Related Experiment Video
Updated: Nov 3, 2025
11:50
A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging
Published on: February 4, 2022
4.3K
Using connectomics for predictive assessment of brain parcellations.
Kristoffer J Albers1, Karen S Ambrosen2, Matthew G Liptrot1
1Department of Applied Mathematics and Computer Science, Technical University of Denmark, Richard Petersens Plads, Building 324, DK-2800 Kgs. Lyngby, Denmark.
Neuroimage
|June 4, 2021
Summary
Evaluating human brain organization is crucial. This study quantitatively assesses brain parcellation atlases using connectomics data, finding resolution more critical than precise parcel boundaries for characterizing functional and structural connectivity.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Brain Imaging
Background:
- Human brain organization is complex and poorly understood.
- Macroscale brain analysis typically relies on cortical units (parcellations), but validation is challenging without a gold standard.
Purpose of the Study:
- To develop a framework for quantitatively evaluating brain parcellation atlases.
- To assess how well existing atlases capture high-resolution functional connectivity (FC) and structural connectivity (SC) data.
- To compare the performance of established atlases against data-driven and geometric parcellations.
Main Methods:
- Utilized a statistical prediction framework to evaluate brain parcellations against Human Connectome Project (HCP) connectomics data.
- Assessed ten existing parcellation atlases, comparing them to data-driven and spatially-homogeneous geometric parcellations.
- Analyzed the representation of both functional connectivity (FC) and structural connectivity (SC) at the cortical unit level.
Main Results:
- Found significant discrepancies between parcellations that characterize FC and SC.
- Identified differences in how well parcellations represent individual versus shared functional connectomes.
- Discovered that simple, spatially homogeneous parcellations performed well for FC and SC, but were outperformed by atlases when parcel size distributions were matched.
Conclusions:
- Parcel resolution is more critical than exact border location for characterizing brain connectivity.
- The choice of fine-grained and coarse representations within atlases significantly impacts their effectiveness.
- Quantitative evaluation using connectomics provides a robust method for assessing brain parcellation quality.
