Contrastive learning for regression in multi-site brain age prediction

TL;DR

This paper introduces a novel contrastive learning regression loss for brain age prediction from MRI scans, significantly improving model robustness and generalization across multi-site neuroimaging datasets.

Contribution

It proposes a new contrastive learning approach tailored for regression tasks in multi-site brain age prediction, enhancing robustness to site-related noise.

Findings

Achieved state-of-the-art performance on the OpenBHB challenge.

Demonstrated superior generalization across diverse datasets.

Improved robustness to site-related noise in MRI data.

Abstract

Building accurate Deep Learning (DL) models for brain age prediction is a very relevant topic in neuroimaging, as it could help better understand neurodegenerative disorders and find new biomarkers. To estimate accurate and generalizable models, large datasets have been collected, which are often multi-site and multi-scanner. This large heterogeneity negatively affects the generalization performance of DL models since they are prone to overfit site-related noise. Recently, contrastive learning approaches have been shown to be more robust against noise in data or labels. For this reason, we propose a novel contrastive learning regression loss for robust brain age prediction using MRI scans. Our method achieves state-of-the-art performance on the OpenBHB challenge, yielding the best generalization capability and robustness to site-related noise.