Segment axes definitions

I am working with the right arm model and need to know how the
coordinate systems on each segment (specifically the scapula and
humerus) is defined. I am trying to define the same coordinate systems
in a finite element model.

It looks like the coordinate system for the humerus has its origin at
the center of the humeral head, is this true? How were each of the
axes defined? With respect to anatomical locations on the humerus, or
another method? Also, the scapula coordinate system seems to be
floating over the acromion; was there a methodical way that this was
set up?

Is there any documentation defining these coordinate systems? I know
of the ISB protocol, but do not know whether the Dutch shoulder model
and the ISB standards are coincident.

Thanks,

Chris

Hi Chris

Please see posting #1445 in this group, this is also about
coordinate systems in the shoulder and there are references to
several other postings on this topic.

Please ask again if this do not answer your questions

Best regards
Søren, AnyBody Support

— In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@…> wrote:
>
> I am working with the right arm model and need to know how the
> coordinate systems on each segment (specifically the scapula and
> humerus) is defined. I am trying to define the same coordinate
systems
> in a finite element model.
>
> It looks like the coordinate system for the humerus has its origin
at
> the center of the humeral head, is this true? How were each of the
> axes defined? With respect to anatomical locations on the humerus,
or
> another method? Also, the scapula coordinate system seems to be
> floating over the acromion; was there a methodical way that this
was
> set up?
>
> Is there any documentation defining these coordinate systems? I
know
> of the ISB protocol, but do not know whether the Dutch shoulder
model
> and the ISB standards are coincident.
>
> Thanks,
>
> Chris
>

Søren,

Thanks for the response. I have a couple additional questions, just to
make sure things are clear.

I understand that the shoulder model is ‘driven’ by angles of the
humerus with respect to the thorax, as opposed to driving distal
segments in a sequential order (thorax to clavicle to scapula to
humerus, for example). After playing with the model (comparing the
scapula position in 0 degrees abduction with 90 degrees abduction) and
reading the code, the scapula position and orientation doesn’t seem to
be affected by the motion of the humerus. Is this correct? How does
the elliptical gliding plane of the scapula come into the model then?
How does this allow for changes in muscle lengths which span the
thorax and scapula then?

Also, the referenced information regarding the coordinate system
information
(http://www.fbw.vu.nl/research/Lijn_A4/shoulder/isg/proposal/protocol.htm)
states that the origin of the scapular coordinate system is located at
the AC joint. In the AnyBody model though, this coordinate system
seems to be floating over approximately where the acromion meets the
main body of the scapula, which is posterior to the AC joint. I am
curious if this difference affects the locations of the muscle origin
and insertion points that are defined in the scapular coordinate
system, such that those points that were originally defined in the
Dutch shoulder model may be slightly translated in the current AnyBody
model.

Thanks again,

Chris

— In anyscript@yahoogroups.com, “AnyBody Support” <support@…> wrote:
>
> Hi Chris
>
> Please see posting #1445 in this group, this is also about
> coordinate systems in the shoulder and there are references to
> several other postings on this topic.
>
> Please ask again if this do not answer your questions
>
> Best regards
> Søren, AnyBody Support

Furthermore, concerning the stationary scapula…

Wouldn’t this affect the muscle lengths (and thus force and moment
predictions) of muscles which originate on the scapula and insert on
the humerus?

