EMG and Muscle activity does not match up well

Hi,

I’ve been working on a model that compares EMG data collected to
muscle activity data calculated by my AnyBody model. I plot the two
over one another (for the muscle of interest) in the ‘AnyChart 2D/3D’
window only after running the calibration sequence. There a two
things I notice that I’m hoping someone could help me rectify after
looking at my model:

1)While the timing of activation of 3 out of the 4 muscles I’m
looking at match rather well, the muscle activity graph of the 4th
(the right tibialis anterior muscle) does not look anything at all
like the EMG data for that muscle. I would like some help figuring
out why this is so, and if there is something I’m doing ewrong that
can be easily rectified.

2)How can I get the muscle activation graphs to better represent the
pattern of the EMG data. Though I was not expecting the calculated
activity to match the EMG exactly in pattern, I was hoping to see the
areas with relatively higher EMG activity represented likewise in the
model’s calculated activity (and a similar corelation in the points
of lower activity). Is it even possible to get results like this with
AnyBody, and if so how can this be done ?

Please take a look at my model. Your insite would be greatly
appreciated. Its posted with the name ‘EMG-Activity.rar’

Jerome

Hi Jerome,

thats an interesting project you are running. I am looking at some
proprioceptive tests too, but with the knee. In terms of your muscle activity,
have you tried using the other recruitement solvers in AnyBody?

I have found that MinMaxOOSolSimplex and OOSolQP can produce some better results
when comparing to EMG.

I hope this helps

best wishes

Peter

Jerome <jromeo_sc_forever@yahoo.com> wrote: Hi,

I’ve been working on a model that compares EMG data collected to
muscle activity data calculated by my AnyBody model. I plot the two
over one another (for the muscle of interest) in the ‘AnyChart 2D/3D’
window only after running the calibration sequence. There a two
things I notice that I’m hoping someone could help me rectify after
looking at my model:

1)While the timing of activation of 3 out of the 4 muscles I’m
looking at match rather well, the muscle activity graph of the 4th
(the right tibialis anterior muscle) does not look anything at all
like the EMG data for that muscle. I would like some help figuring
out why this is so, and if there is something I’m doing ewrong that
can be easily rectified.

2)How can I get the muscle activation graphs to better represent the
pattern of the EMG data. Though I was not expecting the calculated
activity to match the EMG exactly in pattern, I was hoping to see the
areas with relatively higher EMG activity represented likewise in the
model’s calculated activity (and a similar corelation in the points
of lower activity). Is it even possible to get results like this with
AnyBody, and if so how can this be done ?

Please take a look at my model. Your insite would be greatly
appreciated. Its posted with the name ‘EMG-Activity.rar’

Jerome


Users of the AnyBody Modeling System help each other create biomechanical models
in the AnyScript language.Yahoo! Groups Links


Sent from Yahoo! Mail.
A Smarter Email.

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

Thanks for the suggestion Peter. I tried the MinMaxOOSolSimplex and
MinMaxSimplex solvers and neither of them ran in the
InverseDynamicAnalysis. Both gave the same error:

‘Study’ : Muscle recruitment analysis failed, simplex solver
reports that solution does not satisfy all constraints.

I also tried the MinMaxOOSolQP solver but (though it ran) this
yielded even more unfavourable results. Anyone with any other
suggestions ?

— In anyscript@yahoogroups.com, Peter Worsley <peter.worsley@…>
wrote:
>
> Hi Jerome,
>
> thats an interesting project you are running. I am looking at some
proprioceptive tests too, but with the knee. In terms of your muscle
activity, have you tried using the other recruitement solvers in
AnyBody?
>
> I have found that MinMaxOOSolSimplex and OOSolQP can produce some
better results when comparing to EMG.
>
> I hope this helps
>
> best wishes
>
> Peter
>
> Jerome <jromeo_sc_forever@…> wrote: Hi,
>
> I’ve been working on a model that compares EMG data collected to
> muscle activity data calculated by my AnyBody model. I plot the two
> over one another (for the muscle of interest) in the ‘AnyChart
2D/3D’
> window only after running the calibration sequence. There a two
> things I notice that I’m hoping someone could help me rectify after
> looking at my model:
>
> 1)While the timing of activation of 3 out of the 4 muscles I’m
> looking at match rather well, the muscle activity graph of the 4th
> (the right tibialis anterior muscle) does not look anything at all
> like the EMG data for that muscle. I would like some help figuring
> out why this is so, and if there is something I’m doing ewrong that
> can be easily rectified.
>
> 2)How can I get the muscle activation graphs to better represent
the
> pattern of the EMG data. Though I was not expecting the calculated
> activity to match the EMG exactly in pattern, I was hoping to see
the
> areas with relatively higher EMG activity represented likewise in
the
> model’s calculated activity (and a similar corelation in the points
> of lower activity). Is it even possible to get results like this
with
> AnyBody, and if so how can this be done ?
>
> Please take a look at my model. Your insite would be greatly
> appreciated. Its posted with the name ‘EMG-Activity.rar’
>
> Jerome
>
>
> ------------------------------------
>
> Users of the AnyBody Modeling System help each other create
biomechanical models in the AnyScript language.Yahoo! Groups Links
>
>
>
>
>
>
> ---------------------------------
> Sent from Yahoo! Mail.
> A Smarter Email.
>
> [Non-text portions of this message have been removed]
>

sorry jerome,

I should have given you the script descriptions.
you can put a recruitment penalty on the MinMaxOOSolQP by saying;

