SMART: A Flexible, Interpretable, and Scalable Spatio-temporal Brain Atlas from High-Resolution Imaging Data

2026-06-17

AI Summary

Existing spatiotemporal brain mapping methods often rely on black-box generative models, suffering from limited interpretability and scalability. This work proposes the first flexible framework that integrates anatomical priors, region-specific differential equations, and multiscale neuronal cellular automata to disentangle population-level disease dynamics from individual anatomical variability along a continuous pathological timeline. The approach models an interpretable, shared progression trajectory using differential equations and achieves personalized anatomical adaptation through dense diffeomorphic deformations coupled with multiscale cellular automata. Evaluated on five Alzheimer’s disease MRI datasets comprising over 1,300 subjects, the method achieves state-of-the-art performance in both predictive accuracy and temporal consistency, while generating disease progression trajectories with clear anatomical meaning.

Abstract

We introduce SMART, a framework for learning a flexible, interpretable, and scalable spatio-temporal brain atlas from longitudinal high-resolution 3D medical images. Existing approaches to spatio-temporal atlas construction rely on black-box generative models that lack flexibility, limit interpretability, and struggle to scale to high-dimensional data. SMART addresses these challenges by learning a continuous disease-time atlas that decouples global group-wise disease dynamics from their patient-specific anatomical manifestation. Guided by anatomically inspired priors, SMART models interpretable global trajectories of regional progression along a shared disease timeline through region-specific differential equations. Global trajectories are further personalized to individual anatomies via dense diffeomorphic displacements parameterized by a flexible and scalable multi-scale Neural Cellular Automata. Evaluated on five longitudinal MRI datasets in Alzheimer's disease (ADNI-1/GO/2, OASIS-3, AIBL; > 1,300 subjects), SMART produces anatomically meaningful predictions of disease progression and achieves state-of-the-art forecasting accuracy and improved temporal consistency over adversarial and diffusion baselines. Our approach establishes a new paradigm for flexible, interpretable, and scalable modeling of spatio-temporal change in high-dimensional medical image time-series.

Problem

• spatio-temporal brain atlas
• high-resolution imaging
• interpretability
• scalability
• longitudinal medical images

Innovation

• spatio-temporal atlas
• interpretable modeling
• Neural Cellular Automata
• diffeomorphic registration
• disease progression modeling