NeuroBoostedTrees

Resolving neuronal population code and coordination with gradient boosted trees.

This package includes essential tools to analyze neuronal data with gradient boosted trees (using the package XGBoost). More specifically, it can be used to decode signals from population data and study coordination between neuronal populations in a more unbiased way than with typical linear tools (e.g. spike train cross-correlation).

The package was used for the analyses presented in the following paper: Brain-state invariant thalamo-cortical coordination revealed by non-linear encoders.

We put together a Jupyter Notebook to illustrate how to use the package.