Foundation models for neuroscience tutorial

Tutorial repository for the Cajal school in machine learning for neuroscience.

Implements an NDT-1 style encoder from scratch (Ye and Pandarinath, 2021).

Running the tutorial in colab (recommended)

Compiling the tutorial from the .py source

cd into the tutorials folder, and run:

jupytext tutorial_source.py --to ipynb -o Tutorial.ipynb

Working with the scripts

• Fork this repository
• Clone your fork
• Create a fresh conda environment with e.g. Python 3.9, conda create --name fmn-tutorial python=3.11
• cd into the directory and pip install -e .

There are three scripts available in the scripts folder:

• preprocess_data.py: creates pickle files from DANDI sources
• train_autoencoder.py: trains auto-encoders, similar to the tutorial, but on the command line, with better logging
• predict_behavior.py: trains a decoder for behavior from scratch or from a pretrained checkpoint