Static Optimization Muscle Forces

[Ryan Byrne]

Hi everyone,

I'm working on a full body back-lifting model and have run into some trouble during Static Optimization. It's my understanding that coordinate actuators placed at the joints should not be used until the muscles cannot stabilize the model. However, there are many muscles in my model that remain inactive while coordinate actuators see a force. I have checked the muscle lengths, and during the motion they go from being larger than the optimal fiber length to smaller than the optimal fiber length - yet no activation occurs.

I have also messed around with the coordinate actuator optimal force values to be anywhere from 5N to 10,000N. At CA optimal force of 5N, almost all back muscles are inactive and the coordinate actuators have values of around 200-300N. At a CA optimal limit of 10,000N, many muscles are recruited (although a good amount are still inactive even while being stretched) and the coordinate actuators see a force of about 1000-1400N. To me this result doesn't make much sense, as the coordinate actuators are meant to be reserved for when needed, but are activated regardless of how much the muscles are involved in my model. The stretched muscles remaining inactive is also confusing to me. Any help on either of these issues would be much appreciated.

[Dimitar Stanev]

Maybe the model is badly scaled and large passive forces are applied. You can check the values of the forces after running SO.

[Ryan Byrne]

Dear Dimitar,

Thank you for your reply. I think you have pointed me in the right direction. I have checked the passive forces and normalized fiber lengths after running static optimization. I see that there are a few lower extremity muscles and also a few in the iliocostalis that have passive muscles forces of several hundred N. For the lower extremity this is certainly a problem as there's little to no movement of the lower body in my model. However, the normalized fiber lengths in the iliocostalis (around 1.7 max) seem to make sense given that there is about a 40 degree flexion in the lumbar spine.

The main issue I'm having is that I'm expecting the multifidus (MF) muscles to be at least somewhat active during back flexion. Their passive muscle forces vary from 0-6 N (normalized fiber lengths vary anywhere from 0.7-1.5). MuscleAnalysis is telling me that for all of my MF muscles, the activeFiberForce is around 0.5-2 N each, with passive fiber forces ranging anywhere from 0-6 N. However, the excitation, activation and muscle forces reported by SO remain essentially zero which is leading to activation of the coordinate actuators. I guess I'm not understanding how this is possible - perhaps a better understanding of how the excitation of the muscle is determined in SO would help?

[Dimitar Stanev]

I guess I'm not understanding how this is possible - perhaps a better understanding of how the excitation of the muscle is determined in SO would help?

https://simtk-confluence.stanford.edu:8 ... tion+Works

As I mentioned some muscle parameters are badly scaled. SO essentially performs an optimization that finds a set of muscle activations that balance the required joint torques. It may be that the model is actuated without activating the muscles due to the passive forces. For example if you would like to perform a flexion then if the flexors are operating in bad region (e.g. passive forces) then by controlling the extensors you can adjust the flexion accordingly. You can look at the activations of the other group of muscles. Moreover you can look at the moment arm values for the muscles vs coordinates.