Genetic information improves forecasting of neuronal activity
📝 Manuscript under preparation.
Most models of neural activity treat every neuron the same way. This project tests whether adding a neuron's cell type helps forecast its activity in mouse visual cortex, and compares a genetic label (genotype) with a type learned from activity (phenotype).
| Dataset | What it is |
|---|---|
| Allen Brain Observatory – Visual Coding 2P | Two-photon calcium imaging with one Cre line per recording |
| Allen Visual Coding – Neuropixels | Spike recordings with optotagged cell types |
| Bugeon et al. 2022 | Several transcriptomic types recorded together |
Raw data goes in data/ and is not tracked by git.
scripts/ analysis and plotting scripts
results/ short notes and tables
plots/ figures
notebooks/ exploration notebooks
uv venv .venv --python 3.11
uv pip install --python .venv/bin/python allensdk torch scikit-learn matplotlib oasis-deconv "setuptools<70" "statsmodels<0.14.5"