Digital and Intelligent Rehabilitation Technologies in Stroke and Neurological Disorders: A Systematic Review of Artificial Intelligence, Virtual Reality, Gamification, and Emerging Therapeutic Platforms in Neurorehabilitation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper conducts a systematic review of digital neurorehabilitation technologies, including artificial intelligence, virtual reality, gamification, and telerehabilitation, establishing for the first time a cross-technology integration framework. The study comprehensively evaluates multiple types of evidence, such as randomized controlled trials, simulation modeling, and clinical feasibility studies, encompassing diverse neurological populations. It elucidates the mechanisms through which adaptive algorithms and gamification strategies enhance patient engagement, thereby promoting multidimensional functional recovery in motor, balance, and cognitive domains. By examining the differences between home-based and clinical settings and addressing accessibility challenges in resource-limited regions, this study provides both theoretical support and practical guidance for the clinical translation of intelligent rehabilitation technologies.

The shortcomings of this article and the recommendations are as follows:

1. Most of the included studies lack long-term follow-up assessments, which limits the evaluation of the sustained rehabilitation effects of the technologies. Additionally, the absence of real-world application data hinders the assessment of their feasibility for clinical translation.
2. While the review covers multiple technologies and synthesizes evidence from diverse perspectives, it lacks in-depth analysis and dedicated discussion on the synergistic mechanisms and interaction effects between these different technologies.
3. As a cross-technology integrative review, the analysis of existing studies remains insufficient in addressing methodological limitations, sample constraints, and transparency of technological implementation. Moreover, the core theoretical contribution of this review—such as the proposed digital rehabilitation technology ecosystem framework—has not been explicitly articulated in the discussion and conclusion sections.
4. Current rehabilitation platforms lack specifically adapted functionalities for patients with cognitive, visual, or perceptual impairments, and an inclusive assessment framework is absent, thereby limiting the generalizability of the findings.
5. The algorithmic logic and decision-making processes of the AI techniques employed in the study are not illustrated through visual means, while certain simulation-based research relies solely on synthetic data without accompanying clinical trial.

This paper contains some spelling ,grammatical and formatting mistakes that need to be corrected:

1. In Figure 3, last line: The punctuation marks are missing at the end.
2. In Discussion, last line: The punctuation marks are missing at the end.
3. Page 6, In Discussion, Para 2, Line 7: The word 'ARAT' stands for "Action Research Arm Test," which is a standard academic abbreviation and should be fully spelled out the first time it appears in the text.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This article presents a comprehensive and up-to-date systematic review covering AI, VR, gamification, and telerehabilitation technologies in stroke and other neurological diseases. The methodology, compliant with PRISMA 2020, the use of the PROSPERO registry, risk-of-bias (RoB 2, ROBINS-I), and GRADE enhances the methodological strength of the study. Findings are supported by relatively well-structured tables and diagrams. The integration of clinical and simulation studies provides a holistic contribution to the literature.

The comprehensive search strategy (2010–2025) and transparent reporting, the balanced synthesis of clinical, home-based, and simulation/AI model studies, the appropriate and explicit use of risk-of-bias and GRADE, and the combined evaluation of clinical outcomes, usability, and personalization dimensions are the main strengths/essential aspects of the article.

However, in my opinion, there are still some minor areas that could be improved. I ask the following questions:

- It is emphasized that meta-analysis could not be performed due to heterogeneity; More concrete suggestions for future development could be added for subgroup analyses. This would actually strengthen the main focus.

-The topic of AI transparency/explainability could be enhanced with more concrete examples in the discussion.

-Suggestions for shortcomings in long-term follow-up and economic evaluation should be presented in a more structured manner.

-Heterogeneity and the level of evidence should be discussed more clearly in the limitations section.

-Standard terms and abbreviations in the tables should be made consistent.

-If possible, the literature discussion on long-term effects should be expanded.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors

Although the literature review is done really well to this paper, the paper lacks of originality.

It is hard to distinguish what is the contribution of the paper, there are just comparisons to other papers. 

Even though the authors seam more from the health or medical sector, being an engineering journal I would suggest to add some mathematics to the paper, maybe some charts, plots, maybe find and engineer who can help improving your paper.

The papers seams from a magazine, not  scientific paper.

I have also found 29% AI. Almost the full introduction, discussion and conclusion is written by AI.  Please, correct this. Like this this, the paper cannot be published.

There are some other minor issues like:

1. Improve the quality of figures, e.g. Figure 1. It seams a Print Screen, it could be done vectorially also to not loose quality when zooming.
2. Reference citations seam to be not in order.
3. Before refence citations it should be a space, e.g. text [1].
4. The keyword should be separated with a semicolon (;), limited to 5, and chosen from the following list: https://www.ieee.org/content/dam/ieee-org/ieee/web/org/pubs/ieee-taxonomy.pdf

Best Regards

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have fully revised the paper. I think this paper can be accepted.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

accept

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

It is still hard to find to contribution of the paper.

The mathematics and chart part is still really weak.

The AI is now not detected, but it still worries me that this paper was written by AI and paraphrased, so no true contribution of authors.

The minor issues:

1. Pictures seam still Print Screens.
2. OK
3. OK
4. OK

Best Regards

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

1. It is still hard to find to contribution of the paper.

It still feels that there is no original research in the paper, there was done some reviews of other papers only.

2. The mathematics and chart part is still really weak.

In order for me to accept this change I need to see some equations, not "pseudo-formulas".

3. The AI is now not detected, but it still worries me that this paper was written by AI and paraphrased, so no true contribution of authors.

Let say I believe you, but your comment made me laugh a little bit, sorry. Let me quote: "We were also surprised by the earlier report indicating possible AI detection". It is simple if YOU write, there is no AI detection, if the AI writes, than it is AI detection.

4. Pictures seam still Print Screens.

From my point of view it is not acceptable to give high resolution images after acceptance, the paper needs to be published as it is, so in order to be accepted, high quality images are needed. This way it can be tested that the images are really created by the authors and not taken from the internet or generated by AI. If you made the image, than you can recreate them in higher resolution.

Other minor issues:

In the first line of affiliation there is an underlined dot after "Qatar".

Also there are paging issues, too many blank spaces.

Best Regards

Author Response

Please see the attachment

Author Response File: Author Response.docx