Decoding the microstructural correlate of diffusion MRI

Abstract

Diffusion imaging has evolved considerably over the past decade. While it provides valuable information about the structural connectivity at the macro- and mesoscopic scale, bridging the gap to the microstructure at the level of single nerve fibers poses an enormous challenge. This is particularly true for the human brain with its large size, its large white-matter volume and availability of histological techniques for studying human whole-brain sections and subsequent 3D reconstruction. Classic post-mortem techniques for studying the fiber architecture of the brain, such as myeloarchitectonic staining or dye tracing, are complemented by novel histological approaches, such as 3D polarized light imaging or optical coherence tomography, enabling unique insight into the fiber architecture from large fiber bundles within deep white matter to single nerve fibers in the cortex. The present review discusses the benefits and challenges of these latest developments in comparison with the classic techniques, with particular focus on the mutual exchange between in vivo and post-mortem diffusion imaging and post-mortem microstructural approaches for understanding the wiring of the brain across different scales.

The role of diffusion MRI in neuroscience

“…However, although empirical or simulation studies can clearly demonstrate that variation in these tissue properties modulates diffusion parameters, we are still faced with an inverse problem when trying to interpret an observed difference in diffusion parameters (e.g. reduced FA), as there is no one‐to‐one relationship between a given diffusion parameter and the underlying tissue structure . Nevertheless, diffusion parameters provide useful non‐invasive measures of local tissue microstructure, with sensitivity to features including membrane integrity, myelin thickness, axon diameter and packing density.…”

“…One such approach for modelling the diffusion signal (instead of DTI) is to devise a mathematical formalism that is guided by tissue geometry. Yet, anyone who has looked down a microscope at a brain section would have been impressed by the complexity of the white matter geometry . Cells of different shapes and sizes with processes spreading at different scales and orientations depict a mess of geometries at the micrometre scale.…”

“…reduced FA), as there is no one-to-one relationship between a given diffusion parameter and the underlying tissue structure. [33][34][35][36][37] Nevertheless, diffusion parameters provide useful non-invasive measures of local tissue microstructure, with sensitivity to features including FIGURE 1 First drawings of white matter connection in the human brain. Left, brain dissection from de Vieussens; 3 right, brain dissection from Johann Christian Reil 4 membrane integrity, myelin thickness, axon diameter and packing density.…”

“…Yet, anyone who has looked down a microscope at a brain section would have been impressed by the complexity of the white matter geometry. 36,62 Cells of different shapes and sizes with processes spreading at different scales and orientations depict a mess of geometries at the micrometre scale. Even at a lower level of magnification, where cell layers and fibre bundles are visible, complexity arises where fibres cross, disperse and fan within hundreds of micrometres.…”

Diffusion MRI of the brain: The naked truth

“…Integrating complementary information from other imaging modalities could be instrumental to further understand how diffusion MRI measurements are related to the underlying microstructure and physiology in brain tissue. In‐depth discussions on the benefits and limitations of such multimodal strategies are provided by Caspers and Axer, and Liu and Özarslan …”