[REVIEW]: Brainchop: In-browser MRI volumetric segmentation and rendering

[editorialbot]

Submitting author: !--author-handle-->@Mmasoud1<!--end-author-handle-- (Mohamed Masoud) Repository: https://github.com/neuroneural/brainchop Branch with paper.md (empty if default branch): Version: v2.1.0 Editor: !--editor-->@samhforbes<!--end-editor-- Reviewers: @mwegrzyn, @62442katieb, @richford Archive: 10.5281/zenodo.7735848

Status

Status badge code:

HTML: <a href="https://joss.theoj.org/papers/92e800c8ac3b0c431357adaad09cb677"><img src="https://joss.theoj.org/papers/92e800c8ac3b0c431357adaad09cb677/status.svg"></a>
Markdown: [![status](https://joss.theoj.org/papers/92e800c8ac3b0c431357adaad09cb677/status.svg)](https://joss.theoj.org/papers/92e800c8ac3b0c431357adaad09cb677)

Reviewers and authors:

Please avoid lengthy details of difficulties in the review thread. Instead, please create a new issue in the target repository and link to those issues (especially acceptance-blockers) by leaving comments in the review thread below. (For completists: if the target issue tracker is also on GitHub, linking the review thread in the issue or vice versa will create corresponding breadcrumb trails in the link target.)

Reviewer instructions & questions

@mwegrzyn & @62442katieb & @richford, your review will be checklist based. Each of you will have a separate checklist that you should update when carrying out your review. First of all you need to run this command in a separate comment to create the checklist:

@editorialbot generate my checklist

The reviewer guidelines are available here: https://joss.readthedocs.io/en/latest/reviewer_guidelines.html. Any questions/concerns please let @samhforbes know.

✨ Please start on your review when you are able, and be sure to complete your review in the next six weeks, at the very latest ✨

Checklists

📝 Checklist for @mwegrzyn

📝 Checklist for @richford

📝 Checklist for @62442katieb

[editorialbot]

Hello humans, I'm @editorialbot, a robot that can help you with some common editorial tasks.

For a list of things I can do to help you, just type:

@editorialbot commands

For example, to regenerate the paper pdf after making changes in the paper's md or bib files, type:

@editorialbot generate pdf

[editorialbot]

Software report:

github.com/AlDanial/cloc v 1.88  T=15.51 s (46.4 files/s, 35098.9 lines/s)
-------------------------------------------------------------------------------
Language                     files          blank        comment           code
-------------------------------------------------------------------------------
JavaScript                     525          70322          35297         199995
CSS                             48            103           2476         159243
TypeScript                       4            487             12          48783
LESS                            81           1472            610          14630
SVG                              8              0              2           3056
JSON                            20              1              0           1978
Sass                            14             34             34           1768
HTML                             3            477             30           1712
Markdown                         8            487              0            868
Python                           4             41             19            268
TeX                              1             13              0             68
YAML                             1              1              4             22
Jupyter Notebook                 1              0             51             21
reStructuredText                 1              7             11              7
-------------------------------------------------------------------------------
SUM:                           719          73445          38546         432419
-------------------------------------------------------------------------------

gitinspector failed to run statistical information for the repository

[editorialbot]

Wordcount for paper.md is 851

[editorialbot]

Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- 10.1109/IJCNN.2017.7966333 is OK
- 10.48550/arXiv.1511.07122 is OK
- 10.48550/arXiv.1901.05350 is OK
- 10.5281/zenodo.6430433 is OK
- 10.1016/j.neuroimage.2020.117012 is OK

MISSING DOIs

- None

INVALID DOIs

- None

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[samhforbes]

Hi @mwegrzyn, @62442katieb, @richford thanks for agreeing to review. This is our review thread. Please generate your checklist with the above, and raise any issues about the software in the issues tab of the Brainchop repository directly, linking back to here, Any wider issues you'd like to discuss please flag here of course, and let me know if there's anything I can help with. Thanks again!

[mwegrzyn]

Review checklist for @mwegrzyn

Conflict of interest

• [x] I confirm that I have read the JOSS conflict of interest (COI) policy and that: I have no COIs with reviewing this work or that any perceived COIs have been waived by JOSS for the purpose of this review.

Code of Conduct

• [x] I confirm that I read and will adhere to the JOSS code of conduct.

