Hi
My name is Ana, and I am a reconverted aerospace engineer trying to learn about muscles.
I have been reading about Musculotendon Actuator Dynamics, and this are composed of "activation dynamics" and "musculotendon contraction dynamics", which are assumed to be decoupled. If I well understood, the inputs and the outputs are:
- Activation dynamics (first order process) :
Input: excitation u(t)
Output: activation a(t)
- Musculotendon contraction dynamics (first order process):
Input: activation a(t)
Output: muscle force(t) = tendon force (t)
One of the outputs of CMC are the "actuator controls", which are calculated after the Static Optimization step. Are these "controls" referring to excitation u(t), or activation a(t)? I understood this seems to be confusing for a lot of people, so hopefully this is going to be helpful for a lot of people?
I guess the calculation of forces(t) from the activation a(t) is included in the Forward Dynamic step. Is this correct? Is there some documentation/paper to know more about this calculation/equations (musculotendon contraction dynamics)? I am reading Zajac 1989, but I am still having some trouble with this part.
Thanks!
Ana
excitation vs. activation
-
Matthew Millard
- Posts: 24
- Joined: Fri Aug 01, 2008 7:17 am
Re: excitation vs. activation
Hello,
My name is Matt Millard, and I'm a post-doc at OpenSim. I've written the new muscle models you'll find in OpenSim 3.0 (Thelen2003Muscle, Millard2012EquilibriumMuscle and Millard2012AccelerationMuscle), and I'm working on a faster simulating muscle for the next release. I'll intersperse my response to your questions ...
>I have been reading about Musculotendon Actuator Dynamics, and this are composed of "activation dynamics" and >"musculotendon contraction dynamics", which are assumed to be decoupled. If I well understood, the inputs and the outputs >are:
>- Activation dynamics (first order process) :
>Input: excitation u(t)
>Output: activation a(t)
Correct.
Physically excitation is the electrical signal (an impulse train) that the nervous system sends to the muscle. The electrical excitation will cause a single muscle fiber to change chemically, and ultimately causing the fiber to generate a twitch of force (well modeled by a 2nd order critically damped system [1]). When the nervous system electrically stimulates many fibers, many twitches are created in parallel, generating larger forces. The force step-response of a muscle, when many of its fibers are recruited, looks very similar to a first-order step response, and so, a first order model has been used to model muscle `activation'.
>- Musculotendon contraction dynamics (first order process):
>Input: activation a(t)
>Output: muscle force(t) = tendon force (t)
That is correct in the context of a conventional lumped-parameter equilibrium muscle model, where the mass of the muscle is ignored (as is typical in musculoskeletal modeling).
>One of the outputs of CMC are the "actuator controls", which are calculated after the Static Optimization step. Are these >"controls" referring to excitation u(t), or activation a(t)? I understood this seems to be confusing for a lot of people, so >hopefully this is going to be helpful for a lot of people?
CMC computes excitation profiles.
>I guess the calculation of forces(t) from the activation a(t) is included in the Forward Dynamic step. Is this correct?
I believe so.
>Is there >some documentation/paper to know more about this calculation/equations (musculotendon contraction dynamics)? I am >reading Zajac 1989, but I am still having some trouble with this part.
Actually I'm writing a paper right now on the topic that should appear in the ASME J. Biomech Eng that just deals with formulation and simulation of muscle. In the meantime, you can find a description of a conventional muscle model in the appendix of Thelen 2003 [2], or in a recent conference paper [3]. Additionally, complete derivations of the state equations for Thelen2003Muscle, Millard2012EquilibriumMuscle, and the Millard2012AccelerationMuscle can be found in the corresponding doxygen pages for OpenSim 3.0.
References
[1] Bellmare, F., Woods, JJ., Johansson,R., and Bigland-Ritchie,B (1983). Motor-unit discharge rates in maximal voluntary contractions of three human muscles. J. Neurophysiology(50), pp. 1380-1392.
[2] Thelen, DG.(2003), Adjustment of Muscle Mechanics Model Parameters to Simulate Dynamic Contractions in Older Adults. ASME Journal of Biomechanical Engineering (125).
[3] Millard, M., and Delp, S. (2012). A Computationally Efficient Muscle Model. Proceedings of the ASME 2012 Summer Bioengineering Conference. June 20-23, Farjardo, Peurto Rico, USA.
>Thanks!
>Ana
Good luck!
Matt
My name is Matt Millard, and I'm a post-doc at OpenSim. I've written the new muscle models you'll find in OpenSim 3.0 (Thelen2003Muscle, Millard2012EquilibriumMuscle and Millard2012AccelerationMuscle), and I'm working on a faster simulating muscle for the next release. I'll intersperse my response to your questions ...
>I have been reading about Musculotendon Actuator Dynamics, and this are composed of "activation dynamics" and >"musculotendon contraction dynamics", which are assumed to be decoupled. If I well understood, the inputs and the outputs >are:
>- Activation dynamics (first order process) :
>Input: excitation u(t)
>Output: activation a(t)
Correct.
Physically excitation is the electrical signal (an impulse train) that the nervous system sends to the muscle. The electrical excitation will cause a single muscle fiber to change chemically, and ultimately causing the fiber to generate a twitch of force (well modeled by a 2nd order critically damped system [1]). When the nervous system electrically stimulates many fibers, many twitches are created in parallel, generating larger forces. The force step-response of a muscle, when many of its fibers are recruited, looks very similar to a first-order step response, and so, a first order model has been used to model muscle `activation'.
>- Musculotendon contraction dynamics (first order process):
>Input: activation a(t)
>Output: muscle force(t) = tendon force (t)
That is correct in the context of a conventional lumped-parameter equilibrium muscle model, where the mass of the muscle is ignored (as is typical in musculoskeletal modeling).
>One of the outputs of CMC are the "actuator controls", which are calculated after the Static Optimization step. Are these >"controls" referring to excitation u(t), or activation a(t)? I understood this seems to be confusing for a lot of people, so >hopefully this is going to be helpful for a lot of people?
CMC computes excitation profiles.
>I guess the calculation of forces(t) from the activation a(t) is included in the Forward Dynamic step. Is this correct?
I believe so.
>Is there >some documentation/paper to know more about this calculation/equations (musculotendon contraction dynamics)? I am >reading Zajac 1989, but I am still having some trouble with this part.
Actually I'm writing a paper right now on the topic that should appear in the ASME J. Biomech Eng that just deals with formulation and simulation of muscle. In the meantime, you can find a description of a conventional muscle model in the appendix of Thelen 2003 [2], or in a recent conference paper [3]. Additionally, complete derivations of the state equations for Thelen2003Muscle, Millard2012EquilibriumMuscle, and the Millard2012AccelerationMuscle can be found in the corresponding doxygen pages for OpenSim 3.0.
References
[1] Bellmare, F., Woods, JJ., Johansson,R., and Bigland-Ritchie,B (1983). Motor-unit discharge rates in maximal voluntary contractions of three human muscles. J. Neurophysiology(50), pp. 1380-1392.
[2] Thelen, DG.(2003), Adjustment of Muscle Mechanics Model Parameters to Simulate Dynamic Contractions in Older Adults. ASME Journal of Biomechanical Engineering (125).
[3] Millard, M., and Delp, S. (2012). A Computationally Efficient Muscle Model. Proceedings of the ASME 2012 Summer Bioengineering Conference. June 20-23, Farjardo, Peurto Rico, USA.
>Thanks!
>Ana
Good luck!
Matt