# Federated Fine-tuning of SAM-Med3D for MRI-based Dementia Classification

## TL;DR

This study evaluates how different design choices affect federated fine-tuning of foundation models for MRI-based dementia classification, providing insights for practical deployment in decentralized clinical environments.

## Contribution

It systematically assesses the impact of classification head design, fine-tuning strategies, and aggregation methods on federated foundation model performance using brain MRI data.

## Key findings

- Classification head architecture significantly affects performance.
- Freezing the FM encoder yields results comparable to full fine-tuning.
- Advanced aggregation methods outperform standard federated averaging.

## Abstract

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.