Adaptive Neighborhoods in Contrastive Regression Learning for Brain Age Prediction
Abstract
In neuroimaging, accurate brain age prediction is key to understanding brain aging and early neurodegenerative signs. Recent advancements in self-supervised learning, particularly contrastive learning, have shown robustness with complex datasets but struggle with non-uniformly distributed data common in medical imaging. We introduce a novel contrastive loss that dynamically adapts during training, focusing on localized sample neighborhoods. Additionally, we incorporate brain stiffness, a mechanical property sensitive to aging. Our approach outperforms state-of-the-art methods and opens new directions for brain aging research.
Cite
Text
Träuble et al. "Adaptive Neighborhoods in Contrastive Regression Learning for Brain Age Prediction." NeurIPS 2024 Workshops: SSL, 2024.Markdown
[Träuble et al. "Adaptive Neighborhoods in Contrastive Regression Learning for Brain Age Prediction." NeurIPS 2024 Workshops: SSL, 2024.](https://mlanthology.org/neuripsw/2024/trauble2024neuripsw-adaptive/)BibTeX
@inproceedings{trauble2024neuripsw-adaptive,
title = {{Adaptive Neighborhoods in Contrastive Regression Learning for Brain Age Prediction}},
author = {Träuble, Jakob and Hiscox, Lucy V and Johnson, Curtis and Schönlieb, Carola-Bibiane and Schierle, Gabriele S Kaminski and Aviles-Rivero, Angelica I},
booktitle = {NeurIPS 2024 Workshops: SSL},
year = {2024},
url = {https://mlanthology.org/neuripsw/2024/trauble2024neuripsw-adaptive/}
}