Version française

Sunday June 01 2003

Functional brain architecture: the potential of diffusion MRI

Nature Reviews Neuroscience (2003) 4 : 469 -480
The potential offered by water diffusion MRI (dMRI), and the future advances that this technique, coupled with functional MRI, will allow as very high field magnets become available, is discussed in an article published in the journal Nature Reviews Neuroscience (Frédéric-Joliot Hospital Service, DRM). Water diffusion MRI is an unrivalled tool for determining the fine architecture of neurone tissue and how this changes during various physiological or pathological states. In particular, it has been used since the early nineties to evaluate the usefulness of treatments after brain ischaemia. In addition, its specific features enable it to follow, with high accuracy, the tissue paths of brain activation during growth and maturation, and when certain tasks are being carried out.


Introduced in the early eighties, diffusion MRI is based on the finding that the diffusion of molecules can be used to probe tissue structure at a microscopic scale, much finer than that used in classical imaging (i.e., millimetric).  During the diffusion time (about 50 ms), the molecules move in the brain through distances of about 10 µm on average, where they traverse, interact with or are slowed by numerous environing tissue components (cell membranes, fibres, macromolecules, etc.). A study of their movements thus makes it possible to identify the structural characteristics and geometric organisation of the surrounding neuronal tissues, and the modifications these characteristics undergo according to the subject's physiological or pathological state.

One of the best developed medical applications of this method concerns cerebral ischaemia.  Water diffusion is significantly slowed immediately after an ischaemic event, probably owing to changes in membrane permeability or to cell swelling. The assessment of these changes by diffusion MRI gives an immediate indication of what treatment can be usefully applied while there is still time to save brain tissue.

Diffusion is a three-dimensional process, and molecular mobility in tissues is not the same in all spatial directions. These differences in diffusion rate reflect the organisation of neurones and so can be used to map their spatial networks in the brain. Knowledge of the paths of neurone fibres obtained in this way by diffusion MRI, combined with functional MRI data, opens up vast new possibilities for the study of brain connections, potentially a most important application of this technique. It is also a promising tool for exploring different aspects of the brain during its growth and maturation, and for characterising the changes in connections observed in certain situations, e.g., in dyslexia.

Recent data further indicate that diffusion MRI might be used to visualise changes in tissue dynamics associated with neurone activation, so usefully extending the classical methods of functional imaging based on local variations in blood flow.

With the advent of very high field magnets, which will push back the current spatial and temporal limits of MRI, diffusion MRI will certainly allow new advances in both neuroscience and clinical practice.