General checks

• [x] Repository: Is the source code for this software available at the https://github.com/neuroneural/brainchop?
• [x] License: Does the repository contain a plain-text LICENSE file with the contents of an OSI approved software license?
• [x] Contribution and authorship: Has the submitting author (@Mmasoud1) made major contributions to the software? Does the full list of paper authors seem appropriate and complete?
• [x] Substantial scholarly effort: Does this submission meet the scope eligibility described in the JOSS guidelines
• [x] Data sharing: If the paper contains original data, data are accessible to the reviewers. If the paper contains no original data, please check this item.
• [x] Reproducibility: If the paper contains original results, results are entirely reproducible by reviewers. If the paper contains no original results, please check this item.
• [x] Human and animal research: If the paper contains original data research on humans subjects or animals, does it comply with JOSS's human participants research policy and/or animal research policy? If the paper contains no such data, please check this item.

Functionality

• [x] Installation: Does installation proceed as outlined in the documentation?
• [x] Functionality: Have the functional claims of the software been confirmed?
• [x] Performance: If there are any performance claims of the software, have they been confirmed? (If there are no claims, please check off this item.)

Documentation

• [x] A statement of need: Do the authors clearly state what problems the software is designed to solve and who the target audience is?
• [x] Installation instructions: Is there a clearly-stated list of dependencies? Ideally these should be handled with an automated package management solution.
• [x] Example usage: Do the authors include examples of how to use the software (ideally to solve real-world analysis problems).
• [x] Functionality documentation: Is the core functionality of the software documented to a satisfactory level (e.g., API method documentation)?
• [x] Automated tests: Are there automated tests or manual steps described so that the functionality of the software can be verified?
• [x] Community guidelines: Are there clear guidelines for third parties wishing to 1) Contribute to the software 2) Report issues or problems with the software 3) Seek support

Software paper

• [x] Summary: Has a clear description of the high-level functionality and purpose of the software for a diverse, non-specialist audience been provided?
• [x] A statement of need: Does the paper have a section titled 'Statement of need' that clearly states what problems the software is designed to solve, who the target audience is, and its relation to other work?
• [x] State of the field: Do the authors describe how this software compares to other commonly-used packages?
• [x] Quality of writing: Is the paper well written (i.e., it does not require editing for structure, language, or writing quality)?
• [x] References: Is the list of references complete, and is everything cited appropriately that should be cited (e.g., papers, datasets, software)? Do references in the text use the proper citation syntax?

[Mmasoud1]

Dear @samhforbes , thank you for your service, and thanks for respected reviewers for taking the time to review this paper. Would you please point us when to expect your initial decision?

[samhforbes]

Hi @Mmasoud1 Remember that reviews take time to complete and the initial decision will follow that, as I think Arfon has already discussed with you.

@mwegrzyn, @62442katieb, @richford how are things going? Let me know if I can help.

[Mmasoud1]

Dear @samhforbes , according to your last inquiry, I am afraid our respected reviewers are not available.

[62442katieb]

Hello, apologies for the delay, I was out of office. Taking a look at this later today!

[Mmasoud1]

@62442katieb appreciate your time, thank you!

[richford]

Hi @samhforbes, my apologies for the delayed response. I'll take a look at this tomorrow.

[richford]

Review checklist for @richford

Conflict of interest

• [x] I confirm that I have read the JOSS conflict of interest (COI) policy and that: I have no COIs with reviewing this work or that any perceived COIs have been waived by JOSS for the purpose of this review.

Code of Conduct

• [x] I confirm that I read and will adhere to the JOSS code of conduct.

General checks

• [x] Repository: Is the source code for this software available at the https://github.com/neuroneural/brainchop?
• [x] License: Does the repository contain a plain-text LICENSE file with the contents of an OSI approved software license?
• [x] Contribution and authorship: Has the submitting author (@Mmasoud1) made major contributions to the software? Does the full list of paper authors seem appropriate and complete?
• [x] Substantial scholarly effort: Does this submission meet the scope eligibility described in the JOSS guidelines
• [x] Data sharing: If the paper contains original data, data are accessible to the reviewers. If the paper contains no original data, please check this item.
• [x] Reproducibility: If the paper contains original results, results are entirely reproducible by reviewers. If the paper contains no original results, please check this item.
• [x] Human and animal research: If the paper contains original data research on humans subjects or animals, does it comply with JOSS's human participants research policy and/or animal research policy? If the paper contains no such data, please check this item.

