Background: Imaging in the field of epilepsy has been a lasting challenge. In genetic generalised epilepsy (GGE) patients, conventional neuroimaging methods such as the MRI often appear normal and even in focal epilepsy (FE) an obvious structural abnormality is not always visualised. The objective of this exploratory case study is to identify white and grey abnormalities in epilepsy patients based on diffusion weighted imaging in comparison with healthy subjects.
Methods: Diffusion tensor imaging (DTI) and diffusion tensor tractography (DTT) at 3 Tesla is performed on 2 genetic generalised epilepsy patients, 1 focal epilepsy patient and 2 healthy subjects. Four DTI metrics (Fractional Anisotropy (FA), Mean Diffusivity (MD), Radial Diffusivity (RD) and Axial Diffusivity (AD) are compared based on the mean per region of interest between subjects and between hemispheres. A DTT map of the thalamo-cortical radiations is constructed based on a probabilistic tractography (iFOD2) as implemented in MRTrix. Based on this data the structural connectivity matrix is generated. A graph theoretical analysis is performed comparing node degree, global and local efficiency, characteristic path length, (average) clustering coefficient and (average) betweenness centrality. Constructed connectomes are compared for significant different connections between nodes using network based statistics.
Results: Decreased FA and MD values are reported in the hippocampus of the FE patient. DTT showed decreased tract density in the thalamo-cortical radiations in one of the GGE patients and the FE patient, together with a decreased mean streamline length compared to the two healthy subjects, which can be interpreted as a reduced structural connectivity. Structural connectivity is decreased in the FE patient compared to the other four subjects based on graph analysis. This is demonstrated by the relatively high average clustering coefficient and characteristic path length, together with a low node degree, high local efficiency and high clustering coefficient in the thalamus, the basal ganglia and the hippocampus. Conclusion: These results are potentially of aid in identification of white and grey matter abnormalities in epilepsy patients. Especially in the FE brain several abnormalities in both DTI, DTT and graph analysis are observed compared to the healthy and GGE brain. However, for the use of DTI, DTT and graph analysis as a possible imaging biomarker, a standard processing pipeline with validated reference values ranges from the healthy population needs to be developed. Also the biological interpretation of changes in network measures needs to be validated.