On the use of multi-echo NODDI MRI with released intrinsic diffusivity for the assessment of tissue diffusion and relaxation properties in experimental ischaemic stroke.

The multi-echo neurite orientation dispersion and density imaging (MTE-NODDI) model has been proposed to overcome one of the shortcomings of conventional NODDI, namely the echo time (TE) dependence of the compartmental signal fractions, which stems from the intrinsic differences in the compartmental transverse relaxation times (T2). However, the model continues to be constrained by the limitation of having a fixed, brain-wide intrinsic diffusivity, d. The primary aim of this work is to assess the benefits and shortcomings of using MTE-NODDI to investigate the diffusion and T2 properties of ischaemic stroke tissue following middle cerebral artery occlusion (MCAo) in rat models. Given the known alterations in the diffusion properties in ischaemic tissue, a secondary aim is to assess an estimation approach for MTE-NODDI parameters that enables d to be released while also mitigating the consequent model degeneracy. Using the MTE-NODDI parameters, the spatiotemporal evolution of diffusion and T2 properties in ischaemic tissue was characterised from day one to day 23 post-MCAo. The proposed approach enables access to several unique tissue features that would otherwise be obscured by the conventional approach. Importantly, a marked reduction in d was observed, leading to significant changes in other MTE-NODDI parameters compared to the model employing a fixed d. The isotropic signal fraction displayed a significant increase in ischemic tissue, which appears in contradiction with previous works. Regarding the intra- and extra-neurite T2 values, T2,in and T2,en, a significant increment was observed at the ischaemic tissue, while the condition T2,in≥T2,en displayed a tendency to hold in both tissue types. More generally, some parameters, such as the isotropic signal fraction, the intrinsic diffusivity and both compartmental T2 values, display unique, heterogeneous spatiotemporal evolution, where the core and border zones of the ischaemic tissue show different behaviours. Overall, the newly estimated parameters show greater consistency with analogous estimates reported by published models, and are anticipated to significantly enhance the understanding of tissue properties following ischaemic stroke.