Image Quality Improvement and Memory-Saving in a Permanent-Magnet-Array-Based MRI System

Round 1

Reviewer 1 Report

Abstract should be changed. The current abstract is only introduction to the problem. However, an abstract should include objectives, methods, results, and conclusion. Theoretical evaluation and results were derived but not mentioned in the abstract.

A little revision is required for the introduction:

line 30: The relationship between the bulkiness and accessibility does not make sense. 

line 32: explain more about the "point of care and bulkiness"

line 60: spatial encoding magnetic (SEM) field

In Fig. 1, labels (a,b,c) are missed.

Author Response

1. We have shortened the part about problem description and enhance the illustration on method in abstract to make it clearer for readers to gain the information about problem, objectives, method, result and conclusion of the paper. The part on objectives, method, result and conclusion are highlighted in red in the revised manuscript.
2. ‘Bulky’ in line 30 (in the old manuscript) has been removed.
3. ‘Bulkiness’ is replaced by ‘it requires a large space to site’ to have a more detailed description on bulkiness. Please refer to line 28 in the revised manuscript.
4. Thanks! We have fixed it. Please refer to Line 58 in the revised manuscript.
5. Thanks! We have fixed it.

Author Response File: Author Response.pdf

Reviewer 2 Report

In this paper, the authors investigate an approaches for image quality improvement in a PMA based MRI system, and propose an effective approach to obtain both a memory saving reconstruction and a significant improvement in image quality simultaneously in such a system.

The work is well written and organized. The relevance and the effectiveness of this paper is demonstrated by its detailed technical description and a satisfactory experimental section.

Author Response

Thanks for your review!

Reviewer 3 Report

The authors provide a method to acquire MRI images with lower memory requirements than conventional imaging. Although the work is novel, its value is limited, mainly because, to my opinion, the memory usage is not a real problem for low field MRI systems. The authors state in line 382 that their approach reduces the memory from 14GBto 7GB. Normal consumers laptops already have 64GB of memory, so they can easily the apply conventional methods.

Other remarks:

• The article is very long, especially the introduction, I think it can easily be shortend, which makes is easier to read.
• Figure 4 shows images of a brain phantom. The images look however so realistic that either the details of the phantom need to be given, so others can also make such a good phantom, or it is actually a normal MRI scan, in which case this should be clearly mentioned in the text.
• An example of the applicability of this method using real low-field MRI data would significantly improve this manuscript.

Author Response

1. The increase from 7GB to 14 GB is for the case studied in the paper. However, in reality, denser sampling points are needed and the required memory can easily go beyond the memory of a normal laptop (e.g. 64 GB). Therefore, image reconstruction that requires memory that can be handled by a portable computer is crucial. This further explanation has been added to the revised manuscript. Please find it in line 382-386 of the revised manuscript.
2. A permanent-magnet-array-based MRI system using inhomogeneous field to encode the signals for imaging is still a new concept in the MRI society. It is different from the traditional approach where a homogeneous field plus linear gradient fields are used. Different methods have to be used to analyze the performance of such a system, for example, local k-space. For the above reason, we would like to include the details in the introduction to show clearly the differences as well as the corresponding documentations in the literature. Therefore, we prefer to keep the introduction as it is. Minor grammatic modifications have been made in the revised manuscript.
3. We have added some illustration on the caption of Figure 4. Please find in on page 8 of the revised manuscript.
4. Thank you for the suggestion. We will do it in the future to test our system and algorithms and present the data in future publications.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Although the authors have succesfully incorporated most of the suggestions, I still find the value of this manuscript rather limited, especially since the method has not been validated with real data acquired with a low field MR-system (comment number 4 of the previous review).

Author Response

Dear Reviewer 3, 

We understand your concern. However, for the theoretical investigation reported in this manuscript, it is based on realistic setups of a permanent magnet array (PMA)-based system yet ahead of it is physical implementations. The current version of the manuscript carries its values for the understanding of such a system which can be used to guide the improvements of the performances of the system. It will be of interest to the audience who is interested in a PMA-based low-field MRI system.

You have had a very good suggestion for us for this work. We will carry out the experiments and report the corresponding results in the near future. As for now, the experiment is still under preparation and it will take a while to finish. Hope you will understand our situation.

For the above reasons, we prefer to keep the manuscript as it is.

Thank you for your time!

Best, 

the authors