Functionality

• [x] Installation: Does installation proceed as outlined in the documentation?
• [x] Functionality: Have the functional claims of the software been confirmed?
• [x] Performance: If there are any performance claims of the software, have they been confirmed? (If there are no claims, please check off this item.)

Documentation

• [x] A statement of need: Do the authors clearly state what problems the software is designed to solve and who the target audience is?
• [x] Installation instructions: Is there a clearly-stated list of dependencies? Ideally these should be handled with an automated package management solution.
• [x] Example usage: Do the authors include examples of how to use the software (ideally to solve real-world analysis problems).
• [x] Functionality documentation: Is the core functionality of the software documented to a satisfactory level (e.g., API method documentation)?
• [x] Automated tests: Are there automated tests or manual steps described so that the functionality of the software can be verified?
• [x] Community guidelines: Are there clear guidelines for third parties wishing to 1) Contribute to the software 2) Report issues or problems with the software 3) Seek support

Software paper

• [x] Summary: Has a clear description of the high-level functionality and purpose of the software for a diverse, non-specialist audience been provided?
• [x] A statement of need: Does the paper have a section titled 'Statement of need' that clearly states what problems the software is designed to solve, who the target audience is, and its relation to other work?
• [x] State of the field: Do the authors describe how this software compares to other commonly-used packages?
• [x] Quality of writing: Is the paper well written (i.e., it does not require editing for structure, language, or writing quality)?
• [x] References: Is the list of references complete, and is everything cited appropriately that should be cited (e.g., papers, datasets, software)? Do references in the text use the proper citation syntax?

[mwegrzyn]

Dear @samhforbes, thank you for the reminder. I will finish reviewing at the end of this week (March 5). Best, Martin

[62442katieb]

Review checklist for @62442katieb

Conflict of interest

• [x] I confirm that I have read the JOSS conflict of interest (COI) policy and that: I have no COIs with reviewing this work or that any perceived COIs have been waived by JOSS for the purpose of this review.

Code of Conduct

• [x] I confirm that I read and will adhere to the JOSS code of conduct.

General checks

• [x] Repository: Is the source code for this software available at the https://github.com/neuroneural/brainchop?
• [x] License: Does the repository contain a plain-text LICENSE file with the contents of an OSI approved software license?
• [x] Contribution and authorship: Has the submitting author (@Mmasoud1) made major contributions to the software? Does the full list of paper authors seem appropriate and complete?
• [x] Substantial scholarly effort: Does this submission meet the scope eligibility described in the JOSS guidelines
• [x] Data sharing: If the paper contains original data, data are accessible to the reviewers. If the paper contains no original data, please check this item.
• [x] Reproducibility: If the paper contains original results, results are entirely reproducible by reviewers. If the paper contains no original results, please check this item.
• [x] Human and animal research: If the paper contains original data research on humans subjects or animals, does it comply with JOSS's human participants research policy and/or animal research policy? If the paper contains no such data, please check this item.

Functionality

• [x] Installation: Does installation proceed as outlined in the documentation?
• [x] Functionality: Have the functional claims of the software been confirmed?
• [x] Performance: If there are any performance claims of the software, have they been confirmed? (If there are no claims, please check off this item.)

Documentation

• [x] A statement of need: Do the authors clearly state what problems the software is designed to solve and who the target audience is?
• [x] Installation instructions: Is there a clearly-stated list of dependencies? Ideally these should be handled with an automated package management solution.
• [x] Example usage: Do the authors include examples of how to use the software (ideally to solve real-world analysis problems).
• [x] Functionality documentation: Is the core functionality of the software documented to a satisfactory level (e.g., API method documentation)?
• [x] Automated tests: Are there automated tests or manual steps described so that the functionality of the software can be verified?
• [x] Community guidelines: Are there clear guidelines for third parties wishing to 1) Contribute to the software 2) Report issues or problems with the software 3) Seek support

Software paper

• [x] Summary: Has a clear description of the high-level functionality and purpose of the software for a diverse, non-specialist audience been provided?
• [x] A statement of need: Does the paper have a section titled 'Statement of need' that clearly states what problems the software is designed to solve, who the target audience is, and its relation to other work?
• [x] State of the field: Do the authors describe how this software compares to other commonly-used packages?
• [x] Quality of writing: Is the paper well written (i.e., it does not require editing for structure, language, or writing quality)?
• [x] References: Is the list of references complete, and is everything cited appropriately that should be cited (e.g., papers, datasets, software)? Do references in the text use the proper citation syntax?

