Neural Networks Estimate Muscle Force in Dynamic Conditions Better than Hill-type Muscle Models

Abstract

Hill-type muscle models are widely used, even though they do not accurately represent certain muscle mechanics. We explored neural networks to develop new muscle models. We trained neural networks to estimate muscle force from activation, muscle length, and muscle velocity. Training data was recorded using sonomicrometry, electromyography, and a tendon buckle on two muscles of guinea fowl. First, we compared the neural network to a Hill-type muscle model, using the same data for network training and model optimization. Second, we trained a network on a large dataset, in a more realistic machine learning scenario. We found that the neural networks generally yielded higher correlations and lower errors than Hill-type muscle models. Our neural network performed better when estimating forces on the muscle used for training of another bird than on a different muscle of the same bird, which could be explained by inaccuracies in activation and force scaling. We created a force-length and force-velocity relationship for the neural network and found that both amplification factors were underestimated and that both relationships were not replicated well outside of the training data distribution. We conclude that neural networks could provide an accurate alternative to Hill-type muscle models given a suitable training dataset.