Related Experiment Video
Updated: May 3, 2026
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
MANIFOLD-CONSTRAINED EMBEDDINGS FOR THE DETECTION OF WHITE MATTER LESIONS IN BRAIN MRI
Samuel Kadoury1, Guray Erus2, Evangelia Zacharaki3
1Philips Research North America, Briarcliff Manor, NY, USA.
Abstract:
Brain abnormalities such as white matter lesions (WMLs) are not only linked to cerebrovascular disease, but also with normal aging, diabetes and other conditions increasing the risk for cerebrovascular pathologies. Obtaining quantitative measures which assesses the degree or probability of WML in patients is important for evaluating disease burden, and for evaluating its progression and response to interventions. In this paper, we introduce a novel approach for detecting the presence of WMLs in periventricular areas of the brain using manifold-constrained embeddings. The proposed method uses locally linear embedding (LLE) to create "normality" distributions in 12 locations of the brain where deviations from the manifolds are estimated by calculating geodesic distances along locally linear planes in the embedding. A smooth mapping function approximating the relationship between ambient and manifold spaces as a joint distribution maps unseen test images in the intrinsic space. We create a set of low-dimensional embeddings from 876 patches of healthy tissue in 73 subjects and test it on 396 patches imaging both WML and healthy areas in 33 subjects with diabetes. Experiments highlight the need of nonlinear techniques to learn the studied data with detection rates over 85% in true-positives, and the relevance of the computed distance for comparing individuals to a specific pathological pattern.
Insights
This study introduces a new method using manifold-constrained embeddings to detect white matter lesions (WMLs) in the brain. The approach achieved over 85% accuracy in identifying WMLs, aiding in disease assessment.
Area of Science:
- Neuroimaging
- Medical Image Analysis
- Computational Neuroscience
Background:
- White matter lesions (WMLs) are associated with cerebrovascular disease, aging, and diabetes.
- Quantitative measures of WMLs are crucial for assessing disease burden, progression, and treatment response.
- Accurate detection of WMLs is vital for understanding their impact on neurological health.
Purpose of the Study:
- To develop a novel approach for detecting WMLs in periventricular brain regions.
- To utilize manifold-constrained embeddings for WML identification.
- To provide quantitative measures for evaluating WML presence and severity.
Main Methods:
- Locally linear embedding (LLE) was used to create normality distributions in 12 brain locations.
- Geodesic distances along locally linear planes estimated deviations from manifolds.
- A smooth mapping function projected unseen images into an intrinsic space for analysis.
Main Results:
- Experiments demonstrated the necessity of nonlinear techniques for analyzing the data.
- The proposed method achieved a true-positive detection rate exceeding 85%.
- The computed distance proved relevant for comparing individuals against pathological patterns.
Conclusions:
- Manifold-constrained embeddings offer a promising nonlinear technique for WML detection.
- The method provides a valuable tool for quantitative assessment of WMLs in clinical settings.
- This approach can aid in disease burden evaluation and monitoring treatment efficacy.
