Embedded Files
Team members: Aditya Athota, Blair Caccam, Ryan Kochis, Arjun Ray
Mentor: Dr. Frederic Broccard, Institute for Neural Computation
Abstract
Neuromorphic systems attempt to mimic the function, structure, and plasticity of biological neurons. The field of neuromorphic engineering is rapidly evolving and shows great promise for the future of medicine and healthcare. However, relatively little is known about biological neurons’ function in vivo regarding the network structure they build and the parameters that produce their spiking behaviors. Current experimental methods to find these values are very laborious. Here we design a MATLAB pipeline to study statistics of half-center oscillators (HCO), a building block of many central pattern generators (CPGs) that produces rhythmic motor patterns in many animals. Our pipeline is adaptable to explore the parameter space of any mathematical neuron model. We implemented the Izhikevich and Morris-Lecar models to create a database covering all biologically plausible parameters with retrievable spiking statistics, resulting in more than 5 million simulations for each model.
Problem Statement
There is currently no good model of chained Central Pattern Generators (CPG) to observe and study connectivity between interacting oscillator modules and feedback mechanisms. Neuroscientists need biologically accurate models of CPGs that allow observation of their behavior in virtual environments or implemented in hardware.
Other Materials
Project Poster
187_Final_Video.mp4Project Video
Group 15: CPG Implementation in Software
Page Leader: Ryan Kochis
Page updated
Report abuse