Hi,
I am just a begineer to the OpenSim, I was wondering if we can use the Musculoskeletal Model avaiblae in the OpenSim to obtain the muscle forces for fixed pose. For an example, if I rotate my Glenohumeral (GH) Joint to 90 degree internally and have 30 degree elbow flexion followed by 20 degree GH abduction, to get the final resting position of my model.
So, can I get the muscle forces for this final resting position using OpenSim or I can only get the dynamics muscle forces by using MOCAP data for that final position?
Any help or guidelines how to approach this problem is appreciated!!!
Thank You
Kamal
Muscle Forces for Pre-Defined Final Pose
- Kamal Gautam
- Posts: 9
- Joined: Sun Mar 12, 2023 2:18 pm
- Thomas Uchida
- Posts: 1787
- Joined: Wed May 16, 2012 11:40 am
Re: Muscle Forces for Pre-Defined Final Pose
You could use OpenSim to estimate muscle forces for an arbitrary pose by doing the following:
- adjust the pose of the model in the GUI
- lock all the coordinates (see the "Coordinate Controls and Poses" page in the User's Guide: https://simtk-confluence.stanford.edu:8 ... +and+Poses)
- run a short forward dynamic simulation (e.g., using the green "run" button; see the "Toolbar" page: https://simtk-confluence.stanford.edu:8 ... im/Toolbar)
- save the resulting motion
- unlock the arm coordinates
- estimate muscle forces using the Static Optimization or CMC Tool
Additional information about OpenSim's capabilities can be found on the "Overview of OpenSim Workflows" page (https://simtk-confluence.stanford.edu:8 ... +Workflows). However, note that there are, in general, infinitely many solutions to the muscle redundancy problem, so the muscle force estimates you obtain may not be sufficient to answer your research question.
- adjust the pose of the model in the GUI
- lock all the coordinates (see the "Coordinate Controls and Poses" page in the User's Guide: https://simtk-confluence.stanford.edu:8 ... +and+Poses)
- run a short forward dynamic simulation (e.g., using the green "run" button; see the "Toolbar" page: https://simtk-confluence.stanford.edu:8 ... im/Toolbar)
- save the resulting motion
- unlock the arm coordinates
- estimate muscle forces using the Static Optimization or CMC Tool
Additional information about OpenSim's capabilities can be found on the "Overview of OpenSim Workflows" page (https://simtk-confluence.stanford.edu:8 ... +Workflows). However, note that there are, in general, infinitely many solutions to the muscle redundancy problem, so the muscle force estimates you obtain may not be sufficient to answer your research question.
- Kamal Gautam
- Posts: 9
- Joined: Sun Mar 12, 2023 2:18 pm
Re: Muscle Forces for Pre-Defined Final Pose
Thank you very much for the reply. It will be the beginning point for me. I have been going through all the papers published for shoulder models. So, combination of both should work.
- Kamal Gautam
- Posts: 9
- Joined: Sun Mar 12, 2023 2:18 pm
Re: Muscle Forces for Pre-Defined Final Pose
If I am interest only in the elbow flexion, can I just unlock the elbow_flexion coordinate from "MoBL_ARM_Dynamics_Model".
I tried with only elbow_flexion coordiante ON and with all other coordiantes on focusing on the only muscles crossing the elbow (e.g. triceps, biceps and brachalias). I got different results with the variation on the coordiantes locking/unlocking;. The motion file contains elbow angle set to 90 degree from 0 to 0.2 sec. I append a resever acuators, and the reseve forces/moments are very less (0.003).
So, I was wondering whether, I followed the correct algorithm or not? and what is the difference in simulation with coordinates ON and OFF.
I appreciate you help.
