CINeMA: Conditional Implicit Neural Multi-Modal Atlas for a Spatio-Temporal Representation of the Perinatal Brain

2025-06-11

AI Summary

Building high spatiotemporal-resolution perinatal pathological brain atlases is hindered by the scarcity of annotated samples for rare conditions (e.g., ventriculomegaly, agenesis of the corpus callosum). Method: We propose the first conditional implicit neural representation (INR) framework tailored to such rare pathologies. Instead of conventional image registration, our approach jointly models gestational age, postnatal age, and pathology-specific features directly in latent space via INR coupled with conditional diffusion or MLPs. It incorporates multimodal fusion and anatomy-aware loss constraints to ensure end-to-end interpretability, editability, and generalizability. Contribution/Results: The framework accelerates atlas construction by two orders of magnitude (from days to minutes), outperforms state-of-the-art methods in tissue segmentation and fetal age estimation, and enables generation of high-fidelity, anatomically consistent synthetic data for augmentation and analysis.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Natural Language Processing: Generation
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Magnetic resonance imaging of fetal and neonatal brains reveals rapid neurodevelopment marked by substantial anatomical changes unfolding within days. Studying this critical stage of the developing human brain, therefore, requires accurate brain models-referred to as atlases-of high spatial and temporal resolution. To meet these demands, established traditional atlases and recently proposed deep learning-based methods rely on large and comprehensive datasets. This poses a major challenge for studying brains in the presence of pathologies for which data remains scarce. We address this limitation with CINeMA (Conditional Implicit Neural Multi-Modal Atlas), a novel framework for creating high-resolution, spatio-temporal, multimodal brain atlases, suitable for low-data settings. Unlike established methods, CINeMA operates in latent space, avoiding compute-intensive image registration and reducing atlas construction times from days to minutes. Furthermore, it enables flexible conditioning on anatomical features including GA, birth age, and pathologies like ventriculomegaly (VM) and agenesis of the corpus callosum (ACC). CINeMA supports downstream tasks such as tissue segmentation and age prediction whereas its generative properties enable synthetic data creation and anatomically informed data augmentation. Surpassing state-of-the-art methods in accuracy, efficiency, and versatility, CINeMA represents a powerful tool for advancing brain research. We release the code and atlases at https://github.com/m-dannecker/CINeMA.

Problem

• Creating high-resolution spatio-temporal brain atlases for perinatal stages
• Addressing data scarcity in brain studies with pathologies
• Reducing atlas construction time from days to minutes

Innovation

• Latent space operation for fast atlas construction
• Conditional on anatomical features and pathologies
• Generative properties for synthetic data creation