Semi-Supervised Diffusion Model for Brain Age Prediction

Ayodeji Ijishakin, Sophie A. Martin, Florence J Townend, James H. Cole, ANDREA MALASPINA

Keywords: Diffusion Models, Brain Age Prediction, Semi-Supervised Learning, MR Image Analysis, Medical Imaging

TL;DR: A semi-supervised diffusion model is used to predict brain-age T1w clinical-grade MR data, we achieve an 0.83(p<0.01) test accuracy and produce predictions associated with ALS survival; outperforming state-of-the-art approaches.

Abstract: Brain age prediction models have succeeded in predicting clinical outcomes in neurodegenerative diseases, but can struggle with tasks involving faster progressing diseases and low quality data. To enhance their performance, we employ a semi-supervised diffusion model, obtaining a 0.83(p<0.01) correlation between chronological and predicted age on low quality T1w MR images. This was competitive with state-of-the-art non-generative methods. Furthermore, the predictions produced by our model were significantly associated with survival length (r=0.24, p<0.05) in Amyotrophic Lateral Sclerosis. Thus, our approach demonstrates the value of diffusion-based architectures for the task of brain age prediction.

Submission Number: 38