What question did this study set out to answer?

Assess the effectiveness and explainability of federated learning methods in healthcare settings.

March 10, 2026Journal of Biomedical Informatics1 citationsOpen Access

Model Utility And Explainability In Federated Learning For Healthcare

Model utility and explainability in federated learning - A case study in healthcare using fundus oculi datasets

View Full Paper

Ask AI

Bookmark

Authors

Case study evaluates federated learning improvements in medical feature relevance and predictive power.

Assess the effectiveness and explainability of federated learning methods in healthcare settings.
Combined six fundus oculi datasets to simulate diverse federated learning environments
Evaluated three federated learning methods against centrally trained models
Applied explainability techniques to analyze model features and local explanations
Achieved up to 9.97% improvement in model utility with federated learning methods
Identified vertical cup-to-disc ratio as a key feature for glaucoma diagnosis
Demonstrated robustness against biases in fundus datasets
Established high agreement between local explanations and ophthalmologist-annotated attention maps

Penzel et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

March 10, 2026Journal of Biomedical Informatics1 citationsOpen Access

Model utility and explainability in federated learning - A case study in healthcare using fundus oculi datasets

View Full Paper

Ask AI

Bookmark

Authors

Case study evaluates federated learning improvements in medical feature relevance and predictive power.

Assess the effectiveness and explainability of federated learning methods in healthcare settings.
Combined six fundus oculi datasets to simulate diverse federated learning environments
Evaluated three federated learning methods against centrally trained models
Applied explainability techniques to analyze model features and local explanations
Achieved up to 9.97% improvement in model utility with federated learning methods
Identified vertical cup-to-disc ratio as a key feature for glaucoma diagnosis
Demonstrated robustness against biases in fundus datasets
Established high agreement between local explanations and ophthalmologist-annotated attention maps

Penzel et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark