non phisiological knee reaction forces

[Yanis AMMOUCHE]

Dear all,

Lamberto et al in 2015 published an article where they used three musculoskeletal models (London Lower limb model, Lower limb, G2392). They said at one point of the article (see in attachment) that they did not compute knee reaction forces for the models that include a patella because they lead to non physiological resulted. The issue is detailed on this article : http://www.sciencedirect.com/science/ar ... via%3Dihub.

1) Did this problem have a solution implemented on OpenSim ?

2) If not, I'd like to modify the code of the joint reaction forces to get rid of that non phisiological force. I suppose that the joint reaction forces are extracted by computing the differences beetween the intersegmental forces and the muscles forces but I'd like to know where exactly it is implemented.

Thank you.

[Thomas Uchida]

I don't think all models with a patella will necessarily lead to non-physiological results. It seems that one must consider how the patella has been modeled and what data are being sought. Is there any useful information in any of these papers?

- Lerner, DeMers, Delp, Browning (2015). How tibiofemoral alignment and contact locations affect predictions of medial and lateral tibiofemoral contact forces. Journal of Biomechanics 48(4):644–650.
- DeMers, Pal, Delp (2014). Changes in tibiofemoral forces due to variations in muscle activity during walking. Journal of Orthopaedic Research 32(6):769–776.
- Gerus, Sartori, Besier, et al. (2013) Subject-specific knee joint geometry improves predictions of medial tibiofemoral contact forces. Journal of Biomechanics 46(16):2778–2786.

[Yanis AMMOUCHE]

Thank you for your answer.

The first paper consist of 4 simulations of knee contact forces for a patient that has a tibio-femoral alignment of 174°. They did more and more subject specific simulation and compare them with in vivo datas of that patient.

The second one computes the tibiofemoral forces with different objective function for the static optimization and discussion were held about which muscles increase the most the tibiofemoral reaction forces.

Nevertheless, there was no reference to corrections about what has been mentionned in Lamberto et al article.

They say that using a tibio-patellar joint with coupled coordinates beetween the knee flexion and patella function led to non physical transmited forces beetween the patella and tibia.
However, in the model used in the first article (in which I have a strong interest), on the xml file, it is said that "DeMers and Steele redefined the planar patella kinematics as a planar joint between the patella and femur instead of constraints between the patella and tibia. " without any other explanations.

Maybe it is a problem I can adress by postprocessing the datas, but I strongly need the place they code the joint reaction forces.

[Thomas Uchida]

I strongly need the place they code the joint reaction forces.

Have you looked at the JointReaction Analysis?
- API documentation: https://simtk.org/api_docs/opensim/api_ ... ction.html
- Code on GitHub: https://github.com/opensim-org/opensim- ... Reaction.h and https://github.com/opensim-org/opensim- ... action.cpp

[Yanis AMMOUCHE]

Yes, and it seems that they use a function : calcReactionOnParentExpressedInGround as shown in attachment. This function use another one called findMobilizerReactionOnParentAtFInGround.

Nevertheless, I can't access the c++ code of this function even on the API.

Thanks for your answer

[Thomas Uchida]

Yes, and it seems that they use a function : calcReactionOnParentExpressedInGround as shown in attachment. This function use another one called findMobilizerReactionOnParentAtFInGround.
Nevertheless, I can't access the c++ code of this function even on the API.

You can find the implementations of these methods by searching the GitHub repositories.
- calcReactionOnParentExpressedInGround() is in Joint.h: https://github.com/opensim-org/opensim- ... int.h#L235
- findMobilizerReactionOnParentAtFInGround() is a method in Simbody's MobilizedBody class here: https://github.com/simbody/simbody/blob ... y.cpp#L397
There is a Search box at the top of each of these pages, which you can use to search for methods, variables, comments, etc. in the OpenSim and Simbody source code.