Neurite Density and Kurtosis in the Gray Matter of People with Early Schizophrenia.

Abstract

Background: Classical models of diffusion weighted imaging, especially diffusion tensor imaging, are unsuited for application to the cortical gray matter, given the regions high microstructural complexity. As such, most neuroimaging studies thus far have focused on gross structural effects of schizophrenia, such as cortical thickness differences. More recently developed models, such as the neurite orientation dispersion and density imaging (NODDI) model and diffusion kurtosis imaging (DKI), incorporate higher resolution data and may provide more sensitive descriptions of schizophrenia pathology with more specific interpretations.

Methods: We applied the NODDI and DKI models to the cortical gray matter of people with early schizophrenia (n=54) and healthy controls (n=51) from the Human Connectome Project - Early Psychosis dataset. Comparisons between groups were made using region-of-interest and clustering approaches. The effect sizes of these approaches were compared to those of cortical thickness differences. We also investigated the relationship between these parameters and lifetime antipsychotic usage.

Results: Cortical thickness differences were most prominent between groups in terms of global effect size and spatial extent. We also observed a diffuse, right-hemisphere dominant increase in mean kurtosis and isotropic diffusion fraction throughout the gray matter, not fully explained by partial volume effects. Additionally, a lower neurite density index (NDI) correlated with greater lifetime antipsychotic usage.

Conclusions: Increases in mean kurtosis and isotropic diffusion fraction are both markers of schizophrenia, consistent with inflammation models of the gray matter in schizophrenia. NDI reduction, reflecting intraneurite pathology, becomes prominent only in those with greater disease burden.