Related Experiment Video
Updated: Nov 6, 2025
05:14
Standardized Data Acquisition for Neuromelanin-Sensitive Magnetic Resonance Imaging of the Substantia Nigra
Published on: September 8, 2021
4.0K
Cross-Scanner Harmonization of Neuromelanin-Sensitive MRI for Multisite Studies
Kenneth Wengler1, Clifford Cassidy2, Marieke van der Pluijm3,4
1Department of Psychiatry, New York State Psychiatric Institute, Columbia University, New York, USA.
Journal of Magnetic Resonance Imaging : JMRI
|May 7, 2021
Summary
Neuromelanin-sensitive MRI (NM-MRI) harmonization using ComBat effectively removes scanner-specific variations in substantia nigra-ventral tegmental area (SN-VTA) contrast. This method preserves age-related biological variability, crucial for developing reliable biomarkers.
Area of Science:
- Neuroimaging
- Biomarker Development
- Neuroscience
Background:
- Neuromelanin-sensitive magnetic resonance imaging (NM-MRI) measures neuromelanin concentration in the substantia nigra-ventral tegmental area (SN-VTA) complex.
- NM-MRI serves as a proxy for dopaminergic function and holds potential as a noninvasive biomarker.
- Harmonization is essential for combining large-scale, multi-site NM-MRI data for generalizable biomarker development.
Purpose of the Study:
- To develop and validate a method for harmonizing NM-MRI data across different scanners and research sites.
- To ensure the generalizability of NM-MRI as a biomarker by reducing technical variability.
Main Methods:
- A total of 128 healthy subjects from three sites and five MRI scanners were included.
- NM-MRI contrast-to-noise ratio (CNR) was calculated before and after ComBat harmonization.
- Support vector machines (SVM) assessed scanner differences, while support vector regression (SVR) evaluated age-related biological variability.
Main Results:
- Raw NM-MRI CNR showed significant differences across scanners (SVM accuracy = 86.5%).
- ComBat harmonization successfully removed scanner-specific differences (harmonized CNR SVM accuracy = 29.5%, P=0.8542).
- Harmonization did not significantly impact the prediction of age from CNR values (Δr = -0.06, P=0.7304).
Conclusions:
- The ComBat harmonization method effectively eliminates technical variability in SN-VTA NM-MRI contrast across different scanners.
- This harmonization approach preserves biologically relevant information, such as age-related changes in dopaminergic function.
- Harmonized NM-MRI data can facilitate large-scale studies and the development of robust, noninvasive biomarkers for neurological conditions.
