A Practical Multiclass Classification Network for the Diagnosis of Alzheimer’s Disease

Round 1

Reviewer 1 Report

The paper is a thorough well written work wherein the authors have documented both the prior literature as well as their proposed multi-class classification network with an aim of predicting different stages of Alzimers in detail. I have minor comments: For the comparison table, instead of 'ours' please use 'this work'. Also please improve the font sizes in the box plots in fig. 4 for ease to the reader. Good work !

Author Response

Re: applsci-1720726

 Title: A Practical Multiclass Classification Network for Diagnosis of Alzheimer’s disease

Dear Editor-in-chief, Associate Editor, Reviewers,

We would like to thank the editor and all reviewers for the hard work in reviewing our paper. In particular, we highly appreciate the constructive suggestions given by the editor/reviewers, which greatly help us improve the quality of our manuscript. Concerns from the reviewers are mainly about the organization and presentation of the paper. In the revision, we have addressed all the concerns raised by the reviewers carefully and revised the manuscript substantially. We have checked the whole paper for the omission of grammatical errors and improved the visual quality of the graphs to present the pros of it better.

We hope these revisions strengthen the paper and make it meet the criterion of this journal. Our point-to-point responses to comments are given in detail as below. We also highlight the revised contents in the main manuscript by red color and attach them behind our response to each comment to facilitate the review.

Best regards,

Rizwan Khan. et al.

Author Response File: Author Response.pdf

Reviewer 2 Report

MINOR THINGS:

1) There is inconsistent writing of multi-class, e.g. in heading and Keywords is „Multiclass classification“ written, in the text different versions e.g. multi-class (line 10) and multiple class classification (line 73).

2) In the list of authors, commas are sometimes superscripted and once subscripted (first author), please check for consistency.

3) The sign ‡ indicate authors which are contributed equally but only the first author has this sign. Please make clear which authors are contributed equally.

4) There is a bracket at the end of the list of Keywords, please delete.

5) The numbered refence style looks currently like [1][2] and should be changed to this style: [1-2] (see line  24, 33, 49,….)

6) It is recommended to check intensively the used abbrevations and their explanations e.g. „magnetic resonance imaging (MRI)“ was used first in the Abstract and used as MRI in the introduction, but in chapter 2 it is again written „magnetic resonance images (MRIs)“ (line 83). For „SVM“ (line 52) there is no explanation.

7) The use and the linking to the Figures in the text need to be optimized. The Fig 1 and 2 are mentioned in the text for the first time in chapter 3 (line 143 & 145) and Fig 3 is linked in text firstly ( line 115). The interposition of the Figures need to be optimized that they fit better to the written text.

8) Inconsistence in the use of „Alzheimer Disease“, „Alzheimer’s Disease“, „Alzheimer disease“, and AD what need to be optimized.

9) In order to make the manuscript more understandable to a broad readership, the Figures would profit from some more descriptions, e.g. what is presented in Fig. 3: is a MRI scan, of which parameters. Is it a patient with AD or not?

10) Other minor things what need to be corrected: pre-processing vs. prepossessing (line 205); state-of-the-art approaches  vs. state of the arts approaches (line 269); mixture of dots and commas: a), where (b). shows the 3D-views and (c). (line 208); Table (2) vs. Table. 3 (line 294); Fig. 4 vs. Fig 4 ( line 306); Fig 4 i.e„ Group1-Group. Is it „skull stripping“ (line 227) or striping?; Capitalization in Description of Table 3; link to open source machine learning framework PyTorch (line 260); there is information missing of the Table number „shown in Table. ??.“ (line 301); what means data dataset (line 348)?

 

MAJOR THINGS

I assume that the authors have chosen the article type themselves. Therefore, a classical article with introduction, material/methods and results was expected. The structuring is somewhat unfortunate, because in section 2 again general facts about Alzheimer's disease are addressed, which belong to the introduction. Section 2 is like a mini review of related literature. The information that PET is a potent imaging tool to detect Alzheimers is missing. Furthermore, once the image characterists of an AD patient need to be clarified at the beginning since this is the basis for all the AI/learning based processes

In 3.1 the authors write „The proposed practical multi-class classification network, denoted as PMCAD-Net, is designed to classify various stages of Alzheimer’s Disease.“ It seems that there is some information missing. For a better understanding the explanation like: classify stages of AD based on xy MRI data, T2 weighted(?) , visualization of plaques (?), etc.

Overall it is often not clear what is the own work, what is work of other reserachers, what is a step what was performed (Method), what is the result of the own performed work and what discussion of the own work in comparison to existing literature? Thus, the overall structure need revision. After re-structuring and making the red-line clear I would recommend a re-submission.

Comments for author File: Comments.pdf

Author Response

 Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

1. The abstract needs to add numerical results for the accuracy of the classifier and also for evaluation measures.
2. The introduction is reorganized so that it does not contain the workflow of the proposal and is placed in a separate section and explained in detail.
3. Figure 1 and figure 2 are placed in section 3.
4. The construction of deep learning needs clarification and visualization
5. Add a separate section for discussion and comparison of results.
6. Use more evaluation performance such as ROC, AUC, MCC.
7. References are updated for the years 2021 and 2022, and you can use

https://doi.org/10.1016/j.compbiomed.2021.104606

https://doi.org/10.3390/s22031184

 

Author Response

 Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Dear Authors,

please check again the spacing in the authors list between commas and * is not linked to an author. Or do you want this? You used Alzheimer´s disease through the manuscript in line 347 you wrote "Alzheimer’s Disease". Figure 3 is still not very clear to me. In the black box (c) this is not  Axial, Sagittal and coronal views. Maybe this can be separated better?

Regards

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Accepte in present form

Author Response

Thank you very much for your support of our work.