Shoulder joint and bones

Hi !

Still some questions, always on the shoulder model i’m trying to build…

What is the nature of the gleno-humeral joint in Anybody, is it a
ball-in-a-socket ?
If yes, could we define another one like a ball on a curved plane and
is it much complicated ?

The aim is to have a model in which we could suppress one or more
muscle to imulate a tear, and see the theorical consequences compared
to experimental. So, with a ball-in-a-socket joint, the only
difference will be on elevation and global force given to the joint by
muscles but with another joint, we could see other displacement like
migration of bones.

I’m trying to use the arm 3D folder in the BRep. It’s quite ok when
I’m using it with the spine and torso but it’s no use for me.
What is the best and easier way to keep only the arm and shoulder,
with the scapula fixed ?

Thanks a lot

Yoann

Hi Yoann

> What is the nature of the gleno-humeral joint in Anybody, is it a
> ball-in-a-socket ?

Yes the glenohumeral joint is made as a spherical joint.

> If yes, could we define another one like a ball on a curved plane and
> is it much complicated ?

In principle it is possible to define a contact between two ellipsoids using the
elements which are in the system today. I am not sure if i will recomend
trying to do this for the GH joint. The joint should be made as a mechanism and
it will require a good deal of work to implement it for the GH joint, but if you
are interested let us know.

> The aim is to have a model in which we could suppress one or more
> muscle to imulate a tear, and see the theorical consequences compared
> to experimental. So, with a ball-in-a-socket joint, the only
> difference will be on elevation and global force given to the joint by
> muscles but with another joint, we could see other displacement like
> migration of bones.

Ok there is another option than the one mentioned above, it is possible to
switch off the normal reactions in the GH joint. I guess what you would like to
have is one sided reactions in the joint. Instead of the normal reactions in
the joint you can add artificial muscles to the joint that will take up only the
compression forces. This can be done using AnyGeneralMuscles which can be used
to provide compression forces. If you insert such a muscle acting from humerus
towards scapula, then you are some of the way, towards the joint you describe.
This joint however will not be able to take up any sideways reactions, only
compression in the direction of the longitudinal axis of humerus. The strength
of this muscle should be so high that it’s activity will be much lower than the
normal muscles activity envelope. If you can live without the sideways
reactions it is not so big a task to create this model, but it will require some
playing around with the model before it will work.

> I’m trying to use the arm 3D folder in the BRep. It’s quite ok when
> I’m using it with the spine and torso but it’s no use for me.
> What is the best and easier way to keep only the arm and shoulder,
> with the scapula fixed ?

The best way is to keep the spine, but you may use a model with no muscles on
the spine such as “BRep/Aalborg/Bodymodels/SpineNoMusclesRightArm/BodyModel.any”

In principle you may use the arm without the spine, but in this case you will
need to construct manually several segments which are used for muscles
attachments points. This is because muscles such as latissimus dorsi attaches to
the spine in the lumbar region.

Best regards

AnyBody Technology Support

hello !

Excuse me but i don’t think your artificial muscle is what i’d like to
have.
I am not looking for the compression forces, or maybe later.
What i would like to observe is the consequence of a tear, for example
a tear of the rotator cuff muscles.
If in our model the joint is only spherical, the force used to elevate
the shoulder will have to be higher in other muscles, responsible for
abduction. But what about the way of displacement of the humeral head?
Will it go up ? What are the limits of muscles forces? Hom high could
be the arm elevation in this case ?
This kind of question is what we’re working on and we would like to
make an answer by a parallel between our experiments and the Anybody
Shoulder Model.
Is it a little less obscure ?

Thanks

Sincerely,

Yoann

Hi Yoann
> Excuse me but i don’t think your artificial muscle is what
> i’d like to
> have.
> I am not looking for the compression forces, or maybe later.
> What i would like to observe is the consequence of a
> tear, for example
> a tear of the rotator cuff muscles.
> If in our model the joint is only spherical, the force
> used to elevate
> the shoulder will have to be higher in other muscles,
> responsible for
> abduction. But what about the way of displacement of
> the humeral head?
> Will it go up ?

I am sorry. I think I misunderstood the question in the first place.
What you want to do here is basically in the realm of forward
dynamics because you want to change a force and see the consequence
in terms of displacement. Inverse dynamics as AnyBody does it goes in
the other direction: From displacement to computation of the muscle
forces. However, there may be two ways to address your problem with
AnyBody:

  1. If you disable the torn muscle in the model, then you can see how
    the loads will be redistributed to the other muscles presuming an
    unchanged movement, and the system will tell you either by higher
    muscle activities or by failure to fulfil equilibrium how serious the
    lack of the torn muscle is. You can subsequently impose a different
    humeral head position on the model and investigate if this will
    reduce the load on the altered system.

  2. With the new version coming out soon you can set up a posture
    optimization problem so that the system automatically determined the
    position of the humerus that minimized the load on the altered system.

>Hom high could
> be the arm elevation in this case ?

Do you mean how much abduction could the model do in the absence of a
rotator cuff muscle? This could be answered by method 1 above.

Please notice that you may need to redefine the joint reaction forces
in terms of artificial reaction muscles as described in a prevous
reply to obtain the correct limits on the joint stability. I believe
the Dutch shoulder group (van der Helm et al) has done some work on
this issue. Prhaps you know more about it. In any case it is
something that has to be approached with care and it is probably
worthy of a scientific investigation in its own right. If you are
interested, perhaps we can establish a scientific cooperation with
the AnyBody Research Project for this matter. Let us know if you are
interested.

