gaslat & gasmed unexpected wrapping behavior and moment arms in OpenSim 4.0

[Ke Song]

Hi Community,

I have 2 questions on gastrocnemius muscle-tendon units in the OpenSim models as they are loaded in GUI v4.0. I am using the model shared by Lai et al. (https://simtk.org/projects/model-high-flex), which have similar muscle path definitions as earlier models by E. Arnold et al. 2010 and Rajagopal et al. 2016. I noticed that when I set the model's knee flexion to >35 degrees, the lateral gastroc (gaslat_r/l) seemed to start exhibiting weird wrapping behavior; similar issue also appears to medial gastroc (gasmed) at knee flexion > 60 degrees. See snapshot below:

Gas wrapping at 75deg KNEE angle r Lai OS4.0.png (72.48 KiB) Viewed 579 times

These 2 muscles are wrapped over 2 cylinders between origin and insertion (as shown on the other leg in snapshot). It looks like wrapping around the lower cylinder is erroneous at >35deg knee flexion. I am also seeing similar issues with E. Arnold et al. 2010 and Rajagopal et al. 2016 models in OpenSim 4.0 GUI. Interestingly, when I plotted knee moment arms for gaslat and gasmed vs. knee angle, there were no discontinuities. Furthermore, when I loaded these same models in OpenSim 3.3 GUI, these muscle paths looked normal across the whole knee ROM.

Have those of you who also use any of these models also seen similar issues? And as I am only seeing this in OpenSim 4.0 but not 3.3, is this related to any setup in OpenSim 4.0 that I am unaware of?
_____
And as a separate 2nd question, I also found these gaslat and gasmed muscles to have nonzero moment arm lengths for joints that they do not cross: specifically, the 3 hip coordinates and the 3 pelvis rotational coordinates (tilt/list/rotation). In contrary to the wrapping issue mentioned above, this happens at low flexion to hyper-extension knee angle range. Please see an example below:

HIP flexion r moment arm vs. KNEE angle r Lai OS4.0.png (53.37 KiB) Viewed 579 times

I am also seeing this moment arm issue only in OpenSim 4.0 but not in 3.3 (where pelvis/hip moment arms are all 0.0000 for gaslat & gasmed). I am concerned if these hip moment arms would affect force estimation at the hip joint. And, as these moment arms are fairly small (up to 1.5mm in the snapshot), I wonder if I can neglect them? If not, what can I do to fix this issue?

Thank you very much and I look forward to any help and suggestions. Best Regards,

Ke

[Ayman Habib]

Hi Ke,

Thanks for reporting. The issue has to do with muscles that wrap around multiple wrap objects, particularly when you enter a coordinate value or load a motion that switches one object on and another off. If you save the model in this configuration then reload it, it should display correctly. It's the "transition" (optimized to update display during motion) that is not working properly on concurrent/multiple wrapping changes.

This, however, has no effect on results/plots whatsoever.

Hope this helps,
-Ayman

[Ke Song]

Hi Ayman,

Thanks for explaining the wrapping display issue; good to know it will not affect the results. As a follow-up, how about the 2nd issue I described in my original post, that gaslat and gasmed have small but non-zero moment arm lengths about hip and pelvis rotational coordinates, when plotted in OpenSim 4.0 (but are all zeroes in OpenSim 3.3)? Is such unexpected moment arm lengths related to the wrapping display aforementioned, or it's another distinct issue? Do you expect these non-physical moment arm lengths to influence static optimization / CMC results, or they are small enough to neglect?

Ke

[Axel Koussou]

Hi Ke,

I am using OpenSim 4.1 and just as you, I am also seeing a non-zero moment arm for the gastrocnemius-tendon units for hip flexion movement.

Thus, I would like to know if you found what causes this oddness?

Thanks in advance,

Regards