Related Experiment Video
Updated: Sep 21, 2025
11:50
A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging
Published on: February 4, 2022
4.1K
Systematic validation of an automated thalamic parcellation technique using anatomical data at 3T
Brendan Williams1, Etienne Roesch1, Anastasia Christakou1
1Centre for Integrative Neuroscience and Neurodynamics, University of Reading, Reading, Berkshire, RG6 6AL, UK; School of Psychology and Clinical Language Sciences, University of Reading, Reading, Berkshire, RG6 7BE, UK.
Neuroimage
|June 1, 2022
Summary
Automated segmentation of thalamic nuclei shows mixed results. While it delineates major divisions, accuracy varies, especially in posterior regions, requiring careful use in neuroscience research.
Area of Science:
- Neuroscience
- Brain Imaging
- Computational Anatomy
Background:
- The thalamus, crucial for cognition, is often treated as a single unit, overlooking nuclei-specific functions.
- Technical challenges in structural imaging limit the dissociation of individual thalamic nuclei.
- Automated segmentation algorithms offer potential solutions for delineating thalamic nuclei.
Purpose of the Study:
- To systematically validate an automated thalamic nuclei segmentation approach using Human Connectome Project data.
- To assess the accuracy of segmentations against the established Morel stereotaxic atlas.
- To quantify the efficacy of automated segmentation in delineating thalamic nuclei.
Main Methods:
- Utilized an automated segmentation algorithm applying a probabilistic atlas via FreeSurfer software.
- Employed Human Connectome Project data for validation.
- Quantified segmentation quality using Average Hausdorff Distance and DICE coefficient, comparing against the Morel atlas.
Main Results:
- The automated approach successfully generated boundaries for anterior, lateral, posterior, and medial thalamic divisions.
- Segmentation quality was mixed, with better delineation in anterior, lateral, and medial regions compared to the posterior thalamus.
- All segmented nuclei volumes significantly differed from the corresponding regions in the Morel atlas.
Conclusions:
- The automated thalamic nuclei segmentation method shows potential but requires cautious application.
- Users should be mindful of segmentation accuracy, particularly in the posterior thalamus.
- Further refinement is needed for precise studies on the structural and functional roles of individual thalamic nuclei.
