SSL-AD: Spatiotemporal Self-Supervised Learning for Generalizability and Adaptability Across Alzheimer's Prediction Tasks and Datasets

2025-09-12

AI Summary

To address key challenges in Alzheimer’s disease (AD) prediction—including severe scarcity of labeled data, poor cross-dataset generalization, and variable numbers and irregular temporal intervals of input MRI scans—this paper proposes a spatiotemporal self-supervised learning framework for 3D brain MRI. The method introduces a novel extension module capable of handling variable-length longitudinal inputs, jointly optimizing temporal order prediction and contrastive learning while incorporating spatial feature enhancement. Pretraining is conducted across four public datasets. Experimental results demonstrate that the framework outperforms supervised baselines on six of seven downstream tasks, yielding significant improvements in AD diagnosis classification, conversion detection (e.g., MCI-to-AD), and future clinical state prediction. To foster reproducibility and further research, the source code and pretrained models are publicly released.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: 3D Computer Vision
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs

Abstract

Alzheimer's disease is a progressive, neurodegenerative disorder that causes memory loss and cognitive decline. While there has been extensive research in applying deep learning models to Alzheimer's prediction tasks, these models remain limited by lack of available labeled data, poor generalization across datasets, and inflexibility to varying numbers of input scans and time intervals between scans. In this study, we adapt three state-of-the-art temporal self-supervised learning (SSL) approaches for 3D brain MRI analysis, and add novel extensions designed to handle variable-length inputs and learn robust spatial features. We aggregate four publicly available datasets comprising 3,161 patients for pre-training, and show the performance of our model across multiple Alzheimer's prediction tasks including diagnosis classification, conversion detection, and future conversion prediction. Importantly, our SSL model implemented with temporal order prediction and contrastive learning outperforms supervised learning on six out of seven downstream tasks. It demonstrates adaptability and generalizability across tasks and number of input images with varying time intervals, highlighting its capacity for robust performance across clinical applications. We release our code and model publicly at https://github.com/emilykaczmarek/SSL-AD.

Problem

• Addressing limited labeled data in Alzheimer's prediction models
• Improving generalization across diverse Alzheimer's datasets
• Enhancing adaptability to variable scan intervals and counts

Innovation

• Spatiotemporal self-supervised learning approach
• Handles variable-length 3D MRI inputs
• Combines temporal order and contrastive learning