Short-term neuronal and synaptic plasticity act in synergy for deviance detection in spiking networks
Code for "Short-term neuronal and synaptic plasticity act in synergy for deviance detection in spiking networks", Kern & Chao 2023
All code expects an environment with the dependencies listed in requirements.txt.
Set up conf/params.py and conf/<a config module>.py first. The former describes the model and paradigm, while the latter controls details of the simulation. See conf/isi5_500.py for the example used in our paper.
Then, from the base directory, run python isi.py conf/<a config module>.py to run the requested simulations and save the raw data to disk.
The raw data is processed in various, very memory-intensive ways. These are scripted through the process_*.py scripts, which must all be run to display all figures.
The convenience script process_all.py does all of them in the necessary order, but may take a long time to complete.
Invoke each processing script with python process_[...].py conf/<a config module>.py.
Note that the configuration can represent a subset of the data present in order to do a partial analysis, e.g. with N_networks set to a lower number.
If you've run multiple templates or ISIs, process_all can select from these; to do so, run it as python process_all.py <conf> <ISI-in-ms> <template-idx>.
Absent such selection, the scripts will process the first ISI mentioned in the configuration, and the first template, only.
Processing multiple ISIs or templates simultaneously is currently not supported.
Finally, run each of the Fig_*.ipynb* notebooks. Figure PDFs will be placed in paper-1/.
For several figures, both the processed data (for stats across networks) and a demonstrator network's raw data (as examples) are used. To select the sample network/stimulus, change the relevant entries in demonstrator.py.