Only one muscle does all the work

Hello everybody,

I’m working on a rat model which I’m trying to simulate with AnyBody.
The muscle definitions and motion data base on measurements on real
rats. The problem that I have is that one calve muscle is doing
almost all of the activity. Its F0 value is highly exceeded and
although there are sufficient muscle connections, the remaining
muscles are near zero. I know it is hard to analyse such a problem
with so little informations but I thought maybe somebody has already
had similar trouble and knows the root of this problem.

Regards,

Dieter

Hi Dieter,

The problem could come from a wrong moment arm or from boundary
conditions. A solution to find the root of this problem is to add a
reaction to that joint. Add the reaction to one DOF at time and see how
it affects your muscle activity. The reaction that remove the over
activity will give you the direction where the root of the problem
should be.

Best regards,
Sylvain, AnyBody Support.

— In anyscript@yahoogroups.com, “dieterkassgim” <dieterkassgim@…>
wrote:
>
> Hello everybody,
>
> I’m working on a rat model which I’m trying to simulate with AnyBody.
> The muscle definitions and motion data base on measurements on real
> rats. The problem that I have is that one calve muscle is doing
> almost all of the activity. Its F0 value is highly exceeded and
> although there are sufficient muscle connections, the remaining
> muscles are near zero. I know it is hard to analyse such a problem
> with so little informations but I thought maybe somebody has already
> had similar trouble and knows the root of this problem.
>
> Regards,
>
> Dieter
>

Hi Dieter,

Sylvain’s proposal is a good one.

It just ocurred to me that the behavior you describe is also typical
for a case where the muscles are recruited by a linear criterion
rather than the usual min/max criterion in AnyBody. This can happen
if you set the variable RecruitmentLpPenalty too large in the study
section of your model. For more information about this, have a look
at the tutorial named “A study of studies”.

Best regards,
John

> — In anyscript@yahoogroups.com, “dieterkassgim” <dieterkassgim@>
> wrote:
> >
> > Hello everybody,
> >
> > I’m working on a rat model which I’m trying to simulate with
AnyBody.
> > The muscle definitions and motion data base on measurements on
real
> > rats. The problem that I have is that one calve muscle is doing
> > almost all of the activity. Its F0 value is highly exceeded and
> > although there are sufficient muscle connections, the remaining
> > muscles are near zero. I know it is hard to analyse such a
problem
> > with so little informations but I thought maybe somebody has
already
> > had similar trouble and knows the root of this problem.
> >
> > Regards,
> >
> > Dieter
> >
>

Hello John and Sylvain,

thanks for your answers. I also thought that it has something to do
with the optimization criterion. I already use the default
MinMaxSimplex recruitment solver. But I also tried different methods
and “played” with the parameters but it always leads to similar
results.

Well I turned the knee joint reaction type to {On} and indeed the
overreaction has disappeared. But with 81 muscles I still don’t
exactly know where to begin with.

Regards,

Dieter

— In anyscript@yahoogroups.com, “John Rasmussen” <jr@…> wrote:
>
> Hi Dieter,
>
> Sylvain’s proposal is a good one.
>
> It just ocurred to me that the behavior you describe is also
typical
> for a case where the muscles are recruited by a linear criterion
> rather than the usual min/max criterion in AnyBody. This can happen
> if you set the variable RecruitmentLpPenalty too large in the study
> section of your model. For more information about this, have a look
> at the tutorial named “A study of studies”.
>
> Best regards,
> John
>
> > — In anyscript@yahoogroups.com,
“dieterkassgim” <dieterkassgim@>
> > wrote:
> > >
> > > Hello everybody,
> > >
> > > I’m working on a rat model which I’m trying to simulate with
> AnyBody.
> > > The muscle definitions and motion data base on measurements on
> real
> > > rats. The problem that I have is that one calve muscle is doing
> > > almost all of the activity. Its F0 value is highly exceeded and
> > > although there are sufficient muscle connections, the remaining
> > > muscles are near zero. I know it is hard to analyse such a
> problem
> > > with so little informations but I thought maybe somebody has
> already
> > > had similar trouble and knows the root of this problem.
> > >
> > > Regards,
> > >
> > > Dieter
> > >
> >
>

Hello Dieter,

Which muscle model do you use?

