Validation of results from Inverse Dynamics

Hello everyone,

I am currently studying the effect of changes in gait in the kinematics and kinetics of the lower extremities, specifically the tibia. Motion capture data is used to drive the model that we have generated. In total, we have 42 markers, out of which 33 are chosen for the parameter optimization and inverse dynamics.

One of the observations made was that the results (JRFs and angles) are different when different set of markers are chosen for the study (optimized with the same tolerance in both cases), which change the results of the subsequent finite element analysis significantly in some cases.

We have also recorded the emg data, but going through some of the posts, I feel that validating through emg can be difficult. Could someone provide guidance and general remarks on how to interpret results, choose the right markers and validate/verify the results from Anybody.

Thanks.

Hi @v22 and welcome to the forum!

When optimizing marker positions you must decide which markers you believe to be most accurate placed, most often these are the ones that sit on well defined bony landmarks that are easily palpated. These are the ones you would fix in the parameter optimization.
The rest you must decide in which of the three axis they must be optimized.
I cannot give you a exact answer since it depends on the protocol and experiment.
If you have a different protocol than the one in the default model, you can change the position of the model markers so they align with the landmarks you have used in the experiment.

When changing what markers are to be optimized you will get slightly different results as the anthropometry properly also changes slightly for the individual segments.

To validate against EMG measures can as you mention be quite a hassle. Take a look at our wiki page or search other posts here on the forum for previous answers.

Best regards,
Bjørn
AnyBody Technology

Thank you Bjorn for the reply.

Sorry for the delay to get back on this post. I have been running some more cases where for the same gait cycle I choose different marker settings.

It is recommended that the optimization settings be turned off for the bony landmarks and if needed we can change the value of srel. One of the marker settings in the markers.any file we use is shown below:

CreateMarkerDriver RASI (
MarkerName= RASI,
MarkerPlacement=Trunk.SegmentsLumbar.PelvisSeg,
OptX="Off",OptY="Off",OptZ="Off",
WeightX=1.0,WeightY=1.0,WeightZ=1.0,
Model1=MotionAndParameterOptimizationModel, Model2= InverseDynamicModel,
sRelOptScalingOnOff="On"
) = {
sRelOpt = {0.11, -0.05+0.02, 0.145};
};

I had a couple of questions about this.

  1. As per a few of the posts, one way to get the markers at the bony landmarks closer to the markers in the model is manually changing the location by altering sRelOpt . How would this be different than setting the Opt to ON. The model converges in both cases but gives different results. Is turning the optimization ON at these locations a mistake?

  2. Since we are concerned with the analysis of the tibia for our project, would it correct to give the LASI and RASI markers a weight of 10 times as compared to others(even those at the tibia)?

It seems to me that the analysis is sensitive to the marker settings.

Thanks,
Vivek.

Hi @v22

  1. When you alter the sRelOpt it is the model marker you are moving so you should do this if you think that the default position is not corresponding to the position you choose in your experiment. The optimization settings are used to say "i know that the X coordinate of this marker is correct, but i am unsure of the Y coordinate" here you would allow the Y coordinate to be optimized by the study.
  2. When you increase the weight of a marker you are telling the study to minimize the kinematic error for that marker so in your case I would do the opposite. Increase the weight of the markers around the tibia.

Best regards,
Bjørn
AnyBody Technology

This topic was automatically closed 125 days after the last reply. New replies are no longer allowed.