This repository contains code related to an article in preparation on the NineML model description language (http://nineml.net).

A Docker image with everything already installed is available from Docker Hub: https://hub.docker.com/r/nineml/nineml_demo_2016/

Once you have Docker installed:

docker pull nineml/nineml_demo_2016
mkdir myresults
cd myresults
docker run -v `pwd`:/home/docker/projects/nineml_demo_2016/results -t -i nineml/nineml_demo_2016 /bin/bash

This will start a Docker container running the bash shell, in which you can run simulations as described below. The directory myresults on your host machine will be shared with the Docker container; data and image files generated by the simulations will be written there.

cd ~/projects/nineml_demo_2016/9ML-toolkit/examples/Brunel2000
9ML-network -m crk3 brunel_network_alpha_AI.xml  # creates Sim_brunel_network_alpha_AI
./Sim_brunel_network_alpha_AI -d 1200.0 --timestep=0.01 -s "All neurons"

Data are recorded to brunel_network_alpha_AI.dat. Each line consists of the time followed by the indices of the neurons which spiked during that time step.

Or, based on the XML files from the NineML Catalog:

cd ~/projects/nineml_demo_2016/catalog/xml/network/Brunel2000/
9ML-network -m crk3 AI.xml
cd ~/projects/nineml_demo_2016/results
~/projects/nineml_demo_2016/catalog/xml/network/Brunel2000/Sim_AI -d 1200.0 --timestep=0.01 -s "All"

To run a single simulation and plot a figure:

cd ~/projects/nineml_demo_2016/python/brunel2000
python run.py --plot-figure nineml parameters/AI.yml

Replace "nineml" with one of "nest", "pyNN.nest", "pyNN.neuron" or "ninemlpartial" to run one of the other implementations. This produces a figure, showing spike rasters and membrane potentials, as "results/brunel_network_alpha_AI_nineml<timestamp>.png"

To generate a figure similar to Figure 8 in Brunel (2000), run:

python run.py nineml parameters/AI.yml
python run.py nineml parameters/SR.yml
python run.py nineml parameters/SIfast.yml
python run.py nineml parameters/SIslow.yml

Then to plot spike rasters and the instantaneous firing rate, run:

python four_panel_figure.py <datafile1> <datafile2> <datafile3> <datafile4> -o output.png

where <datafile[1-4]> are the ".h5" files produced by the four simulations.

The results and figures will be created in the directory ~/projects/nineml_demo_2016/results directory in the Docker image, which should be mapped to the myresults directory on your host

For more information, contact andrew.davison@unic.cnrs-gif.fr