This repository contains research code and figures of the "SBI for connectomics" paper (link to paper). In the paper, we show how to use simulation-based inference (SBI) to infer parameters of computational models used in connectomics.
The repository is based on a small python package called consbi that allows to simulate different wiring rules in the structural model of the rat barrel cortex.
Additionally, it contains jupyter notebooks with tutorials and code for reproducing the figures shown in the paper.
Binary files with data and results for the work presented in the paper are stored using git-lfs.
The repository for running SBI including detailed tutorials is located at https://github.com/mackelab/sbi.
Please reach out and create an issue if you have any questions or encounter problems with using this repository.
Google colab: You can play around with the tutorials in the browser using the following google colab link:
Locally: clone the repository, cd into the local sbi-for-connectomics folder and run pip install -e .
The repository contains a number of large files that are provided via git-lfs.
See the git-lfs documentation for installation and general info, https://git-lfs.com.
Once git-lfs is installed locally, one can pull the large files contained in the repository just by cloning or pulling it.