Caractérisation spatio-temporelle du signal de diffusion fonctionnel au cours de l'activité cérébrale visuelle et comparaison avec le signal BOLD – DFMRI

Diffusion-weighted magnetic resonance imaging (DMRI) has become a key tool for investigating the structural organization of the human brain over the past twenty years. First, it has first been used in the clinical context of early detection of acute brain ischemia, but also rapidly provided a way to infer assess anatomical connectivity through the mapping of white matter fiber bundles (see Le Bihan 2003 for a review). It is so far the only technique providing such kind of information in vivo. More recently, the DMRI signal has been shown to reflect brain activity (Le Bihan 2006) and diffusion MRI is becoming another means to look at brain function. Diffusion functional magnetic resonance imaging (DfMRI) seems to reflect more directly the activation process, delivering a peak in the activity earlier than with the classical, hemodynamically driven BOLD (Blood-Oxygen Level Dependant) imaging method. The biological mechanisms underlying this new approach are still debated in the community. Based on extensive literature reports we have hypothesized that some changes in the network structure of cellular water could accompany neuronal activation. In particular two different pools of water molecules have been identified within cells, with a slow diffusing pools which would result from interaction of water molecules with cell membranes (Le Bihan 2007). The purpose of this project is twofold. First we would like to investigate the microstructural changes which occur in brain tissue during activation and correlate them with the observed changes in water diffusion. This part which will be conducted in primates and combine several imaging techniques will provide us with a better characterization of the DfMRI signal in terms of temporal and spatial features. Second, we would like to develop a numerical simulation tool based on an accurate model of the biophysical processes underlying water diffusion in cells and tissues. This tool will give us the opportunity to explain the changes in water diffusion behavior which have been observed in different known physiological (brain activation) or pathological (ischemia) situations and to evaluate assumptions or new conditions which may not be tested in vivo, so as to predict new findings or optimize imaging acquisition parameters. Overall results will be used to optimize DfMRI in human subjects.