pyDANT: A Python toolbox for Density-based Across-day Neuron Tracking

pyDANT is a Python toolbox that combines iterative motion correction and density-based clustering to robustly track single neurons across days of high-density recordings.

📄 Published article

Density-based longitudinal neuron tracking in high-density electrophysiological recordings

📚 Read the Documentation

🧮 Check out the MATLAB version (DANT)

⚙️ Installation

This section describes how to install pyDANT.

Install with Anaconda

Anaconda is recommended for managing the pyDANT environment.

conda create -n pyDANT python=3.11
conda activate pyDANT
pip install pyDANT

Install from Python Package Index (PyPI)

You can also install pyDANT directly from PyPI:

pip install pyDANT

Install from Source

If you prefer to install from source, clone the repository and install it manually:

git clone https://github.com/jiumao2/pyDANT.git
cd pyDANT
pip install -e .

🚀 Getting Started

To help you get familiar with the pipeline, we provide two tutorials.

1. Colab Tutorial: Open the pyDANT Colab demo to run pyDANT in a browser without local setup or pre-downloading data.
2. Documentation Tutorial: Download the example dataset for pyDANT, then follow the Python tutorial to run pyDANT locally or process your own recordings.

If you encounter any bugs, have questions, or want to suggest a feature, please open an issue. We look forward to your feedback!

📝 Citation

If you use pyDANT in your research, please cite our article:

@article{Huang2026DANT,
    author = {Huang, Yue and Wang, Hanbo and Cao, Jiaming and Chen, Yu and Wang, Xuanning and Zhao, Yujie and Ren, Hengkun and Zheng, Qiang and Yu, Jianing},
    title = {Density-based longitudinal neuron tracking in high-density electrophysiological recordings},
    journal = {Patterns},
    year = {2026},
    doi = {10.1016/j.patter.2026.101590},
    url = {https://doi.org/10.1016/j.patter.2026.101590},
    note = {Available online 17 June 2026}
}

📚 References & Acknowledgements

pyDANT builds upon and integrates several excellent open-source tools. We extend our gratitude to the authors of the following packages:

• HDBSCAN: Hierarchical Density-Based Spatial Clustering of Applications with Noise. (Campello et al., 2013; McInnes & Healy, 2017).
• Kilosort: Fast spike sorting with drift correction. (Pachitariu et al., 2024).
• DREDge: Robust online multiband drift estimation in electrophysiology data. (Windolf et al., 2025).

📄 License

This project is licensed under the GNU General Public License v3.0 - see the LICENSE file for details.