UNSURF: Uncertainty Quantification for Cortical Surface Reconstruction of Clinical Brain MRIs

2025-05-31

AI Summary

Existing methods—such as voxel-wise Monte Carlo variance—struggle to accurately quantify localization uncertainty in cortical surface reconstruction from clinical brain MRI, especially under arbitrary scan orientations, resolutions, and contrast conditions. To address this, we propose UNSURF: the first framework that directly predicts a voxel-wise signed distance function (SDF) and uses the residual between predicted and ground-truth SDFs as a principled uncertainty metric. By operating in the implicit SDF space, UNSURF circumvents the failure of conventional statistical estimators in non-Euclidean surface domains. It enables automated, multi-granularity quality control—at the subject, region-of-interest, and vertex levels. Experiments demonstrate strong correlation (r > 0.85) between UNSURF’s uncertainty estimates and true surface errors, substantially improving QC efficiency. Moreover, integrating UNSURF-derived uncertainty maps enhances downstream Alzheimer’s disease classification performance, boosting AUC by 3.2%.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations
• Computer Vision: Segmentation

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

We propose UNSURF, a novel uncertainty measure for cortical surface reconstruction of clinical brain MRI scans of any orientation, resolution, and contrast. It relies on the discrepancy between predicted voxel-wise signed distance functions (SDFs) and the actual SDFs of the fitted surfaces. Our experiments on real clinical scans show that traditional uncertainty measures, such as voxel-wise Monte Carlo variance, are not suitable for modeling the uncertainty of surface placement. Our results demonstrate that UNSURF estimates correlate well with the ground truth errors and: extit{(i)}~enable effective automated quality control of surface reconstructions at the subject-, parcel-, mesh node-level; and extit{(ii)}~improve performance on a downstream Alzheimer's disease classification task.

Problem

• Quantifying uncertainty in cortical surface reconstruction from clinical MRI scans
• Evaluating traditional uncertainty measures for surface placement accuracy
• Enhancing quality control and Alzheimer's disease classification performance

Innovation

• Novel uncertainty measure for cortical surface reconstruction
• Discrepancy between predicted and actual SDFs
• Improves Alzheimer's disease classification performance