Diffusion kurtosis imaging and log-normal distribution function imaging enhance the visualisation of lesions in animal stroke models.

Abstract

In this work, we report a case study of a stroke model in animals using two methods of quantification of the deviations from Gaussian behaviour: diffusion kurtosis imaging (DKI) and log-normal distribution function imaging (LNDFI). The affected regions were predominantly in grey rather than in white matter. The parameter maps were constructed for metrics quantifying the apparent diffusivity (evaluated from conventional diffusion tensor imaging, DKI and LNDFI) and for those quantifying the degree of deviations (mean kurtosis and a parameter σ characterising the width of the distribution). We showed that both DKI and LNDFI were able to dramatically enhance the visualisation of ischaemic lesions in comparison with conventional methods. The largest relative change in the affected versus healthy regions was observed in the mean kurtosis values. The average changes in the mean kurtosis and σ values in the lesions were a factor of two to three larger than the relative changes observed in the mean diffusivity. In conclusion, the applied methods promise valuable perspectives in the assessment of stroke.