AnyFolder &Model = .OptModel;
RecruitmentSolver = MinMaxOOSolQP;
nStep=158;
Gravity = {0.001, 0.001,-9.81};
RecruitmentLpPenalty=0;
RecruitmentQpPenalty=1000;
tStart=2.6917+0.1;
tEnd= 4.7872892405;

I have had mixed results when comparing my EMG to muscle activity in AnyBody,
especially at the ankle where it is hard to constrain. Are you taking into
account AnyBody maybe a fraction earlier in muscle recruitment (it doesn’t have
a CNS system).

I hope this helps

peter

Jerome <jromeo_sc_forever@yahoo.com> wrote: Thanks for the suggestion Peter. I
tried the MinMaxOOSolSimplex and
MinMaxSimplex solvers and neither of them ran in the
InverseDynamicAnalysis. Both gave the same error:

‘Study’ : Muscle recruitment analysis failed, simplex solver
reports that solution does not satisfy all constraints.

I also tried the MinMaxOOSolQP solver but (though it ran) this
yielded even more unfavourable results. Anyone with any other
suggestions ?

— In anyscript@yahoogroups.com, Peter Worsley

wrote:
>
> Hi Jerome,
>
> thats an interesting project you are running. I am looking at some
proprioceptive tests too, but with the knee. In terms of your muscle
activity, have you tried using the other recruitement solvers in
AnyBody?
>
> I have found that MinMaxOOSolSimplex and OOSolQP can produce some
better results when comparing to EMG.
>
> I hope this helps
>
> best wishes
>
> Peter
>
> Jerome wrote: Hi,
>
> I’ve been working on a model that compares EMG data collected to
> muscle activity data calculated by my AnyBody model. I plot the two
> over one another (for the muscle of interest) in the ‘AnyChart
2D/3D’
> window only after running the calibration sequence. There a two
> things I notice that I’m hoping someone could help me rectify after
> looking at my model:
>
> 1)While the timing of activation of 3 out of the 4 muscles I’m
> looking at match rather well, the muscle activity graph of the 4th
> (the right tibialis anterior muscle) does not look anything at all
> like the EMG data for that muscle. I would like some help figuring
> out why this is so, and if there is something I’m doing ewrong that
> can be easily rectified.
>
> 2)How can I get the muscle activation graphs to better represent
the
> pattern of the EMG data. Though I was not expecting the calculated
> activity to match the EMG exactly in pattern, I was hoping to see
the
> areas with relatively higher EMG activity represented likewise in
the
> model’s calculated activity (and a similar corelation in the points
> of lower activity). Is it even possible to get results like this
with
> AnyBody, and if so how can this be done ?
>
> Please take a look at my model. Your insite would be greatly
> appreciated. Its posted with the name ‘EMG-Activity.rar’
>
> Jerome
>
>
> ------------------------------------
>
> Users of the AnyBody Modeling System help each other create
biomechanical models in the AnyScript language.Yahoo! Groups Links
>
>
>
>
>
>
> ---------------------------------
> Sent from Yahoo! Mail.
> A Smarter Email.
>
> [Non-text portions of this message have been removed]
>


Users of the AnyBody Modeling System help each other create biomechanical models
in the AnyScript language.Yahoo! Groups Links


Sent from Yahoo! Mail.
A Smarter Email.

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

Hello.
Regarding Q2 you might get some help by reading the post (message)
#2089 and those that follows in the same thread.
/jokke

— In anyscript@yahoogroups.com, “Jerome” <jromeo_sc_forever@…>
wrote:
>
> Hi,
>
> I’ve been working on a model that compares EMG data collected to
> muscle activity data calculated by my AnyBody model. I plot the two
> over one another (for the muscle of interest) in the ‘AnyChart
2D/3D’
> window only after running the calibration sequence. There a two
> things I notice that I’m hoping someone could help me rectify after
> looking at my model:
>
> 1)While the timing of activation of 3 out of the 4 muscles I’m
> looking at match rather well, the muscle activity graph of the 4th
> (the right tibialis anterior muscle) does not look anything at all
> like the EMG data for that muscle. I would like some help figuring
> out why this is so, and if there is something I’m doing ewrong that
> can be easily rectified.
>
> 2)How can I get the muscle activation graphs to better represent
the
> pattern of the EMG data. Though I was not expecting the calculated
> activity to match the EMG exactly in pattern, I was hoping to see
the
> areas with relatively higher EMG activity represented likewise in
the
> model’s calculated activity (and a similar corelation in the points
> of lower activity). Is it even possible to get results like this
with
> AnyBody, and if so how can this be done ?
>
> Please take a look at my model. Your insite would be greatly
> appreciated. Its posted with the name ‘EMG-Activity.rar’
>
> Jerome
>