Individual diffusion-weighted dataset were preprocessed using Connectomist [5] to remove imaging artifacts (motion, eddy current, susceptibility...), thus enabling accurate matching of DW and T1-weighted data using rigid registration. Individual NODDI maps were computed using an inhouse command included in the Connec- tomist software as depicted on Fig. 1: the intracellular fraction (f_{intra}) referring to the space bounded by the membranes of neurites, their orientation dispersion index (ODI) which represents the angular variation of neurite orientation, the CSF fraction (fiso) which quantifies the CSF volume fraction and the Watson concentration parameter (к evolving at the opposite of the ODI). We verified that the ODI values were close to 0 in the white matter where the axons orientation are parallel and close to 1 in the grey matter where the dendrites distribution is more arbitrary.

Figure 1 shows all the individual NODDI maps merged with the corresponding anatomical T1-weighted image for a better anatomical colocalization.

The NODDI model can be easily described using the three compartments Eq. (1) below:

where A is the normalized diffusion attenuation resulting from the contribution of the three compartments: the CSF (A_{iso}), the intracellular space (A_{intra}) and the extracellular space (A_{extra}). f_{iso} denotes the volume fraction of the isotropic diffusion compartment and f_{intra} denotes the volume fraction of the intracellular compartment.

The intra-axonal or intra-dendritic compartment was mathematically modeled using a cylinder geometry. The initial parallel diffusivity was set to 1.7x 10^{_3}m^{2}s^{_1 }and the orientation dispersion of the cylinder direction was modeled using a Watson distribution, the isotropic diffusivity was set to 3.0x10^{_3}m^{2}s^{_1} and the normalized noise standard attenuation was measured equal to 0.03.

Fig. 1 Individual NODDI quantitative maps merged with Tl-weighted MRI