In case you use the 3-element muscle model, it might be that many of the
muscles do not have strength, because of a too long tendon length. In
that case a muscle calibration is necessary.

Regards, Mark

dieterkassgim wrote:
>
> Hello John and Sylvain,
>
> thanks for your answers. I also thought that it has something to do
> with the optimization criterion. I already use the default
> MinMaxSimplex recruitment solver. But I also tried different methods
> and “played” with the parameters but it always leads to similar
> results.
>
> Well I turned the knee joint reaction type to {On} and indeed the
> overreaction has disappeared. But with 81 muscles I still don’t
> exactly know where to begin with.
>
> Regards,
>
> Dieter
> –
> — In anyscript@yahoogroups.com <mailto:anyscript%40yahoogroups.com>,
> “John Rasmussen” <jr@…> wrote:
> >
> > Hi Dieter,
> >
> > Sylvain’s proposal is a good one.
> >
> > It just ocurred to me that the behavior you describe is also
> typical
> > for a case where the muscles are recruited by a linear criterion
> > rather than the usual min/max criterion in AnyBody. This can happen
> > if you set the variable RecruitmentLpPenalty too large in the study
> > section of your model. For more information about this, have a look
> > at the tutorial named “A study of studies”.
> >
> > Best regards,
> > John
> >
> > > — In anyscript@yahoogroups.com
> <mailto:anyscript%40yahoogroups.com>,
> “dieterkassgim” <dieterkassgim@>
> > > wrote:
> > > >
> > > > Hello everybody,
> > > >
> > > > I’m working on a rat model which I’m trying to simulate with
> > AnyBody.
> > > > The muscle definitions and motion data base on measurements on
> > real
> > > > rats. The problem that I have is that one calve muscle is doing
> > > > almost all of the activity. Its F0 value is highly exceeded and
> > > > although there are sufficient muscle connections, the remaining
> > > > muscles are near zero. I know it is hard to analyse such a
> > problem
> > > > with so little informations but I thought maybe somebody has
> > already
> > > > had similar trouble and knows the root of this problem.
> > > >
> > > > Regards,
> > > >
> > > > Dieter
> > > >
> > >
> >
>
>

Hi Mark,

I use the basic AnyMuscleModel for all muscles which are linked to
AnyViaPointMuscles.

Regards,

Dieter

— In anyscript@yahoogroups.com, Mark de Zee <mdz@…> wrote:
>
> Hello Dieter,
>
> Which muscle model do you use?
>
> In case you use the 3-element muscle model, it might be that many
of the
> muscles do not have strength, because of a too long tendon length.
In
> that case a muscle calibration is necessary.
>
> Regards, Mark
>
> dieterkassgim wrote:
> >
> > Hello John and Sylvain,
> >
> > thanks for your answers. I also thought that it has something to
do
> > with the optimization criterion. I already use the default
> > MinMaxSimplex recruitment solver. But I also tried different
methods
> > and “played” with the parameters but it always leads to similar
> > results.
> >
> > Well I turned the knee joint reaction type to {On} and indeed the
> > overreaction has disappeared. But with 81 muscles I still don’t
> > exactly know where to begin with.
> >
> > Regards,
> >
> > Dieter
> > –
> > — In anyscript@yahoogroups.com
<mailto:anyscript%40yahoogroups.com>,
> > “John Rasmussen” <jr@> wrote:
> > >
> > > Hi Dieter,
> > >
> > > Sylvain’s proposal is a good one.
> > >
> > > It just ocurred to me that the behavior you describe is also
> > typical
> > > for a case where the muscles are recruited by a linear criterion
> > > rather than the usual min/max criterion in AnyBody. This can
happen
> > > if you set the variable RecruitmentLpPenalty too large in the
study
> > > section of your model. For more information about this, have a
look
> > > at the tutorial named “A study of studies”.
> > >
> > > Best regards,
> > > John
> > >
> > > > — In anyscript@yahoogroups.com
> > <mailto:anyscript%40yahoogroups.com>,
> > “dieterkassgim” <dieterkassgim@>
> > > > wrote:
> > > > >
> > > > > Hello everybody,
> > > > >
> > > > > I’m working on a rat model which I’m trying to simulate with
> > > AnyBody.
> > > > > The muscle definitions and motion data base on measurements
on
> > > real
> > > > > rats. The problem that I have is that one calve muscle is
doing
> > > > > almost all of the activity. Its F0 value is highly exceeded
and
> > > > > although there are sufficient muscle connections, the
remaining
> > > > > muscles are near zero. I know it is hard to analyse such a
> > > problem
> > > > > with so little informations but I thought maybe somebody has
> > > already
> > > > > had similar trouble and knows the root of this problem.
> > > > >
> > > > > Regards,
> > > > >
> > > > > Dieter
> > > > >
> > > >
> > >
> >
> >
>

