Characterization of cerebral white matter properties using quantitative magnetic resonance imaging stains

Abstract

The image contrast in magnetic resonance imaging (MRI) is highly sensitive to several mechanisms that are modulated by the properties of the tissue environment. The degree and type of contrast weighting may be viewed as image filters that accentuate specific tissue properties. Maps of quantitative measures of these mechanisms, akin to microstructural/environmental-specific tissue stains, may be generated to characterize the MRI and physiological properties of biological tissues. In this article, three quantitative MRI (qMRI) methods for characterizing white matter (WM) microstructural properties are reviewed. All of these measures measure complementary aspects of how water interacts with the tissue environment. Diffusion MRI, including diffusion tensor imaging, characterizes the diffusion of water in the tissues and is sensitive to the microstructural density, spacing, and orientational organization of tissue membranes, including myelin. Magnetization transfer imaging characterizes the amount and degree of magnetization exchange between free water and macromolecules like proteins found in the myelin bilayers. Relaxometry measures the MRI relaxation constants T1 and T2, which in WM have a component associated with the water trapped in the myelin bilayers. The conduction of signals between distant brain regions occurs primarily through myelinated WM tracts; thus, these methods are potential indicators of pathology and structural connectivity in the brain. This article provides an overview of the qMRI stain mechanisms, acquisition and analysis strategies, and applications for these qMRI stains.

FIG. 1.

Schematic illustration of water diffusion interacting with myelinated axons. The apparent diffusion is greatest in the direction parallel to the axons (left side). The diffusion distances and corresponding apparent diffusion coefficient are reduced for more densely packed axons (right side).

FIG. 2.

Quantitative diffusion tensor imaging (DTI) stain maps from a single DTI data set.

FIG. 3.

Diffusion magnetic resonance imaging (MRI) stains from the same hybrid diffusion imaging study and slice–fractional anisotropy (FA), zero-displacement probability (Po), mean-squared displacement (MSD), and q-space inverse variance (QIV). Note the higher apparent noise and heterogeneity in the MSD map in comparison to the QIV.

FIG. 4.

White matter (WM) tract reconstructions for several major WM pathways.

FIG. 5.

Top row: track-density imaging (TDI) examples without super-resolution (left) at the native resolution of the acquired diffusion MRI data (2.3 mm isotropic) and with super-resolution (right) using a grid-size of 125 μm. Note: the same diffusion MRI data and whole-brain tracking dataset (with 2.5 million tracks) were used to create both images; the only difference was the grid-size used to calculate the TDI map. The sub-voxel detail achievable with super-resolution is readily seen. For comparison, the bottom row shows the FA map generated from the same diffusion MRI data used to create the TDI maps, and an anatomical high-resolution three-dimensional (3D) T1-weighted image (1 mm isotropic resolution). The super-resolution TDI map shows not only sub-voxel detail but also novel image contrast (e.g., see high contrast within the thalamus [arrow] and in the optic radiations). Image courtesy F. Calamante.

FIG. 6.

Construction of a structural connectivity matrix (connectome) based upon diffusion-weighted imaging tractography with epsilon-radial nodes. Tractography is performed in normalized space. Nodes are generated using spherical nodes based upon clusters of tract terminations. Tract-counts are used to generate the connectivity matrix.

FIG. 7.

Schematic illustration of magnetization exchange that is detected by magnetization transfer imaging methods.

FIG. 8.

Schematic of the magnetization transfer (MT) saturation process. An intense refocusing (RF) pulse is applied off-resonance, which saturates the magnetization of the macromolecule pool. Rapid exchange between magnetization of the macromolecule proton pool and the free water protons nearby attenuates the free water signal.

FIG. 9.

Example images from an MT ratio (MTR) experiment. The image obtained without any MT saturation (left) shows little contrast between WM and gray matter (GM). Additional application of MT pulse (18-ms Fermi pulse, offset 2.5 kHz, MT flip 1100°, repetition time [TR]=40 msec) causes strong reduction of MR signal in tissues with natural abundance of myelin such as WM (middle), which in turn results in higher values in the corresponding MTR (right).

FIG. 10.

Effect of B1 inhomogeneity on MTR. Uncorrected MTR map (a) demonstrates by slow spatially varying intensity inhomogeneity (a). Correction of MTR using separately acquired B1 map eliminates the intensity bias (b) and leads to improved localization of WM and GM peaks on the corresponding whole brain histograms (c).

FIG. 11.

Quantitative maps or stains of MT effect obtained in a healthy volunteer.

FIG. 12.

Anatomy of major fiber tracts on 3D bound pool fraction maps produced by constrained cross-relaxation imaging. The following anatomic structures are labeled: ILF, inferior longitudinal fasciculus; OR, optic radiations; AC, anterior commissure; CCG, corpus callosum genu; CCS-corpus callosum splenium; AR, auditory radiations; ATR, anterior thalamic radiations; SLF, superior longitudinal fasciculus; SFOF, superior fronto-occipital fasciculus; C, cingulum; MCP, middle cerebellar peduncle. Courtesy of Dr. Vasily Yarnykh. Figure reproduced from Yarnykh and Yuan (2004), with permission from Elsevier.

FIG. 13.

Examples of typical steady-state relaxometry maps acquired at 2 mm isotropic resolution: left, T1; middle, T2; right, myelin water fraction (MWF).

FIG. 14.

Schematic illustration of different water environments within WM. Geometrically restricted compartments exhibit a much shorter T2 due to restricted degrees of translational and rotational freedom.

FIG. 15.

Illustration of strengths and weakness of using tractography for regional brain segmentation. In this case, the trunk of the inferior longitudinal fasciuculus is a tight coherent bundle; however, the regions near the ends of the tract show considerable branching and divergence, which is likely to vary considerably across subjects.

FIG. 16.

Spatial normalization example for DTI data. In this example a two-step normalization is illustrated. (a) Pairs of images (e.g., longitudinal or twin) are normalized with each other, followed by normalization to an overall population averaged template. The spatial normalization employed a diffeomorphic warping algorithm with tensor reorientation. (b) The normalization shows good correspondence of most major WM features, though there is still considerable variation in smaller and more peripheral tracts.