content for your study.

Section: Elderly Stroke Monitoring System Based on Machine Learning and EEG

This appears to be your Methods section, wherein you explain, firstly, the general monitoring system you want to experimentally test and practically apply depending on your findings, right? These represent your methodological approach! Therefore, this should be outlined respectively that the reader can follow your ideas. Please be consistent to the Journal’s general structure – Introduction, Methods, Results, Discussion and Conclusion.

L241-271: Way too long and you’re mixing Methods and Results. Please, revise.

L244-246: Didn’t you tell us this content already before? Try to avoid repetitions.

L254-255: These are results you’re suddenly telling. But this is not fitting here – first give us your whole methods you applied. Afterwards, you jump back to your hypotheses – this very confusing. And back to Methods again (L259-269). You have to re-structure this whole section.

L296-342: I acknowledge your thorough elaboration in this section. However, it appears too detailed for me. Consider reducing the length of re-explaining the equations of Hall’s method by reporting that you applied his method/referencing. Afterwards, report your results of the feature subset extraction.
This part should appear as part of the Methods or Results Section.

Section 4: Experimental Results

Table 5: Put footnotes in the Table/Figure’s caption. Quality of Figure captions could generally get increased. Revise accordingly as Figures/Tables should be self-explanatory.

L505-577: Revise section according to clarity in writing. E.g., L523-524.

I miss a discussion of your experimental results according to the related scientific field. How practically meaningful are your results in comparison to other monitoring systems. What are the clear practical implications in daily application for the medical staff etc.

You try to address that briefly in your Conclusions; however, I’d like to have also a more differentiated discussion of your whole work compared to L579-581. The conclusions and discussing appears suddenly pretty general after being rather explicit before.

I would appreciate an overall Discussion and Conclusion section of the whole work. Reduce the detailed Introduction / Theoretical Background and increase the differentiated comprehensive discussing of your findings specifically and generally.

Author Response

Dear Editors and MDPI Sustainability,

Manuscript ID: applsci-1111758

Title: Machine Learning-Based Elderly Stroke Monitoring System Using Electroencephalography Vital-Signals

We would like to thank editor and reviewers for your meticulous and useful comments. We hereby mention that your constructive criticism has greatly contributed to enhancing level of completion of this manuscript. We did our best efforts to make manuscript concise and clear. We remove less important references, added additional references, and rearranged reference lists. We will describe comprehensive answer in authors’ position about the point of one reviewer one by one faithfully. The ultimate goal of this study is to help medical staff to accurately predict and diagnose stroke diseases by means of the machine learning methodology and vital-signals collected in daily life.

Best Regards,

Dr. Jaehak Yu

Department of KSB Convergence Research, Electronics and Telecommunications Research Institute (ETRI), Republic of Korea (South Korea).

[email protected]

Author Response File: Author Response.docx

Reviewer 2 Report

Interesting study!

Author Response

Dear Editors and MDPI Sustainability,

Manuscript ID: applsci-1111758

Title: Machine Learning-Based Elderly Stroke Monitoring System Using Electroencephalography Vital-Signals

We would like to thank editor and reviewers for your favorable comments. We did our best efforts to make manuscript concise and clear.

Best Regards,

Dr. Jaehak Yu

Department of KSB Convergence Research, Electronics and Telecommunications Research Institute (ETRI), Republic of Korea (South Korea).

[email protected]

Author Response File: Author Response.docx

Reviewer 3 Report

Dear authors,

It is an interesting project with an active research area. Please see the attached file.

Kind regards,

Comments for author File: Comments.docx

Author Response

Dear Editors and MDPI Sustainability,

Manuscript ID: applsci-1111758

Title: Machine Learning-Based Elderly Stroke Monitoring System Using Electroencephalography Vital-Signals

We would like to thank editor and reviewers for your meticulous and useful comments. We hereby mention that your constructive criticism has greatly contributed to enhancing level of completion of this manuscript. We did our best efforts to make manuscript concise and clear. We remove less important references, added additional references, and rearranged reference lists. We will describe comprehensive answer in authors’ position about the point of one reviewer one by one faithfully. The ultimate goal of this study is to help medical staff to accurately predict and diagnose stroke diseases by means of the AI methodology and vital-signals collected in daily life.

Best Regards,

Dr. Jaehak Yu

Department of KSB Convergence Research, Electronics and Telecommunications Research Institute (ETRI), Republic of Korea (South Korea).

[email protected]

Author Response File: Author Response.docx