Meeting Abstract
Current models perform poorly at predicting muscle force during dynamic movements, especially for fast movements. We used a new muscle model to predict muscle force in running guinea fowl. Lateral gastrocnemius (LG) length, activation and force were measured by Daley et al. (2011). Data were recorded while guinea fowl ran on a treadmill modified so that obstacles were encountered at various phases of the step cycle. Trials included level running, 5 and 7 cm obstacles, and speeds from 1.8 – 4.5 m/s. The new model includes a damped contractile element (CE) in series with a spring representing extracellular matrix. The CE is in series and parallel with a damped titin spring. The titin spring wraps around a pulley that represents thin filaments. The CE rotates the pulley, adjusting the length and stiffness of the titin spring. The pulley translates when applied forces stretch or shorten the muscle, which changes the length and force of the titin spring. Free parameters include two spring constants, three damping constants, and an activation factor that varied from trial to trial. Parameters were optimized locally and globally using a high-performance computer. Muscle length and activation are the model inputs, and force is predicted in each time step. Results show that the model accurately predicts in vivo forces during perturbed and level gaits (average R2 = 0.67-0.88). Data and simulations show that force is related, not to onset or amplitude of EMG, but rather to: 1) onset and magnitude of the stretch that occurs when the foot hits the ground during active shortening; and 2) muscle length at ground contact. Results demonstrate how adding titin can improve prediction of in vivo forces during running over level terrain as well as when negotiating obstacles, and also suggests that some muscles depend on small stretches during active shortening for force production.
