While foundation models (FMs) offer strong potential for AI-based dementia diagnosis, their integration into federated learning (FL) systems remains underexplored. In this benchmarking study, we systematically evaluate the impact of key design choices: classification head architecture, fine-tuning strategy, and aggregation method, on the performance and efficiency of federated FM tuning using brain MRI data. Using a large multi-cohort dataset, we find that the architecture of the classification head substantially influences performance, freezing the FM encoder achieves comparable results to full fine-tuning, and advanced aggregation methods outperform standard federated averaging. Our results offer practical insights for deploying FMs in decentralized clinical settings and highlight trade-offs that should guide future method development.
@article{arxiv.2508.21458,
title = {Federated Fine-tuning of SAM-Med3D for MRI-based Dementia Classification},
author = {Kaouther Mouheb and Marawan Elbatel and Janne Papma and Geert Jan Biessels and Jurgen Claassen and Huub Middelkoop and Barbara van Munster and Wiesje van der Flier and Inez Ramakers and Stefan Klein and Esther E. Bron},
journal= {arXiv preprint arXiv:2508.21458},
year = {2025}
}
Comments
Accepted at the MICCAI 2025 Workshop on Distributed, Collaborative and Federated Learning (DeCAF)