All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Congratulations on the acceptance of your manuscript.
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
Authors have addressed the points raised in the previous review well. No further review of the article is required and the article is suitable for acceptance.
NA
NA
Hi,
Kindly address all the comments from expert reviewers appropriately.
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
The manuscript's strength lies in its clear objective and the application of deep learning to Alzheimer's disease detection. However, it could be enhanced by providing a more comprehensive literature review, particularly comparing the proposed methods with recent advancements in the field. For instance, a study by Islam and Zhang (2017) demonstrated a novel deep learning model for Alzheimer’s Disease detection using Brain MRI Data, which could serve as a benchmark for comparison. Additionally, the manuscript should ensure data availability in line with PeerJ's standards for reproducibility, as emphasized by Mamun et al. (2022), who provided datasets and code for replication.
Suggested Improvements:
Include a more detailed comparison with recent studies such as Islam and Zhang (2017) and Mamun et al. (2022).
Ensure the availability of datasets and code, or provide a clear explanation for any restrictions.
While the experimental design is robust, it could be improved by detailing the rationale behind the choice of deep learning models and parameters. The study by Sekhar and Jana (2023) implemented Alzheimer’s disease detection using U-Net and EfficientNet, discussing the choice and tuning of these models. Moreover, the manuscript should include a comprehensive statistical analysis, as performed by Mamun et al. (2022), to support the validity of the findings.
Suggested Improvements:
Elaborate on the selection and optimization of deep learning models, drawing comparisons with studies like Sekhar and Jana (2023).
Provide a detailed statistical analysis, including assumptions and robustness checks.
The findings are promising, but the manuscript should address the potential for overfitting and the model's real-world applicability. Comparative analysis with other models, such as the one by Mamun et al. (2022), which achieved an accuracy of 97.60%, would contextualize the results. Additionally, the manuscript could benefit from discussing the generalizability of the findings, as suggested by Sekhar and Jana (2023), who emphasized the importance of model performance in clinical settings.
Suggested Improvements:
Discuss the generalizability and clinical applicability of the models, referencing studies like Sekhar and Jana (2023).
Include a comparative analysis with other recent models, such as those discussed by Mamun et al. (2022).
The manuscript is a valuable addition to the field, yet it would benefit from a discussion on ethical considerations and broader implications for healthcare integration, as highlighted by Mamun et al. (2022). The potential impact on patient outcomes and healthcare systems should be considered, drawing insights from the ensemble deep learning framework proposed by Sekhar and Jana (2023), which emphasized low parameter count and better training time, important for real-world applications.
In conclusion, while the manuscript presents significant contributions, addressing these points would align it more closely with PeerJ's standards and the current state of research in the field.
The article demonstrates good potential but necessitates significant revisions. The figures and tables within the manuscript have been meticulously examined; however, the code provided in the Github repository's notebooks lacks reproducibility.
While the article adeptly elucidates the utilization of three deep learning ML models, it falls short in justifying their application specifically for Alzheimer's Disease (AD) prediction. For instance, the rationale behind not employing ML models akin to XGBoost or LightGBM remains unexplored. Additionally, while McNemar’s test is widely recognized, employing methods such as combinatorial fusion analysis could have offered a more appropriate means to compare the performance of ML models.
Further insight into the identified targets for AD, as outlined in (https://www.nature.com/articles/s42003-022-03068-7), would have strengthened the article. Line 217-219 requires a paraphrase regarding the data source, while specifying the proportions in each set (testing and training/validation) at Line 256 is crucial. Multiple sections throughout the article necessitate language refinement, prompting the authors to conduct additional checks for clarity and coherence.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.