> This kind of question is what we’re working on and we would like to
> make an answer by a parallel between our experiments and the Anybody
> Shoulder Model.
> Is it a little less obscure ?

Yes. It is a quite usual problem to attack by means of this type of
computational model. Inverse dynamics sometimes requires a reverse
point-of-view, but it is very useful because it enables you to
capture the complexity of the musculoskeletal system. Forward
dynamics would be computationally prohibitive for this type of model.

As indicated above, the definition of the unstable joint may require
some additional thought, and you probably have to explain a bit more
about the physiology of the problem. As I understand the gleno-
humeral joint, the humeral head rests in the glenoid and forms
basically a ball and socket joint with it. The socket is rather
shallow, and if the humeral head slides with respect to the glenoid,
then the joint will dislocate. Are you aiming to model this
dislocation or are you aiming to model a smaller movement within the
normal limits of the GH joint?

Best regards,
AnyBody Support.

Hi !

Thank you, i’ve seen my explanation are quite better now…

> 1. If you disable the torn muscle in the model, then you can see how
> the loads will be redistributed to the other muscles presuming an
> unchanged movement, and the system will tell you either by higher
> muscle activities or by failure to fulfil equilibrium how serious the
> lack of the torn muscle is. You can subsequently impose a different
> humeral head position on the model and investigate if this will
> reduce the load on the altered system.
>
> 2. With the new version coming out soon you can set up a posture
> optimization problem so that the system automatically determined the
> position of the humerus that minimized the load on the altered system.

Your ideas seem to fit what i’m looking for.

> Do you mean how much abduction could the model do in the absence of
> a rotator cuff muscle?

Yes, that was what i meant.

> If you are
> interested, perhaps we can establish a scientific cooperation with
> the AnyBody Research Project for this matter. Let us know if you are
> interested.

That could be a great thing if i’m not able to make my model alone.
I keep it in mind, thank you.

> Are you aiming to model this
> dislocation or are you aiming to model a smaller movement within the
> normal limits of the GH joint?

The aim is to have a small movement but if forces have to dislocate,
let’s see it. It’s a great option.

For the model i wish :
> The best way is to keep the spine, but you may use a model with no
> muscles on the spine such as
> “BRep/Aalborg/Bodymodels/SpineNoMusclesRightArm/BodyModel.any”

Yes, i’ve tried it but there’s a lot of stuff i really don’t need. As
you know, Anybody is quite demandeous when used with muscles and the
less we have, the more we have a quick model.
Let me explain : i need the arm, scapula and clavicula. Muscles needed
are those making relation between arm and bones cited (supraspinatus,
infraspinatus, subscapularis, teres minor, teres major and deltoidus).
Muscles playing with torso or spine are really not use for the moment.
One last thing , the scapula has to be fixed (not fixed with an
artificial node but fixed not to move).

So i would like to try with only the Arm3D to have a “light” model.
I think this comes to complete a little…

Thank you !

Yoann

Hi Yoann

I have written some comments in the bottom of the message

> Thank you, i’ve seen my explanation are quite better now…
>
> > 1. If you disable the torn muscle in the model, then you can see how
> > the loads will be redistributed to the other muscles presuming an
> > unchanged movement, and the system will tell you either by higher
> > muscle activities or by failure to fulfil equilibrium how serious the
> > lack of the torn muscle is. You can subsequently impose a different
> > humeral head position on the model and investigate if this will
> > reduce the load on the altered system.
> >
> > 2. With the new version coming out soon you can set up a posture
> > optimization problem so that the system automatically determined the
> > position of the humerus that minimized the load on the altered system.
>
> Your ideas seem to fit what i’m looking for.
>
> > Do you mean how much abduction could the model do in the absence of
> > a rotator cuff muscle?
>
> Yes, that was what i meant.
>
> > If you are
> > interested, perhaps we can establish a scientific cooperation with
> > the AnyBody Research Project for this matter. Let us know if you are
> > interested.
>
> That could be a great thing if i’m not able to make my model alone.
> I keep it in mind, thank you.
>
> > Are you aiming to model this
> > dislocation or are you aiming to model a smaller movement within the
> > normal limits of the GH joint?
>
> The aim is to have a small movement but if forces have to dislocate,
> let’s see it. It’s a great option.
>
>
> For the model i wish :
> > The best way is to keep the spine, but you may use a model with no
> > muscles on the spine such as
> > “BRep/Aalborg/Bodymodels/SpineNoMusclesRightArm/BodyModel.any”
>
> Yes, i’ve tried it but there’s a lot of stuff i really don’t need. As
> you know, Anybody is quite demandeous when used with muscles and the
> less we have, the more we have a quick model.
> Let me explain : i need the arm, scapula and clavicula. Muscles needed
> are those making relation between arm and bones cited (supraspinatus,
> infraspinatus, subscapularis, teres minor, teres major and deltoidus).
> Muscles playing with torso or spine are really not use for the moment.
> One last thing , the scapula has to be fixed (not fixed with an
> artificial node but fixed not to move).
>
> So i would like to try with only the Arm3D to have a “light” model.
> I think this comes to complete a little…

We do not have a “light arm model” but you can use the one we have for making
one. The procedure is to add AnyReac forces on the joints AC and SC this can be
done in the Brep/Arm3D/Joints. This will have the effect of locking scapula and
clavicle. Then you need to outcomment the muscles spanning these joints this is
done in the Muscles.any file. You can do the same for the lower arm. Before
doing this i would recommend to make a copy of the current Arm3D model.

Best regards

AnyBody Support