Reproducibility assessment of neuromelanin-sensitive magnetic resonance imaging protocols for region-of-interest and voxelwise analyses

Abstract

Neuromelanin-sensitive MRI (NM-MRI) provides a noninvasive measure of the content of neuromelanin (NM), a product of dopamine metabolism that accumulates with age in dopamine neurons of the substantia nigra (SN). NM-MRI has been validated as a measure of both dopamine neuron loss, with applications in neurodegenerative disease, and dopamine function, with applications in psychiatric disease. Furthermore, a voxelwise-analysis approach has been validated to resolve substructures, such as the ventral tegmental area (VTA), within midbrain dopaminergic nuclei thought to have distinct anatomical targets and functional roles. NM-MRI is thus a promising tool that could have diverse research and clinical applications to noninvasively interrogate in vivo the dopamine system in neuropsychiatric illness. Although a test-retest reliability study by Langley et al. using the standard NM-MRI protocol recently reported high reliability, a systematic and comprehensive investigation of the performance of the method for various acquisition parameters and preprocessing methods has not been conducted. In particular, most previous studies used relatively thick MRI slices (~3 ​mm), compared to the typical in-plane resolution (~0.5 ​mm) and to the height of the SN (~15 ​mm), to overcome technical limitations such as specific absorption rate and signal-to-noise ratio, at the cost of partial-volume effects. Here, we evaluated the effect of various acquisition and preprocessing parameters on the strength and test-retest reliability of the NM-MRI signal to determine optimized protocols for both region-of-interest (including whole SN-VTA complex and atlas-defined dopaminergic nuclei) and voxelwise measures. Namely, we determined a combination of parameters that optimizes the strength and reliability of the NM-MRI signal, including acquisition time, slice-thickness, spatial-normalization software, and degree of spatial smoothing. Using a newly developed, detailed acquisition protocol, across two scans separated by 13 days on average, we obtained intra-class correlation values indicating excellent reliability and high contrast, which could be achieved with a different set of parameters depending on the measures of interest and experimental constraints such as acquisition time. Based on this, we provide detailed guidelines covering acquisition through analysis and recommendations for performing NM-MRI experiments with high quality and reproducibility. This work provides a foundation for the optimization and standardization of NM-MRI, a promising MRI approach with growing applications throughout clinical and basic neuroscience.

Keywords: NM-MRI; Neuromelanin; Substantia nigra; Test-retest; Ventral tegmental area; Voxelwise analysis.

Conflict of interest statement

Declaration of competing interest The authors declare no conflicts of interest.

Copyright © 2019 Elsevier Inc. All rights reserved.

Figures

Fig. 1.

Illustration of the step-by-step procedure for placement of the NM-MRI volume. Yellow lines indicate the position of slices used for placing the volume. (A) Sagittal image showing the greatest separation between the midbrain and thalamus. (B) Coronal plane that identifies the most anterior aspect of the midbrain on the image from A. (C) Axial plane that identifies the inferior aspect of the third ventricle on the image of the coronal plane from B. (D) Location of the axial plane from C identified on the image from A. (E) The axial plane denoting the superior boundary of the NM-MRI volume. This axial plane is the axial plane from D shifted superiorly 3 mm.

Fig. 2.

NM-MRI volume placement and NM-MRI images from the 4 NM-MRI sequences tested in this study from a representative subject. (A) Final NM-MRI volume placement. (B–E) Raw signal in axial slices at approximately the same level of the midbrain for each NM-MRI sequence, after motion correction and averaging: (B) NM-1.5 mm, (C) NM-2 mm, (D) NM-3 mm, and (E) NM-3 mm Standard. The intensity scales for the NM-MRI images were set such that the ratio between the lower bound and upper bound was 25%. Red arrows on B – E denote the left SN-VTA complex.

Fig. 3.

Spatial normalization and anatomical masks for analysis of NM-MRI images. (A) Average NM-MRI image created by averaging the spatially normalized NM-MRI images from 10 individuals in MNI space. Note the high signal intensity in the SN-VTA complex. (B) Masks for the SN-VTA complex (yellow voxels) and the CC (pink voxels) reference region (used in the calculation of CNR) are overlaid onto the template in A. These anatomical masks were made by manual tracing on a NM-MRI template from a previous study (Cassidy et al., 2019). Note the hyper-intensity along the edge of the midbrain and that the SN-VTA-complex mask does include the anterior medial edge, which is also not included in the probabilistic masks in D. (C) The same average NM-MRI image from A but different slices. (D) Probabilistic masks for the VTA, SNr, SNc, and PBP as defined from a high-resolution probabilistic atlas (Pauli et al., 2018) overlaid onto the template in C. The color scaling for probabilistic masks goes from P = 0.5 (darkest) to P = 0.8 (lightest). Red arrows in A and C denote the left SN-VTA complex.

Fig. 4.

ICCROI (top row) and CNRROI (bottom row) within the manually traced mask of the SN-VTA complex (Fig. 3B) as a function of acquisition time. Data points denote the median and error bars indicate the 25th and 75th percentiles.

Fig. 5.

ICCASV (top row), ICCWSV (middle row), and CNRV (bottom row) of voxels within the manually traced mask of the SN-VTA complex (Fig. 3B) as a function of acquisition time. Data points denote the median and error bars indicate the 25th and 75th percentiles.

Fig. 6.

Scatterplots of ICCASV and CNRV for each of the NM-MRI sequences and spatial normalization software. Each data point represents one voxel within the manually traced mask of the SN-VTA complex (Fig. 3B). The solid lines indicate the linear fit of the relationship between ICCASV and CNRV.

Fig. 7.

(A) Predictive value (R2) of anatomical position on ICCASV of voxels within the manually traced mask of the SN-VTA complex (Fig. 3B) for NM-1.5 mm sequence and each of the spatial normalization software. Data points denote the median and error bars indicate the 25th and 75th percentiles. (B) Histogram of ICCASV of voxels within the manually traced mask for NM-1.5 mm sequence and ANTs spatial normalization software, which is the best performing method as per A. (C) Histogram of ICCASV of voxels within the manually traced mask for NM-1.5 mm sequence and SPM12 spatial normalization software, which is the worst performing method as per A. Yellow denotes excellent reliability (ICC over 0.75), orange denotes good reliability (ICC between 0.75 and 0.6), red denotes fair reliability (ICC between 0.6 and 0.4), and burgundy denotes poor reliability (ICC under 0.4).

Fig. 8.

The effect of spatial smoothing on ICCASV and CNRV of voxels within the manually traced mask of the SN-VTA complex (Fig. 3B) for different degrees of spatial smoothing. Data points denote the median and error bars show the 25th and 75th percentiles.

Fig. 9.

ICCROI (top row) and CNRROI (bottom row) within the probabilistic masks of the dopaminergic nuclei (Fig. 3D) with different probability cutoffs (0.5, 0.6, 0.7, and 0.8) as a function of acquisition time. Data points denote the median and error bars indicate the 25th and 75th percentiles.

Fig. 10.

Correlations and histograms of the CNRROI values within the 4 dopaminergic nuclei (Fig. 3D) for the lowest (P = 0.5) and highest (P = 0.8) probability cutoffs. The value within each correlation plot is Spearman’s rho.