simtk.org

Hi dear,
We are going to simulate walking in our gait lab, using Hang xu model, to estimate joint reaction force only in dominant knee during the stance phase of the gait. Unfortunately, we have just one force plate, that will be in contact with dominant foot.
The question is that : does reported knee joint reaction force have correct value for the entire stance phase? Considering that only target foot hits the force plate.

Thank you for your help

The answer is yes, the knee joint forces will be correct, as long as you do not use the residual reduction algorithm (RRA). A suitable processing pipeline is: inverse kinematics, inverse dynamics, static optimization, joint reaction analysis.

If you do the analysis this way, you don't need a full body model. Your model can just include pelvis and one leg. This will make the motion capture easier to do and the Opensim analysis a lot faster. A full body model will work, but could actually make the kinematic analysis slightly worse, because the upper body markers will influence the leg motion. You will probably get better leg kinematics if the model does not include anything above the pelvis, and does not include the other leg.

Ton van den Bogert

Thank you very much for your accurate prompt answer.

But reviewing the literature, i found a research performed in similar situation (using one FP and gait 2392 model). They estimate JRF and Muscle force with this pipeline: IK- ID- Static Optimization- RRA- CMC- analysis!
Does it make sense to include RRA and CMC while using SO?!

I would be interested in the reference to that work you mentioned. Maybe there is something I did not think about.

In my understanding, RRA would only be valid in the single stance phase (not the whole stance phase) if you only have one force plate. This is because residual loads can't be reduced to zero unless all external forces are known.

CMC uses dynamic muscle muscle models, and requires RRA to generate a dynamically consistent full body model and movement trajectory. SO is much simpler and does not require RRA and does not require full body model or motion capture.

For walking and other slow movements, SO and CMC will give similar results. Especially if you're interested in JRF rather than details of muscle recruitment. SO is much easier to do, can use one force plate, and less can go wrong.

[1] Lin YC et al. (2012) Comparison of different methods for estimating muscle forces in human movement. Proc Inst Mech Eng H 226(2):103-112.

This is the mentioned article: Nadi A, Karimi MT, Fatoye F, Jafari A. The Effects of Scoliosis on Spinal Muscles Length and Joint Contact Forces. Journal of Mechanics in Medicine and Biology. 2018 Sep 13;18(06):1850022.

what if we decide to estimate JRF and also Muscle force pattern?

Does this pipeline ( IK- ID- SO- analysis) still work?

I don't have access to the full paper so I only read the abstract. They mention "a force plate" which would be insufficient for RRA during double stance. It may be OK if the peak forces occurred during single stance and the double stance phase can be ignored. I would have used SO, and without RRA, though.

According to Lin's paper which I cited, SO and CMC give very similar results for muscle force patterns. For JRF, the similarity will be even greater because Total JRF is not very sensitive to how the load is distributed between muscles. The load distribution in the knee between medial and lateral condyles is more sensitive to muscle force distribution. If you're interested in that, read papers related to the "Grand Challenge": https://simtk.org/projects/kneeloads.

Thanks a lot Dr. van den Bogert
We will definitely use your tips.

Hi, Ton van den Bogert, Maher Mahdavi,
I'm studing opensim, and I saw the reply that if I use one-leg model ,I'll don't need to do RRA. But if without RRA's result, I might get a large residual force. If these forces will effect the knee joint force?
Could you help me out? Thank you very much!