Chris

— In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@…> wrote:
>
> Søren,
>
> Thanks for the response. I have a couple additional questions, just to
> make sure things are clear.
>
> I understand that the shoulder model is ‘driven’ by angles of the
> humerus with respect to the thorax, as opposed to driving distal
> segments in a sequential order (thorax to clavicle to scapula to
> humerus, for example). After playing with the model (comparing the
> scapula position in 0 degrees abduction with 90 degrees abduction) and
> reading the code, the scapula position and orientation doesn’t seem to
> be affected by the motion of the humerus. Is this correct? How does
> the elliptical gliding plane of the scapula come into the model then?
> How does this allow for changes in muscle lengths which span the
> thorax and scapula then?
>
> Also, the referenced information regarding the coordinate system
> information
>
(http://www.fbw.vu.nl/research/Lijn_A4/shoulder/isg/proposal/protocol.htm)
> states that the origin of the scapular coordinate system is located at
> the AC joint. In the AnyBody model though, this coordinate system
> seems to be floating over approximately where the acromion meets the
> main body of the scapula, which is posterior to the AC joint. I am
> curious if this difference affects the locations of the muscle origin
> and insertion points that are defined in the scapular coordinate
> system, such that those points that were originally defined in the
> Dutch shoulder model may be slightly translated in the current AnyBody
> model.
>
> Thanks again,
>
> Chris
>
> — In anyscript@yahoogroups.com, “AnyBody Support” <support@> wrote:
> >
> > Hi Chris
> >
> > Please see posting #1445 in this group, this is also about
> > coordinate systems in the shoulder and there are references to
> > several other postings on this topic.
> >
> > Please ask again if this do not answer your questions
> >
> > Best regards
> > Søren, AnyBody Support
>

Chris,

It is the weekend here, and you will probably receive an answer that
more directly addresses your questions early next week. However, I
think I should say a few words about driving models in general and
the shoulder model in particular for the benefit of all the members.

In general, AnyBody is neither a forward nor an inverse kinematics
system. AnyBody allows you to drive the model by any combination of
forward and inverse type drivers.

You are right that the “natural” way to drive a model is to start
proximaly and progress along the extremities, and to some extent this
is the way the standing model and the free posture model in the
repository work. But this is a modeling issue and not a system issue.

Now about the shoulder model: It does indeed have all the natural
degrees of freedom of a shoulder, i.e. the scapular-thoracic movement
and the gleno-humeral movement. But the repository models do not
provide access to the scapular-thoracic movement through the
mannequin file. This does not mean that the scapula is welded to the
thorax in a kinetic sense - only that the model specifies that this
movement is zero; the model will still recruit the necessary muscle
forces to keep the scapula stationary.

To control the scapular-thoracic movement you must go to the
JointsAndDrivers file of your model. A bit down you will finde the
sterno-clavicular joint driver. To obtain scapular-thoracic movement
you must either control the angle of the clavicle wrt. the sternum as
it is done in the existing JointsAndDrivers setup, or remove one or
more of these existing drivers and replace them by other driver
strategies. A popular way to drive the scapula is to create a linear
measure between the thorax and the acromion and create two drivers on
this measure driving elavation and protraction/retraction.

I hope this clarifies some of the general issues.

Best regards,
John

— In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@…> wrote:
>
> Søren,
>
> Thanks for the response. I have a couple additional questions, just
to
> make sure things are clear.
>
> I understand that the shoulder model is ‘driven’ by angles of the
> humerus with respect to the thorax, as opposed to driving distal
> segments in a sequential order (thorax to clavicle to scapula to
> humerus, for example). After playing with the model (comparing the
> scapula position in 0 degrees abduction with 90 degrees abduction)
and
> reading the code, the scapula position and orientation doesn’t seem
to
> be affected by the motion of the humerus. Is this correct? How does
> the elliptical gliding plane of the scapula come into the model
then?
> How does this allow for changes in muscle lengths which span the
> thorax and scapula then?
>
> Also, the referenced information regarding the coordinate system
> information
>
(http://www.fbw.vu.nl/research/Lijn_A4/shoulder/isg/proposal/protocol.
htm)
> states that the origin of the scapular coordinate system is located
at
> the AC joint. In the AnyBody model though, this coordinate system
> seems to be floating over approximately where the acromion meets the
> main body of the scapula, which is posterior to the AC joint. I am
> curious if this difference affects the locations of the muscle
origin
> and insertion points that are defined in the scapular coordinate
> system, such that those points that were originally defined in the
> Dutch shoulder model may be slightly translated in the current
AnyBody
> model.
>
> Thanks again,
>
> Chris
>
> — In anyscript@yahoogroups.com, “AnyBody Support” <support@>
wrote:
> >
> > Hi Chris
> >
> > Please see posting #1445 in this group, this is also about
> > coordinate systems in the shoulder and there are references to
> > several other postings on this topic.
> >
> > Please ask again if this do not answer your questions
> >
> > Best regards
> > Søren, AnyBody Support
>

John,

Thanks for the reply…especially on a weekend.

