This code accompanies the paper Multi-scale approaches for high-speed imaging and analysis of large neural populations [PLoS Comput Biol. 2017; 13(8):e1005685.]
The scripts were tested on Linux and MacOS with a typical numerical/scientific Python 2.7 installation, e.g. using Anaconda or Canopy.
The scripts make use of CaImAn (formerly Constrained_NMF), hence all its dependencies have to be met. To avoid issues due to newer versions of CaImAn, we recommend to clone/download it from this branch, which further includes minor extensions to produce Fig 8. Please make sure to add the package Constrained_NMF to your $PYTHONPATH, so that the scripts in this repo find it.
The scripts to produce the figures and table have names obvious from the PLoS Comput Biol paper.
They can be run with python table1.py and python fig[1-8].py to show the figures during code execution. If a (sub)directory name is provided as argument, e.g. python fig1.py fig, figures are saved in the directory, e.g. fig, if it exists. For the two-photon data, fig[4-8], you need to execute python run_2P.py first, which could take few hours and saves the results in the subdirectory results.