I am a user of AnyBody version 7.3 and am currently working on a squat activity using the marker-based full-body motion capture model (AMMR 2.3.1 MoCapModel). I have encountered an issue where the Tibialis Anterior muscle activity is almost zero (e^-13) during the activity.
I have also tested the squat example provided in the AMMR, and the Tibialis Anterior muscle activity remains almost zero there as well.
Could you please provide some advice or suggestions on how to address this problem? If you need additional information or if my question needs further clarification, please let me know.
Normally, the tibialis anterior is not a primary muscle involved in squats. It's role is more of a stabilizer. Unfortunately, the stabilizing effect can't really be captured so well in the muscle recruitment formulation.
One scenario in which the tibialis anterior will be activated in the model is if you simulate a deep squat and your weight tends to be more on the posterior side. I think the mocap squat model in the AMMR doesn't do a deep squat. You can instead see the synthetic squat model to simulate a deeper squat motion. Then you would see activation of the tibialis anterior.
Dear Dave,
Thank you very much for your feedback. I appreciate your insights and suggestions.
I have adjusted the maximum knee angle from 100 to 140 degrees, which has resulted in activation of the Tibialis Anterior muscle. However, the activation periods and activation graphs from the simulation still do not align with the experimental results.
Additionally, I have analyzed deep squats using a marker-based full-body motion capture model (AMMR 2.3.1 MoCapModel) with data from 10 subjects (averaged maximum knee angle of 137 degrees). Despite this, the Tibialis Anterior muscle activity remains almost zero.
Is it possible to simulate the model using experimental EMG data and marker data as input?Any guidance or recommendations you can provide on how to integrate EMG data into the simulation would be greatly appreciated.
You could try playing around with different muscle recruitment, but it most likely won't solve the issue. I believe you would have tried that already.
You can use the experimental EMG as an input. You can use the class AnyMuscleActivityBound to create upper and lower bounds on the muscle. You can feed in the EMG data through an interpolation function and add some offset to create an upper and lower bound on the same muscle.
Hi Dave,
I’m grateful for your time and assistance with this.
I had already explored different muscle recruitment strategies, but as expected, they did not resolve the issue. Following your recommendation, I used the AnyMuscleActivityBound class with my experimental EMG data. However, it resulted in constant activation of the Tibialis Anterior muscle throughout the activity, which was unexpected. To address this, I used the AnyMuscleActivityConstraint class based on my experimental results. My first implementation was as follows:
AnyMuscleActivityConstraint TA1 = {
AnyMuscle &TA1 = Main.HumanModel.BodyModel.Right.Leg.Mus.TibialisAnterior1;
Coef = {{1.0}};
AnyFunPolynomial fun_mus_act = {
PolyCoef = {{0, 0.0197, -0.0002}}; // Obtained from experimental EMG
};
Const = fun_mus_act((100/nStep)*iStep); // Cycle 0 to 100
CType = NonNegative;
};
Unfortunately, this did not influence the Tibialis Anterior activation. I also experimented with another approach using the following logic:
AnyMuscleActivityConstraint TA1 = {
AnyMuscle &TA1 = Main.HumanModel.BodyModel.Right.Leg.Mus.TibialisAnterior1;
Coef = {{-1.0}};
Const = {iffun(orfun(ltfun(t,15.0),gteqfun(t,16.5)),0.0,0.6)};
CType = NonNegative;
};
Even with this constraint,the Tibialis Anterior activation remained unchanged. I tried several other functions using a trial-and-error method, and while some values did affect muscle activation, the results lacked a clear logic or pattern.
Am I potentially missing something in the implementation, or is there a specific approach I should follow to achieve the desired results?