General Questions
Posted: Thu Mar 14, 2019 1:42 pm
The following questions were received from Rodrigo Andre Bonacho Mateus <rodrigo.mateus@ist.utl.pt>
I have performed previous works in musculoskeletal modelling before,
however this is the first time I am analyzing crouch gait and I came
across your project and the subsequent articles you published. While I
was checking the data, I had some doubts I wanted to ask you.
Firstly, was the data given in the project .zip file used only for the
article: "Muscle contributions to support and progression during
single-limb stance in crouch gait."? Or did you use it for the later
articles published, such as "Muscle contributions to vertical and
fore-aft accelerations are altered in subjects with crouch gait."?
sides individually, when it comes to crouch gait? If so, which
cautionary measures must be taken to perform this?
weights of the coordinates, specially reagarding the default tracking
weights provided by OpenSim. What was your line of thouhgt to perform
such changes?
- Why did you increase the optimal forces of the residual actuators?
- There were some features you inserted in the setup files, listed
below, that caught my eye and I wanted to ask yoy why did you put them
in.
1. <use_curvature_filter> true </use_curvature_filter>
2. <use_reflexes> false </use_reflexes>
3. <optimizer_algorithm> cfsqp</optimizer_algorithm> (In this case, why
did you change the optimization algorithm?)
I also noted that you did not check the option to compute equilibrium
values for states other than the coordinates or speeds
(<solve_for_equilibrium_for_auxiliary_states> false
</solve_for_equilibrium_for_auxiliary_states>). Is there any specific
reason for this?
You also contrain specific muscles at certain instants in time, as it can be seen in the ControlConstraints file. Why?
I have performed previous works in musculoskeletal modelling before,
however this is the first time I am analyzing crouch gait and I came
across your project and the subsequent articles you published. While I
was checking the data, I had some doubts I wanted to ask you.
Firstly, was the data given in the project .zip file used only for the
article: "Muscle contributions to support and progression during
single-limb stance in crouch gait."? Or did you use it for the later
articles published, such as "Muscle contributions to vertical and
fore-aft accelerations are altered in subjects with crouch gait."?
- There are several .zip files provided on the website and the primary article for each is listed under the download link. There weresome participants that were included in both "Muscle contributions to support and progression during single-limb stance in crouch gait" and "Muscle contributions to vertical and fore-aft accelerations are altered in subjects with crouch gait" but if you read the papers you'll note that the first only included mild crouch gait and single-limb stance, while the later required more force plate data and also included individuals with more crouch severity.
sides individually, when it comes to crouch gait? If so, which
cautionary measures must be taken to perform this?
- For considering right versus left, it depends on the child and their gait symmetry.
weights of the coordinates, specially reagarding the default tracking
weights provided by OpenSim. What was your line of thouhgt to perform
such changes?
- For RRA we based the weights upon iterative testing, but the big difference was that we had low weights for the torso/lumbar because these subjects had limited upper-extremity markers so we knew we had a lot of error in torso kinematics.
- Why did you increase the optimal forces of the residual actuators?
- There were some features you inserted in the setup files, listed
below, that caught my eye and I wanted to ask yoy why did you put them
in.
1. <use_curvature_filter> true </use_curvature_filter>
2. <use_reflexes> false </use_reflexes>
3. <optimizer_algorithm> cfsqp</optimizer_algorithm> (In this case, why
did you change the optimization algorithm?)
I also noted that you did not check the option to compute equilibrium
values for states other than the coordinates or speeds
(<solve_for_equilibrium_for_auxiliary_states> false
</solve_for_equilibrium_for_auxiliary_states>). Is there any specific
reason for this?
You also contrain specific muscles at certain instants in time, as it can be seen in the ControlConstraints file. Why?
- Optimal force of residual actuators can impact whether or not it can find a solution. This research was done in an older version of OpenSim (almost 10 years ago now!) so the use_curvature and use_reflexes were standard in the setup files for those versions. At the time, we used cfsqp as ipopt was just getting implement in OpenSim and wasn't as robust as cfsqp in finding solutions. The equilibrium for auxillary states was the default at the time and not something I considered. As described in the manuscripts, ControlConstraints were applied to help get better alignment with EMG.