Neuromechanics of movement control
Movement is the dynamic expression of our brain that can effect the environment through coordinated musculoskeletal actions. The neural and mechanical systems coevolved to diversify the ability to move efficiently, fast, with high accuracy and diversity. While the main elements are well-described the compounded action of multiple and overlapping mechanisms remains to be poorly understood. The young field of computational neuroscience is built on methodology from computer science and engineering to describe the complexity of experimental datasets. These computational approaches are shared across several modern applied fields of neural engineering, robotics, and active prosthetics.
In this course we will discuss the bottom-up organization of neuromechanical system and build computational models of musculoskeletal elements and neural pathways that control them. In particular, we will build models of muscles and tendons, use mechanical simulation environment to create musculoskeletal models, and add models of sensory feedback from the main proprioceptors. The sensory feedback will be combined with the intrinsic spinal mechanisms and the descending drive. We will use both the inverse and forward dynamic modeling combined with optimization techniques to gain insight into the functional interactions between these components.
We put special focus on learning from rewarding or punishing consequences of self-generated behavior - reinforcement learning. In reinforcement learning, only sparse outcome feedback about success or failure of its own actions is provided to the organism, which constitutes a much harder learning problem than usual supervised learning setting that is often employed in machine learning. In the project, we will work on implementing different neural circuits that are able to perform reinforcement learning with spiking neurons. As a demonstration, we target a spiking neural network that can learn a classical arcade pong game just by experiencing ball hits and misses without providing any further prior knowledge how to control the game.
- Introduction to Matlab
- Passive Walker control: passive walker model (opensim)
- IDM neuromusculoskeletal models of locomotion: lower-extremity gait model (model1.2 opensim), upper-extremity model (opensim)
- FDM neuromusculoskeletal models of locomotion: lower-extremity gait model (model1.2 opensim), upper-extremity model (opensim)
- Central Pattern Generator
Dr. Sergiy Yakovenko, West Virginia University School of Medicine, WV, USA