Also, thanks for the direction on how to implement scapulo-thoracic
movement…I will look into this further. Why is this not included in
the model though? This seems like a significant deviation from
reality and it’s ramifications on other model predictions could be
very much affected.

Thanks again,

Chris

— In anyscript@yahoogroups.com, “AnyBody Support” <support@…>
wrote:
>
> Chris,
>
> It is the weekend here, and you will probably receive an answer
that
> more directly addresses your questions early next week. However, I
> think I should say a few words about driving models in general and
> the shoulder model in particular for the benefit of all the
members.
>
> In general, AnyBody is neither a forward nor an inverse kinematics
> system. AnyBody allows you to drive the model by any combination of
> forward and inverse type drivers.
>
> You are right that the “natural” way to drive a model is to start
> proximaly and progress along the extremities, and to some extent
this
> is the way the standing model and the free posture model in the
> repository work. But this is a modeling issue and not a system
issue.
>
> Now about the shoulder model: It does indeed have all the natural
> degrees of freedom of a shoulder, i.e. the scapular-thoracic
movement
> and the gleno-humeral movement. But the repository models do not
> provide access to the scapular-thoracic movement through the
> mannequin file. This does not mean that the scapula is welded to
the
> thorax in a kinetic sense - only that the model specifies that this
> movement is zero; the model will still recruit the necessary muscle
> forces to keep the scapula stationary.
>
> To control the scapular-thoracic movement you must go to the
> JointsAndDrivers file of your model. A bit down you will finde the
> sterno-clavicular joint driver. To obtain scapular-thoracic
movement
> you must either control the angle of the clavicle wrt. the sternum
as
> it is done in the existing JointsAndDrivers setup, or remove one or
> more of these existing drivers and replace them by other driver
> strategies. A popular way to drive the scapula is to create a
linear
> measure between the thorax and the acromion and create two drivers
on
> this measure driving elavation and protraction/retraction.
>
> I hope this clarifies some of the general issues.
>
> Best regards,
> John

Hello Chris,

— In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@…> wrote:
> Why is this not included in the model though?
> This seems like a significant deviation from
> reality and it’s ramifications on other model
> predictions could be very much affected.

It is important to stress that the lack of readily accessible drivers
for the scapular-thoracic movements does not influence the
equilibrium of the models; the equilibrium situation is correct and
has not been simplified. So users of these models just have to judge
whether the posture and movement are realistic.

I think there are a couple of reasons why we did not implement
vairables in the mannequin file for the scapular-thoracic movement:

  1. There is not single obvious way of driving this degree of freedom
    but rather a variety of ways people may want.

  2. The Dutch shoulder group has shown that the scapular-thoracic
    movement is small in most cases of small arm elevations.

  3. These models are mere examples. Users are supposed to use them as
    a basis for generating their own models.

Best regards,
John

John,

Thanks for the explanations.

I realize that all shoulder models have their pros and cons, and that
there is no perfect, all-inclusive model. I can also understand the
reasons for not implementing scapulo-thoracic movement. But, I don’t
quite understand your comments concerning the equilibrium:

>It is important to stress that the lack of readily accessible drivers
>for the scapular-thoracic movements does not influence the
>equilibrium of the models; the equilibrium situation is correct and
>has not been simplified. So users of these models just have to judge
>whether the posture and movement are realistic.

Could you expand on this, specifically ‘the equilibrium situation is
correct and has not been simplified’?

Suppose scapulo-thoracic movement was implemented (say with that of
Inman et al, 1944, for example) and this model (with ST movement) and
the model without ST movement were compared in terms of muscle forces
at 90deg of abduction in the frontal plane. Wouldn’t the muscle forces
be different because of different muscle lengths and moment arms? If
this example doesn’t help explain to what you are referring, disregard it.

