microcircuit 0.1rc2

PyNEST implementation of the cortical microcircuit model of Potjans & Diesmann (2014).

PyNEST implementation

Installing the python package microcircuit

The PyNEST implementation of the model is provided in the form of a python package microcircuit.

We recommend installing the python package inside a python environment:

• Create a python environment

  python -m venv venv
  

• Activate the python environment:

  source venv/bin/activate
  

• Update pip:

  pip install -U pip
  

The microcircuit python package can be installed by

git clone https://github.com/inm-6/microcircuit-PD14-model
pip install microcircuit-PD14-model/PyNEST

or, without cloning the entire repository, by

pip install "git+https://github.com/INM-6/microcircuit-PD14-model.git/#egg=microcircuit&subdirectory=PyNEST"

Software requirements

• NEST (see NEST installation)

  We recommend installing NEST locally within the virtual environment:

  git clone https://github.com/nest/nest-simulator
  ## for a specific <VERSION> (e.g. <VERSION>=v3.9), use
  ## git clone --depth 1 --branch <VERSION> https://github.com/nest/nest-simulator
  cd nest-simulator
  mkdir build
  cd build
  pip install -r ../requirements_pynest.txt
  cmake ..
  make
  make install
  

• Python 3.x

• docopt-ng, matplotlib, numpy, psutil, ruamel.yaml, scipy (handled by python package dependencies)

Testing

Executing

pytest

runs the unit test(s) in microcircuit-PD14-model/PyNEST/tests.

Usage

After installation, the microcircuit python package can be imported in a python application using

import microcircuit

See this examples for a more detailed illustrations of how the package can be used.

Memory requirements

scaling factor (= N_scaling = K_scaling)	Memory
0.1 (default)	490 MB
0.2	1200 MB
0.5	4400 MB
1	14 GB

Performance benchmarking

Recent performance benchmarking results for the microcircuit model can be found here.

Implementation details

This implementation uses the iaf_psc_exp neuron and the static_synapse synapse models provided in [NEST]. The network is connected according to the fixed_total_number connection rule in NEST. The neuron dynamics is integrated in a time-driven manner using exact integration with a simulation step size sim_resolution (Rotter & Diesmann, 1999).

The PyNEST implementation runs with NEST 3.9 (Terhorst et al., 2025).

Simulation parameters (defaults)

Name	Value	Description
sim_resolution	0.1 ms	simulation time resolution (duration of one simulation step)
t_presim	500 ms	duration of pre-simulation phase (warm-up)
t_sim	1000 ms	duration of simulation phase
rec_dev	spike_recorder	recording device

References

Rotter & Diesmann (1999), Exact digital simulation of time-invariant linear systems with applications to neuronal modeling. Biological Cybernetics 81(5-6):381-402. doi:10.1007/s004220050570

Terhorst et al. (2025). NEST 3.9. Zenodo. doi:10.5281/zenodo.17036827

License

This project is licensed under GNU General Public License v2.0 or later. For details, see here.