Generative Aging of Brain Images with Diffeomorphic Registration

Abstract

Analyzing and predicting brain aging is essential for early prognosis and accurate diagnosis of cognitive diseases. The technique of neuroimaging, such as Magnetic Resonance Imaging (MRI), provides a noninvasive means of observing the aging process within the brain. With longitudinal image data collection, data-intensive Artificial Intelligence (AI) algorithms have been used to examine brain aging. However, existing state-of-the-art algorithms tend to be restricted to group-level predictions and suffer from unreal predictions. This paper proposes a methodology for generating longitudinal MRI scans that capture subject-specific neurodegeneration and retain anatomical plausibility in aging. The proposed methodology is developed within the framework of diffeomorphic registration and relies on three key novel technological advances to generate subject-level anatomically plausible predictions: i) a computationally efficient and individualized generative framework based on registration; ii) an aging generative module based on biological linear aging progression; iii) a quality control module to fit registration for generation task. Our methodology was evaluated on 2662 T1-weighted (T1-w) MRI scans from 796 participants from three different cohorts. First, we applied 6 commonly used criteria to demonstrate the aging simulation ability of the proposed methodology; Secondly, we evaluated the quality of the synthetic images using quantitative measurements and qualitative assessment by a neuroradiologist. Overall, the experimental results show that the proposed method can produce anatomically plausible predictions that can be used to enhance longitudinal datasets, in turn enabling data-hungry AI-driven healthcare tools.

References in corpus (13)

• Generative Adversarial Networks: An Overview
• VoxelMorph: A Learning Framework for Deformable Medical Image Registration
• TransMorph: Transformer for unsupervised medical image registration
• Deep Learning in Medical Image Registration: A Review
• Unsupervised Learning of Probabilistic Diffeomorphic Registration for Images and Surfaces
• End-to-End Unsupervised Deformable Image Registration with a Convolutional Neural Network
• Learning a Probabilistic Model for Diffeomorphic Registration
• SynthMorph: learning contrast-invariant registration without acquired images
• Learning Conditional Deformable Templates with Convolutional Networks
• Generating Diffusion MRI scalar maps from T1 weighted images using generative adversarial networks
• Brain Imaging Generation with Latent Diffusion Models
• Neuromorphologicaly-preserving Volumetric data encoding using VQ-VAE
• Generative Aging of Brain Images with Diffeomorphic Registration

Cited by in corpus (3)

• Generative Aging of Brain Images with Diffeomorphic Registration
• A deformation-based morphometry framework for disentangling Alzheimer's disease from normal aging using learned normal aging templates
• Synthesizing Individualized Aging Brains in Health and Disease with Generative Models and Parallel Transport