[62442katieb]

Went through the paper and checklist and I have a few notes: The docs have a lot of information, with snippets of example code throughout, and the README + paper have a nice schematic of how Brainchop works, but it might be helpful to have a start-to-finish example solving a real-world problem.

The paper mentions that being browser-based avoids privacy issues. As it seems the authors intend this tool to be used by clinicians, it would be helpful to more explicitly spell out in the paper and docs how this avoids privacy issues (e.g., nothing is saved to an external server, the tool can be installed and run locally, etc.) as not all clinicians are familiar with the implications of browser-based vs. server-side processing.

I'm testing out Brainchop in Firefox (v109.0.1) on an M1 MacBook Pro and many of the segmentation options won't run, with error messages such as: This model needs texture size of minimum 16384, while current browser supports only 8192 and This option needs a dedicated graphics card. The system requirements are noted in the wiki, but it would be helpful to present those limitations more clearly in the paper and/or by denoting which segmentation options have specific system requirements on the browser client. Some options suggest I should use Safari instead, which would be helpful information to have up front (esp. as Figure 1 suggests Chrome and Firefox are the browsers of choice). As the authors intend this tool to increase accessibility, making these requirements/limitations more clear in the tool and in the paper would be helpful for users, especially those with limited access to computational resources.

The paper would benefit from explicitly comparing Brainchop to existing automated segmentation tools, in addressing the state of the field. The authors make it clear that the main benefit of this tool over others is its lack of installation requirements and ease of use, but that point would be more compelling if the authors compared it to the installation/usability of the tools it aims to supersede.

Very cool tool, overall! I think a few tweaks and it's good to go!

[richford]

@emdupre, I went through the paper and source code. Overall, it is very impressive work. I opened a few issues on the source repository related to UX and the paper itself. They are also linked in this thread.

I cannot verify the performance claims in the documentation because I don't have that exact hardware. But they are very plausible given the performance I saw on my M1 MacBook.

@62442katieb also brought up some good point that I'd like to echo. Namely,

• The paper should say a bit more about the "state of the field," even if that means discussing "offline" segmentation tools.
• The documentation would benefit from a start to finish example (including preprocessing) using a NIfTI file from an openly available repository.

I enjoyed reading the paper and test driving the tool. With a few minor modifications, I will give this submission an enthusiastic approval!

[Mmasoud1]

Many thanks to our respected reviewers for their valuable feedback. We work actively to address their comments and respond to the opening issues.

[samhforbes]

Thanks both for your reviews! @Mmasoud1 do update us here when these issues have been addressed! @mwegrzyn please let us know your thoughts as well when you've had a chance to look.

[mwegrzyn]

Dear @Mmasoud1,

thank you for your work on brainchop. The following are some comments I have regarding the package and the manuscript:

use of brainchop website

• [ ] the website is well-designed. However, depending on screen size, not all elements are visible. For example, only part of the papaya viewers are visible on a smaller screen
• [ ] regarding hardware and software requirements, the manuscript states that the tool might be particularly beneficial in rural areas and developing countries. However, it seems to me that the tool demands a lot from the user’s machine. In fact, I was not able to use it on various desktop and laptop computers (both Linux and Windows), and only could use it partly running Chrome (but not Firefox). Many of the models were inaccessible to me (“This model may need dedicated graphics card”). Performing “Subvolume GWM (failsafe)” took minutes to complete. Due to this, I did not check the “Performance: If there are any performance claims of the software, have they been confirmed?” part of the JOSS checklist, as I think that the requirements of the package should be communicated in a more explicit way

validation of brain extraction tools

• [ ] I was wondering how well the different brain extraction algorithms perform and how they were pre-trained. More information on this would be helpful. In the acknowledgements I read “The authors would like to thank Kevin Wang and Alex Fedorov for discussions and pre-trained Meshnet models.”. Could you elaborate further? Thank you

• [ ] I was not able to use all Segmentation Options of the tool, due to my insufficient hardware and/or software setup, but when using “Extract the Brain (FAST)” on one of my own images (https://openneuro.org/datasets/ds001419/versions/1.0.1/file-display/sub-01:anat:sub-01_T1w.nii.gz) I was not satisfied with the result. It contains both false-positives (a substantial part of the skull labelled as brain) and false-negatives (a substantial part of orbitofrontal brain areas not labelled as brain). This is shown below. I am used to imperfect brain extraction, for example from the FSL brain extraction tools, but there I can manually tweak the parameters to get better results. How can these issues be addressed in brainchop? Can I change certain settings to improve results?