Thank you for your time,

Chris

Hi Chris

I think that what John meant is that the forces and support
conditions for the Scapula is ok in the model. So if the Scapula is
driven to the right position by the user the force equilibrium will
be ok. It is possible to have scapula-thoracic motion by driving for
example the rotation of the SC joint.

Currently there are no Scapula-thoracic movement implemented that
will automatically link this to the motion of the mumerus, the user
has to drive this manually to ensure that the motion is correct.

So if the user drives the scapula to the right position either by
implementing a rhythm or by driving the joints angles, the results
will be ok.

I hope this made it more clear

Best regards
Søren, AnyBody Support

— In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@…> wrote:
>
> John,
>
> Thanks for the explanations.
>
> I realize that all shoulder models have their pros and cons, and
that
> there is no perfect, all-inclusive model. I can also understand the
> reasons for not implementing scapulo-thoracic movement. But, I
don’t
> quite understand your comments concerning the equilibrium:
>
> >It is important to stress that the lack of readily accessible
drivers
> >for the scapular-thoracic movements does not influence the
> >equilibrium of the models; the equilibrium situation is correct
and
> >has not been simplified. So users of these models just have to
judge
> >whether the posture and movement are realistic.
>
> Could you expand on this, specifically ‘the equilibrium situation
is
> correct and has not been simplified’?
>
> Suppose scapulo-thoracic movement was implemented (say with that of
> Inman et al, 1944, for example) and this model (with ST movement)
and
> the model without ST movement were compared in terms of muscle
forces
> at 90deg of abduction in the frontal plane. Wouldn’t the muscle
forces
> be different because of different muscle lengths and moment arms?
If
> this example doesn’t help explain to what you are referring,
disregard it.
>
> Thank you for your time,
>
> Chris
>

Hi Chris

Here are some answer to your question about the cooridnate system of
the scapula. I checked the model and the weblink concerning the
location of the origin of the scapula cooridnate system.

In this link it says that the origin of the sapula coordinate system
is the bonylandmark AA (Angulus Acromialis) as i read the document
(“Origin G_AA”). There is a note though saying that in a previous
version of the document AC was used instead of AA. I tried to add an
AnyDrawRefFrame in the scapula segment (arm3d/seg.any file) and
judgeing from the picture it looks ok, the cooridnate system appears
at AA.

One small note: The body parts models are not static they are
constantly being improved by users, so they may deviate from the
original data, if the users find this needed to resolve a certain
issue in the model.

Best regards
Søren, AnyBody Support

— In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@…> wrote:
>
> Furthermore, concerning the stationary scapula…
>
> Wouldn’t this affect the muscle lengths (and thus force and moment
> predictions) of muscles which originate on the scapula and insert
on
> the humerus?
>
> Chris
>
> — In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@> wrote:
> >
> > Søren,
> >
> > Thanks for the response. I have a couple additional questions,
just to
> > make sure things are clear.
> >
> > I understand that the shoulder model is ‘driven’ by angles of the
> > humerus with respect to the thorax, as opposed to driving distal
> > segments in a sequential order (thorax to clavicle to scapula to
> > humerus, for example). After playing with the model (comparing
the
> > scapula position in 0 degrees abduction with 90 degrees
abduction) and
> > reading the code, the scapula position and orientation doesn’t
seem to
> > be affected by the motion of the humerus. Is this correct? How
does
> > the elliptical gliding plane of the scapula come into the model
then?
> > How does this allow for changes in muscle lengths which span the
> > thorax and scapula then?
> >
> > Also, the referenced information regarding the coordinate system
> > information
> >
>
(http://www.fbw.vu.nl/research/Lijn_A4/shoulder/isg/proposal/protocol
.htm)
> > states that the origin of the scapular coordinate system is
located at
> > the AC joint. In the AnyBody model though, this coordinate system
> > seems to be floating over approximately where the acromion meets
the
> > main body of the scapula, which is posterior to the AC joint. I
am
> > curious if this difference affects the locations of the muscle
origin
> > and insertion points that are defined in the scapular coordinate
> > system, such that those points that were originally defined in
the
> > Dutch shoulder model may be slightly translated in the current
AnyBody
> > model.
> >
> > Thanks again,
> >
> > Chris
> >
> > — In anyscript@yahoogroups.com, “AnyBody Support” <support@>
wrote:
> > >
> > > Hi Chris
> > >
> > > Please see posting #1445 in this group, this is also about
> > > coordinate systems in the shoulder and there are references to
> > > several other postings on this topic.
> > >
> > > Please ask again if this do not answer your questions
> > >
> > > Best regards
> > > Søren, AnyBody Support
> >
>

