The OpenSim Project and the National Center for Simulation in Rehabilitation Research (NCSRR) at Stanford invite you to join our next webinar, featuring Adrian Lai from Simon Fraser University and Allison Arnold from Harvard University.


Title: Why Are Antagonists Co-activated in My Simulations? Simulating Cycling and Other “High Flexion” Tasks
Speaker: Adrian Lai from Simon Fraser University and Allison Arnold from Harvard University
Time: Thursday, October 5, 2017 at 10:00 a.m. Pacific Daylight Time


Pedalling is a popular task for exercise and rehabilitation, and pedalling on an instrumented ergometer provides a meaningful paradigm for lab-based studies of neuromuscular function. Like walking, pedalling involves reciprocal flexion/extension of the hips, knees and ankles; however, pedaling avoids the loading associated with repeated foot-ground contact, and it minimizes balance demands. For these reasons, the biomechanics of pedalling has been studied for decades.

We were surprised, therefore, when our initial attempts to generate subject-specific simulations of pedalling, in OpenSim, failed to reproduce subjects’ measured EMG recordings. This webinar is motivated by our efforts to understand why, and our goal is to share practical tips for analyzing tasks, like pedaling, that involve substantial hip and knee flexion.

We found that existing musculoskeletal models have three major limitations. First, they overestimate the passive fiber forces developed by the hip and knee extensors and these large passive forces can lead to anomalous compensatory muscle activity in simulations. Another limitation is that the 3D paths of some muscles are poorly represented over the joint angles achieved by subjects during pedalling. Yet another limitation is that the gastrocnemii and other muscles become too short to generate active force during portions of the cycle that involve substantial hip and/or knee flexion.

Using illustrative examples from pedalling, walking and running, we will show that the anomalous co-activation of antagonist muscles, observed in our initial simulations, can be resolved if the passive fiber forces generated by the underlying model are diminished.

Our refined model is available from ( and is suitable for generating simulations of movements involving up to 120° of hip flexion and 140° of knee flexion.

Visit our website for more information and registration. The website also includes links to recordings of past webinars: The OpenSim Webinar Series is funded by the NIH National Center for Simulation in Rehabilitation Research (NCSRR). Find out more about the NCSRR and the webinar series by visiting our website