Shoulder complex

Hi guys,

I now work with the Bergmann shoulder model of AnyBody 4.0 and AMMR v1.0.

I see that the scapula and the clavicula are joined by some ligaments and that the scapula is driven by contact node at the thorax. Using this configuration, is this possible to simulate the whole shoulder complex just by moving the clavicula ? Have you tried to determine if the scapulohumeral rhythm obtained by moving only the clavicula correlates with values found in literature ?

Also, in the SIN Fourier function used to simulate the GH abduction, I see that the signal varies from 0 to 45. Is there a mean to make it varies from 15 to 45 instead ?

Thanks

Pierre

Hi Pierre,

I am pretty sure that you won’t get a correct scapular motion by simply moving the clavicula. This is because of the conoid ligament that is driven to a fixed value. I think it would not be fixed in a full shoulder complex motion.
So you will need some data about the scapula position as well (at least one dof, you can maybe keep the 2 contact points on the rib cage).

For the fourier driver it is possible to make it start from an other position than 0 deg. The following piece of code will add this initial angle to de driver:

AnyKinMeasureLinComb CombMeasure =
{
Coef = {{1}};
Const = {<InitialAngle>*pi/180} ;
AnyKinMeasureOrg <name> =
{
MeasureOrganizer = {<direction>};
AnyKinMeasure &ref = <driven measure>;
};
};

AnyKinEqFourierDriver <name> = {
Type = Sin;
Freq = ;
A = {{0, <amplitude>}*pi/180};
B = {{0,0}};
AnyKinMeasure &CombMeasure = .CombMeasure;
Reaction.Type = {Off};
};

Best regards, Sylvain.

Hi Sylvain,

If I understand it right, I should add an upward rotation to the scapula, in order to get closer to the real movement of the shoulder complex. What should I do when I get a KKT singular matrix error ?

I have another question about the HumerusMuscleGeometry. In this part of the model, you define the ‘art_tub_minus_tub_majus’ node in which you also define a sphere surface. What is the goal of that surface ? Is this for muscle wrapping ? Because when I simulate a permanent contact between the acromion and the humerus, this sphere creates an error of penetration in some parts of the deltoid muscle. What are the consequences of having this error on the recruitment process ? Can I still consider the results or I should get rid of the error before ?

Thank you

Pierre

Hi Pierre,

Well it is hard to say from here what motion you should exactly add to the scapula. I think you can just try to move the clavicula only first and see what it looks like. As i understand you work with the BergmannGH model and perform only abduction, so it migth not be that bad at the end.
I think a combination of abduction and flexion would be more critical and this approach probably not valid.
It is up to you to decide if the aproximation is good enough.

The KKT singular matrix error is general solver error message, it can be cased by wrong boundary conditions, missing reactions, muscle configuration unable to carry some dof, etc…

The sphere on the ‘art_tub_minus_tub_majus’ node is a wrapping surface for the deltoid. The penetration of a via point in the surface will affect he muscle path, so it is to be avoided in general. However if it is very small then you can live with it, but if the penetration is far into the surface then the muscle path will be wrong and so migth the results. In the last case you have to get rid of the penetration.

Best regards, Sylvain.

Hi Sylvain,

I tried both approaches for the scapular kinematics and it gave me similar muscular forces at the GH joint, but different scapular motions. I should maybe check the forces in the scapular muscles to see which one gives the most representative results. Also, the values that I use to drive the shoulder are perhaps not entirely valid for my case, because they represent an average of several patients, each with its own morphology.

I asked you the question about the KKT matrix because I had this error while feeding the scapular motion with a ‘AnyKinEqInterPolDriver’ (with 22 points). Then I looked at the corresponding acceleration curves and found that they were not equal to zero. To avoid having these acceleration, I determined a polynomial function of sufficient degree to represent each movement and the accelerations are now closer to zero and the KKT matrix error is gone.

For the ‘art_tub_minus_tub_majus’ surface sphere, I want to keep it, because it’s a good representation of the muscle pathway and thickness. This is much more important when the massive tear is simulated in the model. Please have a look at the attached file. It’s a simple 2D representation of what I need in my massive tear model.

In the normal shoulder model, I see that the wrapping surface doesn’t touches the muscles, as they pass through the via points that are more outside. Knowing that, will I get the same results if I get rid of this surface wrapping in the muscles definitions ? Would it be a better idea to just decrease its radius or to move up the origin of the muscle to avoid the penetration?

Thank you very much.

Pierre

Hi Pierre,

I think it’s ok to use the avarage data of several patient. You could use the very specific data of one patient if you also used the coresponding specific anthropometric dimension but i understand this isn’t the case so it’s ok.

High acceleration would definitly case problem. It usually end up in overloaded muscles but it can also creates this singular KKT matrix error. So it is agood idea to filter the data or as you did fit an idealised polynomial curve.

For the wrapping surface i think you should not remove but just modify sligthly either the motion or the muscle geometry to avoid the penetration.

Best regards, Sylvain.

Hi Sylvain,

If I change the motion of the humeral head, the contact will be lost. I had another idea, but I’m not sure if it’s valid. Basically, I would put the “art_tub_minus_tub_majus” wrapping surface on the scapula gh node instead of on the humerus gh node, to avoid the surface to move with the humerus. I would place it as close as possible to where it was on the humerus, when the latter is in contact with the acromion. I will have to move down the surface to avoid the penetration, but it should be more representative, when compared to a removal of the wrapping surface.

However, this would only be valid with a fixed radius wrapping sphere. If I want to change the shape of the humeral head (e.g. use an ellipsoid), I will have to find another issue. Please check the attached file to understand what I mean when I talk about changing the shape of the humeral head for an ellipsoid.

Thank you

Pierre

Hi Pierre

I have not seen the attached file, but i do not think it sounds as a good idea to change the segment from humerus to scapula. It may solve the issue of the wrapping line but it will also change the force flow so the contact force will be moved from scapula to humerus, i am unsure about the effect of this.

Best regards
Søren