Keywords: Deformation-Based Morphometry, Normal Aging, Alzheimer’s Disease, Deep Learning, Brain Morphology
This repository contains the source code accompanying the paper:
Fu, Jingru, et al. "Decomposing the effect of normal aging and Alzheimer’s disease in brain morphological changes via learned aging templates", Scientific Reports, 2025.
- Add the link to the paper
- Share the pre-trained weights (in
./models) - Add the Jupyter notebook demo (
step-by-step_example.ipynb)
We used the OASIS-3 dataset in this project. To preprocess the data in accordance with our study, please refer to https://github.com/Fjr9516/Nii2NPZ.
If you use this dataset, please cite the following and refer to the corresponding Data Use Agreement .
- OASIS-3: Longitudinal Neuroimaging, Clinical, and Cognitive Dataset for Normal Aging and Alzheimer Disease. Pamela J LaMontagne, Tammie L.S. Benzinger, John C. Morris, Sarah Keefe, Russ Hornbeck, Chengjie Xiong, Elizabeth Grant, Jason Hassenstab, Krista Moulder, Andrei Vlassenko, Marcus E. Raichle, Carlos Cruchaga, Daniel Marcus, 2019. medRxiv. doi: 10.1101/2019.12.13.19014902
We offer pretrained Atlas-GAN weights trained on the OASIS-3 Dataset for simulating normal aging. Additionally, we provide inference scripts for extracting the learned diffeomorphic registration module and template generation module. (in ./models)
The extracted healthy templates are also available in NIfTI format, spanning ages from 60 to 90. (in ./example/CN_templates)
@article{fu2025decomposing,
title={Decomposing the effect of normal aging and Alzheimer’s disease in brain morphological changes via learned aging templates},
author={Fu, Jingru and Ferreira, Daniel and Smedby, {\"O}rjan and Moreno, Rodrigo},
journal={Scientific Reports},
volume={15},
number={1},
pages={11813},
year={2025},
publisher={Nature Publishing Group UK London}
}
@misc{fu2023deformationbased,
title={A deformation-based morphometry framework for disentangling Alzheimer's disease from normal aging using learned normal aging templates},
author={Jingru Fu and Daniel Ferreira and Örjan Smedby and Rodrigo Moreno},
year={2023},
eprint={2311.08176},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
This repository is developed based on the Atlas-GAN project and makes extensive use of the VoxelMorph library.