• [ ] How will the pre-trained models perform for brains with atrophy, lesions or for developing brains?

manuscript

• [ ] I found the manuscript well-written and helpful. The motivation is clear to me. I would like to know however, how the tool compares to other software packages, for example MricroGL, SPM, FSL, FreeSurfer, ANTs, and/or fMRIprep pipelines). Due to this, I did not check the “State of the field: Do the authors describe how this software compares to other commonly-used packages?” part of the JOSS checklist
• [ ] line 58: “Papaya (Papaya dev. team, 2019) viewers is used” → this should read “Papaya (Papaya dev. team, 2019) viewer is used”
• [ ] line 61: “Three.js (Cabello & et, 2010) ” → this should read “Three.js (Cabello et al., 2010) ” and in the references, “& et, al.” should be corrected

documentation in wiki

• [ ] contribution guidelines and ways to get help are well-explained. However, I was not able to find much information on the different Models for the Segmentation Options. Due to this, I did not check the “Functionality documentation: Is the core functionality of the software documented to a satisfactory level” box on the JOSS checklist. Could you elaborate on the different models and the function to “Load tfjs Models” (see screenshot below)?
• [ ] The mocha testing does show an error for #normalizeVolumeData(): https://neuroneural.github.io/brainchop/test/runner.html

I hope you find these comments helpful. Please let me know if you have any questions regarding the points I raised. Thank you for your time and consideration.

Best, Martin Wegrzyn

[Mmasoud1]

@mwegrzyn Thank you for your time and valuable comments, we will submit our full responses shortly, hopefully by tomorrow after addressing all the comments, but for now please double check with Mocha comment again since the error doesn't appear with me nor my collaborator and also it wasn't reported by respected reviewers. This is what I have after running the test now :

I may suggest for now to make sure please that WebGL context is not lost and working before retesting for Mocha or try another browser please if you had the same issue again. Thank you.

[mwegrzyn]

Dear @Mmasoud1,

thank you for your message. I am looking forward to you revision. I followed the link in the wiki to https://neuroneural.github.io/brainchop/test/runner.html and get the following message (on Debian with Firefox and on Windows with Firefox/Chrome/Edge):

I hope this helps.

Best, Martin Wegrzyn

[samhforbes]

Hi @mwegrzyn thanks for flagging these, and I know @Mmasoud1 will be keen to address the points you raise. It sounds like a couple of these issues should be raised as issues on the software repo since other users may run into the same sort of problems.

[Mmasoud1]

Dear @62442katieb , thank you for your time and valuable comments.

Example usage: We added a showcase example to the doc.

Data privacy statement: In the statement of needs section, we added the suggested phrase in line 28, also we added similar phrase to the wiki home page and brainchop welcome screen.

Performance & Functionality : Despite the high diversity of H/W and S/W resources that exist on the users' side, we did our best to give them feasible recommendations. We did our best to make the end-user be able to run at least one model successfully, but, since S/W and H/W are renewable, and their capabilities improve periodically in terms of cost, speed and computational power, we expect more declining in the arising resources issue over time.

A verified list of S/W and H/W that run successfully with each brainchop model is provided with this link. This link also is mentioned at UI and Wiki to give end-users more information about each model tested resources. We also modified the UI can popup warning messages for the need of a dedicated graphics card or minimum texture size to point end-user for browser resources window showing end-user browser resources and the aforementioned link of verified S/W and H/W for each model.

Aso, we updated the System Requirements page on wiki to reflect more information and recommendations for our end-user.

On our end, GeForce GTX 1050 Ti was able to run all the models. In general, internal or built-in graphics chip (e.g. HD, UHD) can run mostly light models such as: Full Brain GWM (light), Compute Brain Mask (FAST), or Extract the Brain (FAST), but they are not able to run larger models such as Full Brain GWM (large), FS aparc+aseg Atlas 104 (failsafe) or Cortical Atlas 50. We also noted that some users, despite the GPUs that they may have, didn't properly configure it to be used with their browsers ending that their browser can only detect built-in graphics chips on the motherboard, and they didn't notice that. That is why we added the Browser Resources icon on the taskbar to point them out for this issue.

