SimCortex: Collision-free Simultaneous Cortical Surfaces Reconstruction

2025-07-09

AI Summary

Existing cortical surface reconstruction methods struggle to simultaneously preserve complex geometric details and enforce strict topological constraints, often resulting in self-intersections, overlaps, and topological errors. To address this, we propose an end-to-end differentiable framework that enables simultaneous reconstruction of four cortical surfaces—bilateral white matter and pial surfaces—for the first time. Our method integrates a nine-class tissue prior, static velocity field modeling, and the square-root velocity scaling (SRVS) strategy to construct a multi-scale diffeomorphic deformation field, while incorporating subject-specific initial meshes to improve anatomical consistency. Evaluated on standard benchmarks, our approach significantly reduces surface self-intersection rates (by 82% on average) and overlap errors, achieving state-of-the-art performance in both Hausdorff distance and mean surface distance. This work establishes a robust, topology-guaranteed paradigm for high-fidelity neuroanatomical surface modeling.

Technology Category

• Machine Learning: Learning with Manifolds
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Computer Vision: Diffusion Models for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Accurate cortical surface reconstruction from magnetic resonance imaging (MRI) data is crucial for reliable neuroanatomical analyses. Current methods have to contend with complex cortical geometries, strict topological requirements, and often produce surfaces with overlaps, self-intersections, and topological defects. To overcome these shortcomings, we introduce SimCortex, a deep learning framework that simultaneously reconstructs all brain surfaces (left/right white-matter and pial) from T1-weighted(T1w) MRI volumes while preserving topological properties. Our method first segments the T1w image into a nine-class tissue label map. From these segmentations, we generate subject-specific, collision-free initial surface meshes. These surfaces serve as precise initializations for subsequent multiscale diffeomorphic deformations. Employing stationary velocity fields (SVFs) integrated via scaling-and-squaring, our approach ensures smooth, topology-preserving transformations with significantly reduced surface collisions and self-intersections. Evaluations on standard datasets demonstrate that SimCortex dramatically reduces surface overlaps and self-intersections, surpassing current methods while maintaining state-of-the-art geometric accuracy.

Problem

• Reconstructs cortical surfaces without overlaps or self-intersections
• Ensures topological correctness in brain surface reconstruction
• Improves accuracy in MRI-based neuroanatomical analyses

Innovation

• Deep learning framework for simultaneous cortical surfaces reconstruction
• Generates collision-free initial surface meshes from segmentations
• Uses multiscale diffeomorphic deformations with stationary velocity fields