Connectome-constrained models of the fruit fly visual system
Tutorial material for working with FlyVis models at the Winter School on Computational Approaches in Biological Sciences (SJCABS).
| Tutorial | Topic | Description |
|---|---|---|
| 0 | Using the model | Load a pretrained model, stimulate it, plot neural responses, and update parameters with gradient descent |
| 1 | Mechanism discovery | Apply UMAP dimensionality reduction and Gaussian mixture clustering to discover computational strategies |
| 2 | Deep stimulus design | Find optimal naturalistic stimuli and generate artificial optimal stimuli using gradient-based optimization |
Run this in the first cell of your notebook:
!pip install "git+https://github.com/sjcabs/flyvis_tutorial.git"
!flyvis download-pretrainedRequires Python 3.11.
# Clone the repository
git clone https://github.com/sjcabs/flyvis_tutorial.git
cd flyvis_tutorial
# Create a conda environment (recommended)
conda create -n flyvis_tutorial python=3.11
conda activate flyvis_tutorial
# Install the package
pip install -e .
# Download pretrained models
flyvis download-pretrainedThis tutorial was adapted from the FlyVis visual system tutorial by Janne Lappalainen.
Lappalainen, J. K. et al. Connectome-constrained networks predict neural activity across the fly visual system. Nature 634, 1132–1140 (2024). https://doi.org/10.1038/s41586-024-07939-3
FlyVis Documentation: https://turagalab.github.io/flyvis/