Federated Fine-Tuning of SAM-Med3D for MRI-Based Dementia Classification

Kaouther Mouheb

doi:10.1007/978-3-032-05663-4_7

Federated Fine-Tuning of SAM-Med3D for MRI-Based Dementia Classification

2026-01-01

0

PRE

AI

K

Kaouther Mouheb *

M

Marawan Elbatel

J

Janne M. Papma

G

Geert Jan Biessels

J

Jurgen A.H.R. Claassen

H

Huub A. M. Middelkoop

B

Barbara C. Van Munster

W

Wiesje M. van der Flier

I

Inez Ramakers

S

Stefan Klein

E

Esther E. Bron

DOI：10.1007/978-3-032-05663-4_7

Original

Original request for help

Save

Abstract

En 中文

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.

Keywords:

Federated learning

Foundation models

Dementia

MRI

Federated Fine-Tuning of SAM-Med3D for MRI-Based Dementia Classification

Abstract

Journal

Organization

Cited Papers

Citing Papers