Connectome-based models of disease propagation are used to probe mechanisms of pathology spread in neurodegenerative disease. We present our network spreading model toolbox that allows the user to compare model fits across different models and parameters.
- Network Diffusion Model (NDM) models diffusive spread of pathology between connected brain regions (Raj et al.)
- Fisher-Kolmogorov-Petrovsky-Piskunov (FKPP): network spreading plus uniform local production of pathology (Weickenmeier et al.)
- Weighted-FKPP: spreading plus regionally-varying production weighted by a vector of choice (He et al.)
-
Interaction-weighted FKPP:
an extension of the weighted-FKPP model with extra coupling parameter
$\gamma_{regional}$ that determines maximum influence of the regional weights on production (McAlinden at al.)
Weighting the production rates in the FKPP model by different values across the brain can improve the ability of the model to capture pathology patterns, by taking into account different regional characteristics during disease progression [1,2]. We define this new model as weighted-FKPP, and show that weighting pathology production with regional amyloid improves the model fit to tau-PET data [3].
Equations and simulated pathology over time are provided in the figure below:
First, clone the repository into your desired location. Then, from within network_spreading_models directory:
# Create and activate conda environment
conda env create --name network_spreading_models --file environment.yml
conda activate network_spreading_models
# Install the package in development mode
pip install -e .
# Verify installation
python -c 'import src; print("Installation successful!")'The uv package manager may be more reliable for installing on Linux. First, you will need to install uv using the instructions here.
# Create virtual environment (Python 3.12)
uv venv -p 3.12
source .venv/bin/activate # On Windows: .venv\Scripts\activate
# Install dependencies and package
uv pip sync requirements.txt
uv pip install -e .
# Verify installation
python -c 'import src; print("Installation successful!")'Note: Both installation methods include ipython, which is required for running the tutorial notebook.
The tutorial in notebooks/tutorial.ipynb should get you familiar with the different models in the toolbox. If you have any further questions, don't hesitate to raise an issue!
This toolbox originated from project for CMICHACKS 2023. The team included: Ellie Thompson, Anna Schroder, Tiantian He, Neil Oxtoby, James Cole, Antoine Legouhy and Xin Zhao. If you use the toolbox for your project, please cite the AAIC abstract below.
[1]. He, T., Schroder, A., Thompson, E., Oxtoby, N. P., Abdulaal, A., Barkhof, F., & Alexander, D. C. (2023). Coupled pathology appearance and propagation models of neurodegeneration. The Organization for Human Brain Mapping (OHBM) 2023 Annual Meeting, [Oral Presentation]
[2]. He, T., Thompson, E., Schroder, A., Oxtoby, N. P., Abdulaal, A., Barkhof, F., & Alexander, D. C. (2023, October). A coupled-mechanisms modelling framework for neurodegeneration. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 459-469). Cham: Springer Nature Switzerland. https://link.springer.com/chapter/10.1007/978-3-031-43993-3_45
[3]. Thompson, E., Schroder, A., He, T., Legouhy, A., Zhao, X., Cole, J.H., Oxtoby, N.P. and Alexander, D.C., (2024, July). Demonstration of an open-source toolbox for network spreading models: regional amyloid burden promotes tau production in Alzheimer's disease. In Alzheimer's Association International Conference. ALZ.https://alz-journals.onlinelibrary.wiley.com/doi/full/10.1002/alz.093791
[4]. McAlinden, D., Thompson, E., Soskic, S., Oxtoby, N. P. (2025). Predicting the spread of Alzheimer's disease pathology using brain connectomes, amyloid beta and tau during different cognitive stages. MSc Project report, available at SSRN: http://dx.doi.org/10.2139/ssrn.5707002