/blueshell-py

Implements an efficient, biologically plausible spiking neural network in a few lines of Python.

Primary LanguagePythonMIT LicenseMIT

Blue Shell

Implementation of a spiking neural network. Uses neurons described by Eugene Izhikevich in:

E. M. Izhikevich, "Simple model of spiking neurons," in IEEE Transactions on Neural Networks, vol. 14, no. 6, pp. 1569-1572, Nov. 2003, doi: 10.1109/TNN.2003.820440.

Model can be run via main.py. Fiddle with model parameters by updating pool-config.yaml.

Currently no learning is implemented. A basic simulation can be run which will produce a plot of the neuron voltages over time, like this:

Voltages

Other resources

This repo is for demo/educational purposes. If you haven't already, check out the more fully-featured spiking net simulators for Python that already exist. Some good places to start are BindsNET and Brian.