This is a legacy repository dedicated to the reproducibility of our findings from Braganza et al., "Quantitative properties of a feedback circuit predict frequency-dependent pattern separation" eLife, 2020 http://dx.doi.org/10.7554/eLife.53148 that were generated with pyDentate. If you want to use or build on pyDentate it is highly recommended to use the active pyDentate repository.

Follow these steps to run this legacy version of pyDentate:

1. Install Anaconda
2. Install NEURON

   There are many ways to install NEURON. I prefer the conda-forge distribution

   conda install -c conda-forge/label/cf201901 neuron

3. Install elephant

   pip install elephant

4. Compile the NEURON mechanisms in /mechs_7-6
5. Download the pyDentateeLife2020 repository and unpack
6. Open paradigm_pattern_separation.py and add the path to your compiled mechanisms to dll_files variable
7. Run paradigm_pattern_separation.py

paradigm_ files are scripts that create one or more networks, run them and save the results. net_ files contain the network classes that are used by scripts to create the networks. Other files define for example the input pattern generator (burst_generator_inhomogeneous_poisson.py) or the single cells specifications (granulecell.py with granuelcellparams.txt). ouropy provides a backend for the networks by defining the logic to create and connect networks. I custom made ouropy to move away from NEURONS Section logic to a higher-level logic of cells, populations, connections and finally networks. Networks in net_ files inherit from ouropys GenNetwork class. For more details on pyDentate make sure to look at the active pyDentate repository.

Daniel Müller-Komorowska - Institute of Experimental Epileptology and Cognition Research