Welcome to COMOB

The project for collaborative modelling of the brain grew out of the 2022 Cosyne tutorial. The aim is to work together, openly on a large scale computational brain modelling project.

Our first project (of hopefully many) is to use surrogate gradient descent methods to train spiking neural network models of sound localization. You can check out the project here.

How to get involved

We do our work in public on GitHub. Anyone can view the code, figures, etc. You can also join our discussion pages, create and comment on issues and so on. If you want to contribute code to a project, please submit a pull request with your contribution. Once an existing member has reviewed your pull request, you’ll be added to the organization.

Each project has an associated repository, and everyone works on Jupyter notebooks in their own branch of this repository, periodically merging their changes into the main repository. From these notebooks, we occasionally refactor common code into Python packages that are maintained by the most active members of the project. Changes to these packages get reviewed before being merged. We generate a JupyterBook from the notebooks in the repo so that you can always check out a project’s webpage to see the current state of the project including all notebooks.

We will occasionally organise local hackathons. Please feel free to add yourself to our ethermap so that we can decide where are good hub locations to hold the hackathons.

When a project matures, we will attempt to organise the mass of results into a coherent, linear paper. Project organizers will get first/last author positions as a reward for their extra work, and the remaining authors will be randomly ordered. Everyone who contributed code/text/figure gets to be an author regardless of whether or not their contribution is directly included in the final paper.

If you want to get involved, go straight to our current, first project on spiking neural networks and see how you can contribute.

Why are we doing this?

We were inspired by the Polymath Project which managed to successfully solve previously unsolved mathematical problems in a similar open fashion. Science ought to be a collaborative enterprise, not a competition. This project is an experiment to see if this approach can be successful in the field of computational neuroscience, which, like mathematics, only requires a computer and access to the internet to contribute to.