A Simple Model to Predict Loads within Muscle-Tendon Complexes of the Shoulder during Fast Motions

Round 1

Reviewer 1 Report

I would like to gratulate the authors for their fine work. I think this work adds to the understanding of the loading of the shoulder joint and specific muscles during different activities in regards to injury. I have no comments or recommendations and suggest publishing this manuscript in its current form.

Author Response

Dear Reviewer,

thank you for your motiviating words!

Sincerely Yours

Falk Mörl

Reviewer 2 Report

The work is interesting. I suggests adding references to Anatomy Trains or a fascia network before publishing. This may affect your results. Please add a paragraph – conclusions. In the references, please add pages in the following positions: 14 / 25 / 26

"Can the thicker, more outer-positioned, synergistic MTC, such as the deltoid muscle, provide support or protection for the smaller, more proximal positioned supraspinatus muscle, if there are high forces acting on the arm? "  I think that is an interesting question.
It shows possible patterns of muscle overload. In my opinion, the contribution is great. The work in is clearly written. The conclusions are supported by results. The work is well written. 

Author Response

Dear reviewer,

thank you for your motivating words regarding our manuscript. 

We now give conclusions in the last paragraph (line 337f).

Considering the references, number 14 has pages (3007–3014). Reference 25 is a master thesis. All it is about is the test and biomechanical modeling of an hand-to-tool interaction, from which we extracted the parameters of our contact model. Number 26 is a book, so too, there is no reference for pages. 

Sincerely Yours

Falk Mörl

Reviewer 3 Report

 

Modeling of the shoulder motion is a complex but highly necessary. The authors used a Hill-type muscle mechanical model and a Hatze-type muscle activation model to predict shoulder motions, forces. The authors set out to address the question: ‘Can the thicker, more outer-positioned, synergistic MTC, such as the deltoid muscle, provide support or protection for the smaller, more proximal positioned supraspinatus muscle, if there are high forces acting on the arm?’ I have only a few comments I ask the authors to address:

Comments

1. The authors need to more critically and in a detailed manner motivate their use of Hill model and Hatze model. In particular, what are the benefits and drawbacks using these models versus OpenSim type of modeling. In particular, it is unclear how compatible are the Hill and Hatze model temporally and spatially to predict shoulder motion.

 

2. It is unclear why the authors used males and females. Such a sample increases anatomical variation and there could be hypotheses generated that the predictions are gender-specific.

 

3. The authors need to provide photos or schematic drawings of the experimental tasks, as it is not easy to fully comprehend what the experimental situations were.

 

4. Perhaps I missed it, but I cannot see the quantification of error / difference between experimental and simulated data in joint position (shoulder kinematics) and forces. The authors need to state explicitly the magnitude of these errors.

Author Response

Dear reviewer,

thank you for your comments, which actually improved our manuscript. Please find our response below your comments. We also document the line numbers, where changes have been made.

1. The authors need to more critically and in a detailed manner motivate their use of Hill model and Hatze model. In particular, what are the benefits and drawbacks using these models versus OpenSim type of modeling. In particular, it is unclear how compatible are the Hill and Hatze model temporally and spatially to predict shoulder motion.

### answer
Regarding the motivation to use the Hill-type approach, we added a sentence in the introduction around line 70:
"Hill-type muscle models allow to simulate a lot of muscle functions explored experimentally,..."
A drawback is the calculation of transverse forces which for Hill-type muscle models is just at a starting point (not implemented in the model here). Lateral muscle forces may play an important role at the shoulder joint, for example the hat-like deltoideus muscle may produce compression on the joint. We added a short paragraph in the discussion regarding this topic (line 308f). 
###

2. It is unclear why the authors used males and females. Such a sample increases anatomical variation and there could be hypotheses generated that the predictions are gender-specific.

### answer
In a very detailed anatomic modelling approach like the propagation of CT-scan parameters to individual model anatomy, the difference between different subjects and gender-specific properties may play a role. Here, the simple shoulder model is just a hinge joint driven by muscle models and/or external forces. No individual anatomic parameters were implemented, it is just one example for all. Despite this, the model predicts all investigated parameters in the range of the data form the subjects extracted experimentally. 
###

3. The authors need to provide photos or schematic drawings of the experimental tasks, as it is not easy to fully comprehend what the experimental situations were.

### answer
done, see our new figure 1

4. Perhaps I missed it, but I cannot see the quantification of error / difference between experimental and simulated data in joint position (shoulder kinematics) and forces. The authors need to state explicitly the magnitude of these errors.

### answer
If a model is built to replicate a given motion exactly, differences between model and experimental data are very important. This is not the case in comparison to our experimental data, the subjects produced shoulder angle and contact force value on a broad range of variability. Data of a single experimental task will fit to the corresponding model data to more than 95%. This would be true only for few tasks of a single subject. By changing the parameters of the model (stimulation, stimulation duration, force capacity of the muscle model etc.) such fit would be possible for each experimental task. Such sensitivity analysis may answer your question 2, but is no topic of this study. Hence, calculating differences between model and experimental data has no benefit in this context. 

Sincerely yours

Falk Mörl