Embedded Files
Significance
Analytical methods of studying these multi-neuron networks are difficult to utilize due to their nonlinear nature and high dimensionality, so a computational study such as ours is essential in understanding the dynamics of such a system. The database generated from our parameter searching simulation can be queried to find models that have desirable properties for application in further engineering studies and dynamical models. In vivo, central pattern generators are merely a single part in a larger system comprising of peripheral circuits, muscle mechanics and sensory feedback, such as the one pictured below. Closing the loop in the system with sensory feedback can significantly change the properties of the CPG itself, which warrants the development of more complex models and future studies, and the functional models generated by our simulations is a stepping stone in that direction.
Figure 1: Gait Control System for Locomotion, Modified from Prochazka, Ellaway [1]
Future Directions and Applications
Our work can be expanded on to include CPGs with different neuron and synapse models (Hodgkin-Huxley, Fitzhugh-Nagumo, Hindmarsh-Rose, etc.) as well as more complex topologies as opposed to a linear chain of HCOs to further study how topology affects the activity of the network. Furthermore the effects of different input currents to the system can be studied as we only implemented a noisy step current that had a global effect on the network. The aggregation of spiking neural network data can promote the development of ML models that can predict the properties of a spiking neural network as well as generate new network topologies with desirable properties. Functional central pattern generators have seen application in biomimetic robotics as efficient control mechanisms for rhythmic movements and show great promise for use in prosthetics and in the near future can interface with biological neurons in more advanced neuroprosthetics.
Simulation of human gait controlled by CPG in OpenSim [2]
Schematic of therapeutic exoskeleton [3]
References
Prochazka, A., & Ellaway, P. (2012). Sensory Systems in the Control of Movement. Wiley Online Library, 2(4). doi:https://doi.org/10.1002/cphy.c100086
Shachykov, A., Shuliak, O., & Henaff, P. (2019). Closed-loop Central Pattern Generator Control of Human Gaits in OpenSim Simulator. 2019 International Joint Conference on Neural Networks (IJCNN), 1–8. https://doi.org/10.1109/IJCNN.2019.8852163
Fang, J., Ren, Y., & Zhang, D. (2014). A robotic exoskeleton for lower limb rehabilitation controlled by central pattern generator. 2014 IEEE International Conference on Robotics and Biomimetics (ROBIO 2014), 814–818. https://doi.org/10.1109/ROBIO.2014.7090432
Group 15: CPG Implementation in Software
Page Leader: Aditya Athota
Page updated
Report abuse