Thank You
- Thomas Uchida
- Posts: 1787
- Joined: Wed May 16, 2012 11:40 am
Re: Muscle Forces for Pre-Defined Final Pose
If you are just starting out, you might want to check out the examples and tutorials here: https://simtk-confluence.stanford.edu:8 ... +Tutorials. In particular, the "The Strength of Simulation: Estimating Leg Muscle Forces in Stance and Swing" example might be helpful (https://simtk-confluence.stanford.edu:8 ... n+in+Swing).
Perhaps you mean locking/unlocking coordinates? If you lock a coordinate, its value will not change during a forward simulation. There is more information about locking/unlocking coordinates on the "Coordinate Controls and Poses" page here: https://simtk-confluence.stanford.edu:8 ... +and+Poses. I suggested locking the coordinates and running a forward simulation because you don't have any data to track. You can quickly generate data where the model is stationary by locking all the coordinates and running a short forward simulation. The resulting states can then be used with Static Optimization or CMC to estimate the muscle forces that would be necessary to produce this motion (i.e. without the constraints that are created when the coordinates are locked).and what is the difference in simulation with coordinates ON and OFF.
- Kamal Gautam
- Posts: 9
- Joined: Sun Mar 12, 2023 2:18 pm
Re: Muscle Forces for Pre-Defined Final Pose
Hi Thomas,
Thank you very much for the reply and pointing me to the tutorials. I have already gone through those tutorials but don't know why I am still struggling with it.
I have used MoBL-ARMS Upper Extremity Dynamics Model, I have locked the coordinates as shown in Figure below to generate a FD for 0.01 sec. Then I have used that FD motion for estimating muscle force distribution using SO for two cases: (1) only the elbow_flexion coordinate unlocked and
(2) all the coordiantes unlocked (except r_z).
The results for the elbow muscles activation and forces for two different case is as shown in Figure. Here, when I considered only the elbow coordinate the force by the BRA (Brachialis muscle) is higher while when I unlock all the coordinated the forces on the tricpes and other biceps muscle get's higer for same position (90 degree elbow flexion and -90 pro/sup).
So, I was wondering why there is so much difference on the two results, since the moment around the elbow is same i.e.,3.178 Nm on both the cases.
Thank you for the time.
Thank you very much for the reply and pointing me to the tutorials. I have already gone through those tutorials but don't know why I am still struggling with it.
I have used MoBL-ARMS Upper Extremity Dynamics Model, I have locked the coordinates as shown in Figure below to generate a FD for 0.01 sec. Then I have used that FD motion for estimating muscle force distribution using SO for two cases: (1) only the elbow_flexion coordinate unlocked and
(2) all the coordiantes unlocked (except r_z).
The results for the elbow muscles activation and forces for two different case is as shown in Figure. Here, when I considered only the elbow coordinate the force by the BRA (Brachialis muscle) is higher while when I unlock all the coordinated the forces on the tricpes and other biceps muscle get's higer for same position (90 degree elbow flexion and -90 pro/sup).
So, I was wondering why there is so much difference on the two results, since the moment around the elbow is same i.e.,3.178 Nm on both the cases.
Thank you for the time.
- Attachments
-
- Locked Coordiantes Value.png (21 KiB) Viewed 721 times
-
- Muscle Forces and Activation.png (21.82 KiB) Viewed 721 times
-
- opensim_snapshot.png (313.4 KiB) Viewed 721 times
- Thomas Uchida
- Posts: 1787
- Joined: Wed May 16, 2012 11:40 am
Re: Muscle Forces for Pre-Defined Final Pose
There would be no difference in these two situations if each muscle actuated only one coordinate; however, many muscles cross more than one joint. The solution will need to coordinate the muscles such that the desired motion is achieved at all joints.
- Kamal Gautam
- Posts: 9
- Joined: Sun Mar 12, 2023 2:18 pm
Re: Muscle Forces for Pre-Defined Final Pose
I got your point.
So, unlocking all the coordinate is more of a practical where the muscles crosses more than one joint.
Thank You
So, unlocking all the coordinate is more of a practical where the muscles crosses more than one joint.
Thank You