Probabilistic numerics for common neuroscience models. This is the code for the manuscript by Oesterle et al.
Contains all the notebooks to recreate the figures in the manuscript mentioned above. To run any of notebooks, you need to install NeuroProbNum first. This is described below.
Probabilistic solvers and neuroscience models, implemented in Python 3 and Cython.
First, download and install NeuroProbNum and its requirements.
git clone https://github.com/berenslab/neuroprobnum.git
cd neuroprobnum
pip3 install -r requirements.txt
pip3 install -e .To get an impression of how NeuroProbNum works open and run the tutorial notebook with Jupyter Notebook. Make sure that the Hodgkin-Huxley Cython model is compiled without an error in the first part of the notebook, otherwise the the model can not be simulated.
Contains the implemented probabilistic and deterministic solvers.
The probabilistic solvers were implemented based on the work by Conrad et al. 2017 and Abdulle and Garegnani 2020.
Contains all implemented models in python and Cython.
The STG model was implemented based on Prinz et al 2003 and Prinz et al 2004. The Izhikevich neuron model was implemented based on Izhikevich 2004.
Contains python files to conveniently generate, save and load data for all the model with different solvers.
Contains several python utility functions.
For example plot_utils.py is used to create consistent figures across the notebooks.