simtk.org

Dear OpenSim Team and Users,

When conducting human gait simulations in OpenSim, we usually apply the ground reaction forces collected by the force plates to the calcaneus, which is feasible when the foot modeling requirements are not high. However, my current research focus is on the foot, and I need to calculate the moment of force at the metatarsophalangeal joint. I have tried separately applying the ground reaction force to the calcaneus and the toes, and there is a significant difference in the moment of force at the metatarsophalangeal joint between the two cases. I believe that applying the ground reaction force to the calcaneus and toes respectively according to the actual force distribution is the correct approach, but I have encountered many issues during the specific implementation process:

1. When the swinging leg just contacts the force plate and the heel touches the ground, the ground reaction force needs to be applied to the calcaneus. As the center of mass moves forward, the ground reaction force gradually shifts from the calcaneus to the toes. How can I determine the time when the ground reaction force shifts from the calcaneus to the toes?

2. After solving the first problem, how can I split the force plate data into two parts, with the first part acting on the calcaneus and the second part acting on the toes?

3. Apart from applying the ground reaction force to the calcaneus and toes respectively based on the actual force distribution, are there any other methods to accurately obtain the moment of force at the metatarsophalangeal joint?

Looking forward to your reply.

Regards.

Lan

Hi Lan,

I have seen two primary ways of dividing up the GRF between the calcaneus and metatarsals, and while typing this out I had an idea for a possible third approach.

The first method treats the MTP joint as a single hinge. When the center of pressure (CoP) is posterior to the axis of rotation of this hinge on the foot, the entire GRF is assigned to the calcaneus (really, the whole mid/rearfoot) segment. When the CoP is anterior to this axis, the entire GRF is assigned to the metatarsal segment. Usually the inertial properties of the metatarsals are assumed to be negligably small. This means that the MPT joint torque is zero when the GRF is posterior to the MTP joint. Papers that take this approach usually reference Stefanyshyn and Nigg 1997 as the canonical paper.

The second method uses a planar pressure system (e.g. Loadsol or Matscan) to parse out how much of the total GRF to assign to the various segments of the foot. This method has the advantage of allowing more than two foot segments, but of course requires collecting both GRF and plantar pressure data (and synchronizing them, which is often the hardest part).

I suppose a third method could be to use Moco and set up your model so it tracks the foot kinematics and the GRF, and let the optimizer "figure out" how to parse out the force among the various segments of the foot. Antoine Falisse's paper from last year might be a good read if you are interested in that approach.