Related Experiment Video
Updated: Aug 9, 2025
09:06
Whole-brain Segmentation and Change-point Analysis of Anatomical Brain MRI—Application in Premanifest Huntington's Disease
Published on: June 9, 2018
12.2K
Robust machine learning segmentation for large-scale analysis of heterogeneous clinical brain MRI datasets
Benjamin Billot1, Colin Magdamo2, You Cheng2
1Centre for Medical Image Computing, University College London, London WC1V 6LJ, UK.
Summary
SynthSeg+ is a new AI tool for analyzing diverse brain MRI scans. It enables robust whole-brain segmentation and cortical parcellation, unlocking quantitative morphometry from clinical data.
Area of Science:
- Neuroimaging
- Artificial Intelligence
- Medical Image Analysis
Background:
- Millions of clinical brain MRI scans are acquired annually, representing a vast, underutilized resource for research.
- Existing automated algorithms struggle with the high variability in clinical MRI data, limiting their research application.
- Quantitative morphometry of brain structure is crucial for understanding neurological conditions and aging.
Purpose of the Study:
- To introduce SynthSeg+, an AI segmentation suite designed for robust analysis of heterogeneous clinical brain MRI datasets.
- To enable automated whole-brain segmentation, cortical parcellation, and intracranial volume estimation from diverse clinical scans.
- To develop a tool for detecting faulty segmentations in low-quality MRI data.
Main Methods:
- Development of an AI-driven segmentation suite, SynthSeg+, capable of handling variations in MR contrasts, resolutions, orientations, and artifacts.
- Implementation of whole-brain segmentation, cortical parcellation, and automated quality control for segmentation accuracy.
- Validation of SynthSeg+ performance across seven experiments, including a large-scale aging study.
Main Results:
- SynthSeg+ demonstrated robust performance in segmenting heterogeneous clinical brain MRI data.
- The tool accurately replicated known atrophy patterns in an aging study using 14,000 scans.
- Automated detection of faulty segmentations was successfully implemented, particularly for low-quality scans.
Conclusions:
- SynthSeg+ provides a robust and automated solution for analyzing large-scale, heterogeneous clinical brain MRI datasets.
- The tool facilitates quantitative morphometry, unlocking the potential of routinely acquired clinical scans for research.
- SynthSeg+ is publicly available as a ready-to-use tool to advance neuroimaging research.
