Neuroscience Data Analysis Notebooks

A collection of Jupyter notebooks exploring neuronal waveforms, retinal ganglion cell (RGC) images, dimensionality reduction, clustering, and other neuroscience data-analysis questions.

The notebooks are exploratory workflows with different input datasets and dependencies. Start with a topic below rather than running every notebook in filename order.

Featured workflows

Neuronal waveform clustering

• 04_PCA_clustering_compare_methods.ipynb standardizes waveform samples, applies PCA, and compares k-means with hierarchical clustering using cluster-quality metrics and plots.
• 05_cluster_characterization.ipynb examines exported cluster assignments, waveform features, method agreement, and differences between clusters.

These notebooks expect a cleaned waveform CSV and, for characterization, intermediate CSV outputs under local data/ and outputs/ directories. Those required files are not all distributed in this repository. The separate neuropixels_analysis repository contains an import-and-cleaning notebook; MATLAB_neuropixels_clustering contains a focused MATLAB/Python waveform project with example results.

Retinal ganglion cell image analysis

• 01_RGC_import_preview.ipynb previews cropped images and creates a contact sheet.
• 02_RGC_CNN_embeddings_clustering.ipynb uses a pretrained ResNet18 feature extractor, PCA/UMAP, and unsupervised clustering.
• 03_RGC_cluster_characterization_revised.ipynb reviews cluster structure and morphology-related outputs.

This sequence expects local cropped images and generated files that are not included in the repository.

Other analyses

Additional notebooks explore CSD and traumatic brain injury, retinal imaging, GluSnFR signals, evoked potentials, and model-based workflows. Treat these as individual analyses; inspect each notebook's stated inputs and assumptions before running it.

Getting started

Create a Python environment with JupyterLab and the packages imported by your chosen notebook, then launch JupyterLab from the repository root:

jupyter lab

For the waveform clustering notebooks, common dependencies include NumPy, pandas, Matplotlib, SciPy, and scikit-learn. Image and Bayesian notebooks require additional packages. There is no complete root-level dependency specification for every notebook.

Repository notes

• The root-level .mat file is an experimental data artifact; its provenance, sharing permission, and role in a reproducible example should be checked before wider distribution.
• Some notebooks are drafts or alternate versions. The links above identify useful entry points; they do not imply that all notebooks form one end-to-end pipeline.
• Generated outputs and some source datasets are absent, so a fresh clone will not execute every notebook without supplying the corresponding data.