Søren,

Hmm. I did the same thing of adding AnyDrawRefFrame in the scapula
segment, in addition to adding AnyDrawNode to the AA node. The
location of the reference frame origin and the AA node are not the
same; the reference frame still seems to be medial and superior to the
AA. Are you using just a spine (no muscles) and shoulder model? Which
.any files may affect these positions?

Chris

— In anyscript@yahoogroups.com, “AnyBody Support” <support@…> wrote:
>
> Hi Chris
>
>
> Here are some answer to your question about the cooridnate system of
> the scapula. I checked the model and the weblink concerning the
> location of the origin of the scapula cooridnate system.
>
> In this link it says that the origin of the sapula coordinate system
> is the bonylandmark AA (Angulus Acromialis) as i read the document
> (“Origin G_AA”). There is a note though saying that in a previous
> version of the document AC was used instead of AA. I tried to add an
> AnyDrawRefFrame in the scapula segment (arm3d/seg.any file) and
> judgeing from the picture it looks ok, the cooridnate system appears
> at AA.
>
> One small note: The body parts models are not static they are
> constantly being improved by users, so they may deviate from the
> original data, if the users find this needed to resolve a certain
> issue in the model.
>
> Best regards
> Søren, AnyBody Support

Hi Chris

I am sorry about the confusion here. The origin of the scapula
cooridnate system is in the AC point, this can be seen by looking at
the sRel value of the AC poitn which is {0,0,0}. This is according
to the previous version of the document, and not the current
version!, as i wrote yesterday.

At some point it was decided to slightly move the AC joint. For this
reason a new node was created named “acj”. So this is the point used
for defining the AC joint on scapula.

As i mentioned earlier, these model are not static, they are
constantly being improved by the users. If there are good reasons
for it they may deviate from the original data.

I hope this made things more clear, otherwise please write again.

Best regards
Søren, AnyBody Support

— In anyscript@yahoogroups.com, “chrisjgatti” <cjgatti@…> wrote:
>
> Søren,
>
> Hmm. I did the same thing of adding AnyDrawRefFrame in the scapula
> segment, in addition to adding AnyDrawNode to the AA node. The
> location of the reference frame origin and the AA node are not the
> same; the reference frame still seems to be medial and superior to
the
> AA. Are you using just a spine (no muscles) and shoulder model?
Which
> .any files may affect these positions?
>
> Chris
>
>
> — In anyscript@yahoogroups.com, “AnyBody Support” <support@>
wrote:
> >
> > Hi Chris
> >
> >
> > Here are some answer to your question about the cooridnate
system of
> > the scapula. I checked the model and the weblink concerning the
> > location of the origin of the scapula cooridnate system.
> >
> > In this link it says that the origin of the sapula coordinate
system
> > is the bonylandmark AA (Angulus Acromialis) as i read the
document
> > (“Origin G_AA”). There is a note though saying that in a
previous
> > version of the document AC was used instead of AA. I tried to
add an
> > AnyDrawRefFrame in the scapula segment (arm3d/seg.any file) and
> > judgeing from the picture it looks ok, the cooridnate system
appears
> > at AA.
> >
> > One small note: The body parts models are not static they are
> > constantly being improved by users, so they may deviate from the
> > original data, if the users find this needed to resolve a
certain
> > issue in the model.
> >
> > Best regards
> > Søren, AnyBody Support
>