View Full Version : Third Grand Challenge Competition Data Released

09-08-2011, 08:26 PM
Dear Musculoskeletal Modeling Research Community,

We are glad to inform you that the data for the Third Grand Challenge Competition to Predict In Vivo Knee Loads has now been posted on the Simtk.org website (https://simtk.org/home/kneeloads). We apologize that we were unable to post the competition data sooner, but two bugs in the Vicon Nexus software prevented us from completing the data processing as quickly as we had anticipated. The competition will again be held at the ASME Summer Bioengineering Conference (http://www.asmeconferences.org/SBC2012/index.cfm), which will take place this year at the beautiful El Conquistador Resort in Puerto Rico.

An overview of the competition data, including the two gait trials to be used for the third competition, is contained in the file “Competition Data Description.pdf.” The data this year include pre-surgery MR data of the knee region, which may facilitate the determination of subject-specific muscle lines of action and muscle cross sectional areas. The post-surgery CT data is also of extremely high quality and spans the hip through the ankle, which allowed us to create a subject-specific geometric model that includes a hemi-pelvis, talus, and part of the calcaneus. Strength data and fluoroscopic motion data are still being processed and will be uploaded at a future date.

We are going to make three modifications to how the competition is run this year:
1) We will release the in vivo contact force measurements for the competition gait trials to all competitors shortly after abstract and contact force submission is closed. Competitors will then evaluate the sensitivity of their medial and lateral contact force predictions to model parameter values that they deem worthy of investigating, with the goal of determining which parameter values need to be changed to achieve medial and lateral contact force estimates consistent with the experimental data. During the competition workshop at the conference, competitors will be required to present their original blinded contact force predictions, their updated unblinded contact force predictions, and a brief discussion of what they needed to change to get their predictions to match the experimental measurements more closely.
2) We are going to calculate accuracy using RMS errors and R^2 values so that predictions that follow the correct trends, but not necessarily the correct values, will be given greater consideration. Accuracy of competition submissions will still be evaluated based solely on the blinded predictions.
3) We will allow previous competition winners to participate this year to encourage continued innovation in the development and assessment of new musculoskeletal modeling methods.

Please note that we have also released the instrumented knee data for the gait trials from the first two competitions. Please cite the reference indicated on the competition website in any publications that use any of the competition data or models.

We hope to see broad participation from the musculoskeletal modeling research community for this year’s competition. Please post any questions you have about the competition on the Public Forum of the Simtk.org competition website.

Best regards,

B.J. Fregly, University of Florida
Darryl D’Lima, Shiley Center at Scripps Clinic
Thor Besier, University of Auckland