MMD-MSD: A Multimodal Multisensory Dataset in Support of Research and Technology Development for Musculoskeletal Disorders

Round 1

Reviewer 1 Report

The manuscripts presents a way to automatically detect improper sitting postures using gold-standard and user-labelled data. The overall structure of the work is good, and is of high relevance in automated decision-support and personalized healthcare. 

Some suggested improvements:

1. In the introduction, please add more literature review on the use of machine learning in similar problems, why those methods were used, problems faced and solutions found. This is relevant to the manuscript, and without this review, the reader is left with many fundamental questions. Example, in line 50, the authors can follow-up with reviewing the literature more in-depth. 

2.  In the Methods section, it will help the readers if a brief summary of the ML methods are provided, with some equations, their fundamental algorithmic nature. 

3. More details on the 10-fold cross-validation can be provided to help readers understand how to design and handle datasets in this work, when combined with ML methods.

Questions:

1. How can we understand the temporal effects  of bad posture? If we envision real-time decision-support, can we rank bad postures that should be corrected first, based on the ranking? As an example, top ranks are given to postures that adversely affect the human faster than the lower ranked postures. 

2. How can we design more personalized algorithms using population level dataset? Can the authors discuss this in the Conclusion section, maybe referring to some literature, and discussing how this study can be modified toward a more personalized approach?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

A very interesting and interestingly designed study with great potential with good use and development of the material. The manuscript presented is a well described database for "future use". However, the journal Algorithms is for presenting new solutions, tools, mathematical analysis, allogrithms and not for presenting databases. Perhaps it is worth focusing on analyzing some aspects and describing them in more detail for this journal presenting just an algorithm, or if the essence of the publication is to present a database then perhaps it is worth considering publishing in other journals like Big Data, BioMed, BioMedInformatics, BioTech, BioSensors? Because in the current version the manuscript does not fully fit into the aims and scopes of the journal.
What I miss in the conclusion is the potential use of the data/solution, i.e. with this study what can we introduce, what will it translate into or what will it serve?

For references, practically all of the articles cited are from the last 5 years.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript develops a new multimodal multisensory dataset (MMD) via simultaneous recordings of diverse physiological signals relevant to sitting posture. As subjects are also carrying out cognitive tasks while these signals are recorded, task performance and cognitive load are additionally measured. The aim of the dataset is utility in applications that aid in preventing musculoskeletal disorders (MSD), such as those in the realm of machine learning.

The paper is well written, sufficiently detailed, and clearly organized. The experimental apparatus and scientific claims appear sound. I have just several minor suggestions for improvement:

-For completeness, it would be useful to include potential limitations of the developed dataset in the conclusion section.

-The abbreviation MMD is used in the introduction before it is defined in the main text (outside the abstract).

-There are some small grammatical errors and typos that should be correct.

Ex: In the section Method, we -> In the Materials and Methods Section, we

In Section Results, we  -> In Results Section, we

English requires minor editing.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

This study aims to utilize computer aided technology along with multisensory data for primary research on Musculoskeletal Disorder. The manuscript is well written and the idea given here could possibly be used for early stage diagnosis of various diseases. In my opinion the manuscript should be published. However,  the results are not appropriately presented to gain the full benefit of this study. Therefore a major revision is required as mentioned below

1) Include some of the emerging applications of this methodology in healthcare.

2) Some recent trends in medical imaging and disease diagnosis should also be included in section-1

https://link.springer.com/article/10.1186/s12891-017-1428-1

https://www.mdpi.com/2079-6374/12/12/1181

3) Section-3 requires graphical data illustration for the acquired datasets mentioned in Table-1. Valuable insights and key features must be highlighted

4) The results must be comprehensively discussed. 

5) Figure-1 must be appropriately scaled. The axes must be labeled properly.

6) The significance of multisensory data from  must be qualitatively and quantitatively analyzed.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I accept the author's corrections and response.