• White matter is composed by axons and glial cells

• Most MRI techniques show a sharp contrast between grey and white matter

• Macroscopic white matter abnormalities can be detected on T2-weighted MRI

• Magnetization transfer and multi-compartment T2 relaxometry offer indices sensitive to myelin

• Diffusion MRI is sensitive to fiber density and orientation

Clinical and basic scientists interested in white matter biomarkers based on MRIAfter this lecture, participants should be able to understand the composition of white matter, the most common pathology occurring in the white matter, and what are the most suitable MRI techniques to detect such pathology. They should also be aware of the potential limitations of these techniques.The white matter of the human brain is composed by tightly packed myelinated and non-myelinated axons and glial cells. The glial cells include oligodendrocytes, astrocytes, microglia, and oligodendrocyte progenitor cells (1). Pathology in the white matter thus consists mainly of demyelination, axonal degeneration and loss, and gliosis. Even pushing the resolution of MRI to its current limits, a single voxel contains a huge amount of these cells and therefore non-invasive white matter imaging is confined to exploiting the indirect average signal generated by their combination. Nevertheless, MRI offers a wide range of techniques that could reveal different aspects of white matter. This lecture will explore the established techniques and their limitations.T2-weighted MRI is extremely sensitive to any change to white matter, and it is routinely used to detect abnormalities, such as demyelinating lesions, neoplastic masses, vascular lesions, and abnormalities of infectious origin. The main limitation of conventional T2-weighted imaging is its limited specificity. In the clinic, T2-weighted imaging is often complemented by gadolinium-enhanced T1-weighted imaging, which provides information about blood-brain barrier leakage. High-resolution T1-weighted MRI offer an exquisite contrast between grey and white matter, at least at field strengths <= 3T. The source of such sharp contrast is believed to be the myelin in the white matter and its high lipid and protein content (2). These images can be used to segment the brain in separate tissues, and therefore to quantify white matter atrophy. While it is possible to achieve an excellent contrast with the cortex, the subcortical grey matter is more difficult to isolate, and can lead to errors in the segmentation, especially without the support of priors. Recently, quantitative techniques based on magnetization transfer (MT) have been developed to address this issue (3). MT provides a contrast mechanism based on the exchange of magnetization occurring between groups of spins characterised by different molecular environment (4), and is particularly sensitive to the myelin in the white matter. While MT imaging has been available for 3 decades now, the traditional approach to quantify it was based on the so-called MT ratio (MTR), i.e., the percentage difference of two images, one with off-resonance saturation and one without. By increasing the number of acquisitions to 3, it becomes possible to separate the contributions of MT and T1, and therefore to reduce the impact of other factors, including the T1-shortening effects of iron, on image contrast. But the role of MT in characterising the white matter is not limited to improve the grey-to-white matter contrast. Analytical models of the MT-weighted signal exist (5), where each pool is characterised by their longitudinal relaxation rates (RA and RB), their transverse relaxation times (T2A and T2B), and their spin densities (M0A and M0B). The exchange rate constant between the pools is R. Assuming that myelin is the main contributor to MT in white matter, the ratio M0B/M0A, known as F, or as the relative density of the macromolecular pool, is believed to reflect myelin content. Myelination and demyelination have been extensively studied by MRI. Alongside quantitative MT, alternative techniques, such as multi-compartment T2 relaxometry (6) enable the quantification of the myelin water fraction (MWF), a proxy of myelination. Finally, diffusion MRI (dMRI) is based on the random motion of water molecules within tissue. As such a motion is affected by tissue structure; dMRI can indirectly reflect white matter microstructure. dMRI has evolved over the years from simple models such as diffusion tensor imaging (DTI) to more complex, multi-compartment models. Thanks to its sensitivity to tissue orientation, dMRI also enables the reconstruction of the main white matter tracts with diffusion tractography. Recent work has demonstrated that dMRI can be used to infer the distribution of axonal radius and fiber density within an image voxel (7). Finally, it was recently suggested that MT parameters other than F might be sensitive to inflammation, such as activated microglia and astrocytosis (8).A significant amount of effort has been dedicated to validate MRI biomarkers of myelination. Animal and post mortem data suggest that proxies derived from both MT and T2 relaxometry are correlated by myelin content from histology (9,10). Despite these observations, F and measures of myelination derived from short T2 mapping appear largely uncorrelated (11) , prompting the debate on which technique better estimates myelin in vivo. By contrast dMRI parameters are related to fiber density. Indices derived from DTI are heavily affected by orientation dispersion, and novel models that account for these factors are recommended. MRI techniques for imaging the white matter range from standard clinical scans to very sophisticated quantitative approaches, each offering a different view of white matter structure and pathology. The main limitation of these techniques is the lack of specify for a specific substrate. This is particularly problematic as often demyelination, inflammation and oedema can occur at the same time. Novel methods under development fall essentially into 2 separate categories: ultra-high field techniques, such as quantiative susceptibility imaging and chemical-exchange saturation transfer, and multiparametric techniques, where several indices ae combined to improve specificity.