Hi Dieter

I do not know your joint configuration, but if you have muscles spanning
multiple joints it could be, that a muscle creating a flexion moment around
the knee at the same time will create for example a eversion moment of the
ankle. If the momentarm for the eversion is too large there might not be
other muscles around this joint to counter balance it. This may prevent this
muscle from creating a knee flexion moment, so you might see activity only
in muscles spanning one joint.

So the explanation to problem can be related to the ratios between the
moment arms of the different joints, but it could also be that some muscles
are missing in the model.

I think you should try to identify which dof is causing the problem, by
introducing reaction forces.

Then look at the moment arms on the muscles spanning this joint, and the
muscle configuration.

Hope this helps you move on.

Best regards

Søren, AnyBody Support


From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of dieterkassgim
Sent: 15 April 2008 13:08
To: anyscript@yahoogroups.com
Subject: [AnyScript] Re: Only one muscle does all the work

Hi Mark,

I use the basic AnyMuscleModel for all muscles which are linked to
AnyViaPointMuscles.

Regards,

Dieter

— In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com, Mark
de Zee <mdz@…> wrote:
>
> Hello Dieter,
>
> Which muscle model do you use?
>
> In case you use the 3-element muscle model, it might be that many
of the
> muscles do not have strength, because of a too long tendon length.
In
> that case a muscle calibration is necessary.
>
> Regards, Mark
>
> dieterkassgim wrote:
> >
> > Hello John and Sylvain,
> >
> > thanks for your answers. I also thought that it has something to
do
> > with the optimization criterion. I already use the default
> > MinMaxSimplex recruitment solver. But I also tried different
methods
> > and “played” with the parameters but it always leads to similar
> > results.
> >
> > Well I turned the knee joint reaction type to {On} and indeed the
> > overreaction has disappeared. But with 81 muscles I still don’t
> > exactly know where to begin with.
> >
> > Regards,
> >
> > Dieter
> > –
> > — In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
<mailto:anyscript%40yahoogroups.com>,
> > “John Rasmussen” <jr@> wrote:
> > >
> > > Hi Dieter,
> > >
> > > Sylvain’s proposal is a good one.
> > >
> > > It just ocurred to me that the behavior you describe is also
> > typical
> > > for a case where the muscles are recruited by a linear criterion
> > > rather than the usual min/max criterion in AnyBody. This can
happen
> > > if you set the variable RecruitmentLpPenalty too large in the
study
> > > section of your model. For more information about this, have a
look
> > > at the tutorial named “A study of studies”.
> > >
> > > Best regards,
> > > John
> > >
> > > > — In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com>
ps.com
> > <mailto:anyscript%40yahoogroups.com>,
> > “dieterkassgim” <dieterkassgim@>
> > > > wrote:
> > > > >
> > > > > Hello everybody,
> > > > >
> > > > > I’m working on a rat model which I’m trying to simulate with
> > > AnyBody.
> > > > > The muscle definitions and motion data base on measurements
on
> > > real
> > > > > rats. The problem that I have is that one calve muscle is
doing
> > > > > almost all of the activity. Its F0 value is highly exceeded
and
> > > > > although there are sufficient muscle connections, the
remaining
> > > > > muscles are near zero. I know it is hard to analyse such a
> > > problem
> > > > > with so little informations but I thought maybe somebody has
> > > already
> > > > > had similar trouble and knows the root of this problem.
> > > > >
> > > > > Regards,
> > > > >
> > > > > Dieter
> > > > >
> > > >
> > >
> >
> >
>

[Non-text portions of this message have been removed]