State of the field: In the statement of needs section, we added the state of the field paragraph in lines 29-34.

Brainchop as the first volumetric brain segmentation tool that runs completely on the user side provides a strong proof of concept of the browser capability to process volumetric neuroimaging and satisfies also several important needs such as : Accuracy due to volumetric segmentation, privacy and data residency due to user side processing, easiness-to-use or usability due to zero installation.

Data: we shared our loading sample and also mentioned our showcase open dataset. The end-user can also Reproduce our segmentation results in no time and with few clicks if they meet System Requirements or follow the verified H/W and S/W configuration which is flexible list of multiple H/W and S/W options for each model.

Functionality documentation : In addition to the information about brainchop architecture, pre-trained and inference pipelines, etc, we also added more details about segmentation models , model browsing options, and System Requirements.

Best,

[Mmasoud1]

Dear @richford, thank you for your time, valuable comments and opening bug issues.

Add author contributions : Contribution section added lines 74-78.

Add to acknowledgments : Acknowledgments section modified.

Sidebar does not scroll on home page issue 18 : We added a scroll bar to the left mini forms of tool options e951fbd

https://user-images.githubusercontent.com/36082112/223699793-757f0f08-3151-4c44-826d-968c8037a9bd.mp4

Segmentation model info popup cuts off content issue 19 : We pushed Fixes for the bug 9f5cae7

State of the field: In the statement of needs section, we added the state of the field paragraph in lines 29-34.

Example usage: We added a show case example to the doc.

Best,

[Mmasoud1]

Dear @mwegrzyn, thank you for your time, and valuable comments.

Use of brainchop website:

I tried to mediate this stylish conflict between Webix and Papaya by adjusting the padding. This makes MRI planes visible for screen width in range > 1350 and subject also for more future improvements to fit smaller sizes.

Performance : For Verified S/W and H/W please check this link

In our case, GeForce GTX 1050 Ti was able to run all the models.

Internal or built-in graphics chip (e.g. HD, UHD) can run mostly light models such as: Full Brain GWM (light), Compute Brain Mask (FAST), or Extract the Brain (FAST), but in general they are not able to run larger models such as Full Brain GWM (large), FS aparc+aseg Atlas 104 (failsafe) or Cortical Atlas 50.

We tried to make our end-user able to run at least one model successfully. Please note, we are performing volumetric segmentation for the first time in the browser, this is normally done on desktop applications. Taking into consideration also please the improvement in H/W and S/W (i.e. browser) in terms of speed and memory each year, the resources issue will decline with time. In our case we use GeForce GTX 1050 Ti which is an average GPU compared to state of the art GPUs and still it performs very well with all models. However, we noted also that some users, despite the GPUs that they may have, didn't properly configure it to be used with their browsers ending that their browser can only detect built-in graphics chips on the motherboard, and they didn't notice that. That is why we added the Browser Resources icon on the taskbar to point them out for this issue.

Validation of brain extraction tools

I added two deeper Meshnet models ( 11 filters per layer ) for better brain extraction and masking accuracy : Extract the Brain (High Acc) and Compute Brain Mask (High Acc). Both models are tested with Apple M1 and GeForce GTX 980.

https://user-images.githubusercontent.com/36082112/223703125-0e1d6d89-8a3f-4b2f-9cd3-b93c88cf7d78.mp4

In the UI, we provided also end-user with a browser resources window showing end-user browser resources and a link to verified S/W and H/W for each model.

For more info about Meshnet training please refer to this previous work of our lab here. That also referenced in our wiki. We also plan to integrate training pipeline with brainchop repo for advanced users in version 3.0.0.

Although fast brain extraction model (i.e. Extract the Brain (FAST) ) has less accuracy than the aforementioned model, it is more suitable to run on machines with Internal or built-in graphics chip (e.g. HD, UHD) that normally unable to run 11 filters per layer.

Manuscript:

State of the field: In the statement of needs section, we added the state of the field paragraph in lines 29-34.

Typos: Suggested typos corrections are applied for Papaya and three.js. Thank you!.

Documentation in wiki:

Load tfjs Models: Description section has been added to wiki, also we added to the wiki a brief functionality description of each model currently in use with the UI. These is also accessible within the UI by clicking on the info icon next to the model list.

