simtk.org

Hi,

I have a quick question about the peroneus longus muscle activation in static optimaization. Using gait 2392 and Rajagopal generic model, I have averaged 50 stances of walking in healthy adults to get ankle joint contact force using static optimization. The magnitude and temporal aspects of axial ankle joint contact force look good compared to previosuly published data. But is it normal to get activation level of less than .05 out of 1 for the peroneus longus during stance phase of walking? I know it is an assitant plantarflexor muscles compared to triceps surae but I think it is too small.

I tried to find some previously published articles if there is any related data out there but couldn't find any studies reported peroneus longus activation data either for outcomes or validation.

Is that potentially because I'm locking subtalar and MTP joints in the residual reduction algorithm? Or the generic model is not sensitive to catch it?

Thanks,

You may want to try without residual reduction. Residual reduction is not required for static optimization, and could sometimes make things worse.

I took a look at my own results with static optimization and the Delp 1990 musculoskeletal model [1]. Muscle activation results are provided as an Excel file in the supplementary material of [1]. This shows peak activation of about 60% in the Peroneus Longus. The image below is the average and standard deviation of all subjects. I shifted the time axis by 5% to compensate for the real time filter delay.
The pattern is similar to the large plantarflexors which makes sense because (as you pointed out) they are synergists.

Another possible source of the difference is that I used a volume-weighted cost function for the static optimization. There is a mathematical proof that this is required to make the result invariant to anatomical discretization [2]. With volume weighting, the cost of activating the PL is much smaller than the cost of activating the large plantarflexors, so the PL activation will be larger than without volume weighting.

The good news is that the ankle joint reaction force is not very sensitive to how the joint moment is distributed between synergistic muscles. Different cost functions will probably give very similar ankle joint forces.

Ton van den Bogert

[1] van den Bogert AJ, Geijtenbeek T, Even-Zohar O, Steenbrink F, Hardin EC (2013) A real-time system for biomechanical analysis of human movement and muscle function. Medical and Biological Engineering and Computing 51:1069-1077. https://link.springer.com/article/10.10 ... 013-1076-z
[2] Holmberg LJ, Klarbring A (2012) Muscle decomposition and recruitment criteria influence muscle force estimates. Multibody Syst Dyn 28:283-289.

Thank you so much for your advice. I will definetely try these.

If the Opensim GUI does not offer the option of a volume or strength-weighted cost function (I did not see it mentioned in the documentation), you will have to do it with the Opensim API. For each frame of data, the following steps would be needed:

1. Ask Opensim to calculate the joint angles (inverse kinematics).
2. Ask Opensim to calculate the muscle moment arms at those angles.
3. Ask Opensim to calculate the joint moments (inverse dynamics).
4. Ask Opensim what the strength (Fmax) and fiber length of each muscle is and calculate muscle volumes: Fmax * fiber length. For efficiency, this can be done outside of the data processing loop.
5. Use fmincon (Matlab) or equivalent tool to solve the static optimization problem.