M3 2134
Speaker
Dr. Wilten Nicola | Bioengineering, Imperial College London
Title
Large network approaches in plastic spiking neural networks: mean-field analysis, inverse mean-field methods, and macroscopic chaos
Abstract
We often want to determine how large networks of neurons behave, or control a network of neurons so that the network performs a specific output or dynamics. In this work, we discuss mean-field methods for predicting the dynamics for large networks of spiking neurons with plastic weights. Additionally, we demonstrate some new results on determining the weight matrix for a network of neurons such that the macroscopic dynamics of the network is set to a specific dynamical system. We use the Neural Engineering Framework and the FORCE method to resolve the weight matrices that fix the dynamics for a variety of neuron and network types.