simtk.org

You have been contacted by a member of Simtk.org.

From: Felicia Uhl
The e-mail was sent via the "contact" link on Simtk.org by the above member. It is not screened by Simtk.org or its staff.


*** MESSAGE BEGINS ***

Dear Mr. Fregly,
In the readme file of the GCC2 eTibia data there is the following equation given for medial and lateral contact forces:
F_medial = -0.876*F_ML + 0.494*F_SI - 0.0189*M_VV
F_lateral = 0.876*F_ML + 0.505*F_SI + 0.0189*M_VV
I wanted to know if these equations are also valid for the other datasets using the same instrumented tibia implant(GCC 3,5,6)?
In the respective paper those equations were evaluated for nromal gait, medial thrust gait, stair and a combination.
Are the equations only valid for those motions or for all motions?
Kind regards,
F. Uhl


-------------- Response below ----------------
Hi Felicia,

Subjects from GC 1 and GC 4 are the same (original EKnee design), and subjects from GC 2 and GC 6 are the same (newer eTibia design). Subjects from GC 3 and GC5 are different but also have the newer eTibia design.

All the medial and lateral contact force calibration curves we have generated are for multiple walking trials only - we have not done any calibrations for non-walking trials to see if the calibrations work well for tasks other than walking.

We have posted the eTibia calibration equations for the subjects from GC2/6, GC3, and GC5 on the Knee Grand Challenge website. Look in the "eTibia Data" folder under "Motion Data/Overground Gait Trials" for the README.txt file with the calibration equation in it. Each subject has his/her own coefficients in the calibration equations, so while the form of the equations is the same, the coefficient values are slightly different (though very similar still).

Hope that information helps.

Kind regards,

B.J.

BJ,

Thanks for the info the the eTibia calibration equations. I could not find any equations in the sixth grand challenge data set README.txt files. However, I was able to find the equations Felicia mentioned above in the 2nd grand challenge data set, so I assume these are OK to use for subject DM for the 6th grand challenge as well.

F_medial = -0.876*F_ML + 0.494*F_SI - 0.0189*M_VV
F_lateral = 0.876*F_ML + 0.505*F_SI + 0.0189*M_VV

I converted the eTibia data to N and Nmm and applied these calibration equations for the bouncy5 and smooth1 trials of the competition. However, the regression equations don't indicate which direction is positive for each load component (ie is F_ML positive in the medial direction?). This leads to some ambiguity because the provided coordinate system for the eTibia is for the left knee, but this subject has an instrumented right knee. If I apply the equations directly to the given eTibia data (converted to N and Nmm), I am seeing some strange results (measured lateral force is higher than medial force).

Can you give some guidance as to applying these equations to the data for the sixth competition?

Thanks,
Colin

Colin,

The regression terms F_ML, F_SI, and M_VV are the same as the terms Fx, Fz, and Ty provided in the contact force data CSV files, and in the eTibia coordinate system.
Also notice that the coordinate system for the eTibia is the same for a left knee and a right knee. If implanted in the right knee, +Fx would just happen to be pointing medially.
The bottom line is you can simply apply the regression equations to the provided data once it is converted to N and N mm. Also notice that this subject had a valgus alignment so the F_lateral/F_medial contact force ratio will be larger than usual.

I hope that answers your question,

Ilan