Mocha test: We fixed it taking into consideration that issue may arise due to slightly different floating point calculations between machines.

Best,

[richford]

@Mmasoud1, thanks for your revisions.

@emdupre, the authors have responded to all of the issues that I raised and I recommend this submission for publication!

[Mmasoud1]

Thank you @richford.

[mwegrzyn]

Dear @Mmasoud1, thank you for your response and the thoughtful revision!

Dear @samhforbes, I have now checked all the boxes in my reviewer checklist and recommend the submission for publication.

Best, Martin Wegrzyn

[Mmasoud1]

Thank you @mwegrzyn

[Mmasoud1]

@62442katieb, do you have any question or comments on our corrections? I will be away for traveling starting next week and I hope you can kindly have a chance to review them before that if possible please.

[62442katieb]

@62442katieb, do you have any question or comments on our corrections? I will be away for traveling starting next week and I hope you can kindly have a chance to review them before that if possible please.

Looks good to me!

[Mmasoud1]

Thank you @62442katieb!

[Mmasoud1]

Thanks both for your reviews! @Mmasoud1 do update us here when these issues have been addressed!

@samhforbes, we are done!, we appreciate the dedicated time and helpful comments by all respected reviewers that increased our work quality. Let me know please if I can help with anything next. Thank you!

[Mmasoud1]

@editorialbot check references

[editorialbot]

Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- 10.1109/IJCNN.2017.7966333 is OK
- 10.48550/arXiv.1511.07122 is OK
- 10.48550/arXiv.1901.05350 is OK
- 10.5281/zenodo.6430433 is OK
- 10.1016/j.neuroimage.2020.117012 is OK
- 10.1087/20150211 is OK
- 10.1002/hbm.460020402 is OK

MISSING DOIs

- None

INVALID DOIs

- https://doi.org/10.1006/nimg.1998.0395 is INVALID because of 'https://doi.org/' prefix

[Mmasoud1]

@editorialbot check references

[editorialbot]

Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- 10.1109/IJCNN.2017.7966333 is OK
- 10.48550/arXiv.1511.07122 is OK
- 10.48550/arXiv.1901.05350 is OK
- 10.5281/zenodo.6430433 is OK
- 10.1016/j.neuroimage.2020.117012 is OK
- 10.1087/20150211 is OK
- 10.1002/hbm.460020402 is OK
- 10.1006/nimg.1998.0395 is OK

MISSING DOIs

- None

INVALID DOIs

- None

[Mmasoud1]

@editorialbot generate pdf

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[Mmasoud1]

@editorialbot check references

[editorialbot]

Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- 10.1109/IJCNN.2017.7966333 is OK
- 10.48550/arXiv.1511.07122 is OK
- 10.48550/arXiv.1901.05350 is OK
- 10.5281/zenodo.6430433 is OK
- 10.1016/j.neuroimage.2020.117012 is OK
- 10.1087/20150211 is OK
- 10.1002/hbm.460020402 is OK
- 10.1006/nimg.1998.0395 is OK

MISSING DOIs

- None

INVALID DOIs

- None

[Mmasoud1]

@editorialbot generate pdf

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[samhforbes]

Thanks @richford for the review you completed, that's really helpful.

[samhforbes]

Thanks also @mwegrzyn and @62442katieb . @62442katieb Can you please confirm you're happy with the changes made and how they addressed the issues you raised by completing your reviewer checklist?

[Mmasoud1]

Dear @samhforbes , please kindly note that @62442katieb had completed her reviewer checklist. I see all of them checked on my end.

[samhforbes]

Hi @Mmasoud1 thanks for flagging this - I'm not sure what's up with my browser, so I apologise.

Great! Looks like everyone is really pleased, and having had a little play with it, so am I. Can you please check the version number is accurate, confirm the author list, and then archive it somewhere with a stable DOI (Zenodo, or figshare for example) - making sure the authors and title match that of the paper. Then could you please post the DOI here.

[Mmasoud1]

Hi @samhforbes , thank you for the suggested changes.

I reviewed the release version number with the latest changes from the review, and checked the paper author list (including ORCIDs and affiliations). Also, I archived the latest version on Zenodo with same title, author list(including ORCIDs) and MIT license :

Version: 2.1.0 DOI: 10.5281/zenodo.7735848