Abstracts

Rationale: Until recently, functional imaging techniques have been developed predominantly to improve visualization of epileptogenic foci in grey matter for focal-onset epilepsy. However, few techniques assess the degree of ictal-associated connectivity between different zones in the epileptogenic cortex through white matter (WM) axonal paths, thus neglecting a crucial step towards understanding epileptic networks. Probabilistic subtracted post-ictal Diffusion Tensor Imaging (pspi-DTI) is a novel imaging technique with the objective of identifying transient water diffusion changes during the early post-ictal state in such pathways. Methods: A post-ictal (up to 4hrs following seizure termination) and an inter-ictal (no electrocerebral seizures for at least 24hrs) high resolution DTI datasets were acquired. Imaging parameters consisted of 2mm-thick oblique slices in 60 non-collinear directions with a diffusion factor of 900 s/mm2 in a 3T MRI scanner. Eddy current and motion corrections were applied to these volumetric datasets using TORTOISE v2.5.1. The resulting volumes were registered to a MPRAGE MRI dataset. Fractional anisotropy (FA) and mean diffusivity (MD) measures were obtained using FSL FMRIB's Diffusion Toolbox (Oxford, UK, 2016). Differences in FA and MD between the post-ictal and inter-ictal volumes were computed using a three-dimensional probabilistic kernel-based spatial filtering operation. Using the kernel in both post-ictal and inter-ictal volumes, a central voxel and a sample of its surrounding neighborhood voxels were selected. Subsequently, a one-tailed paired t-test was performed in the sample obtaining the associated p-value. The kernel was then translated to each location in the datasets to create a resulting p-value volume. Regions with statistically significant changes (p < 0.01) were identified. Results: Our pspi-DTI technique detected tracts with a statistically significant transient decrease in FA and increase in MD following stereotypic complex partial seizures without generalization secondarily for each patient. These tracts were concurrently detected with deterministic tractography. The method was tested in 8 different patients, who were candidates for responsive neurostimulation (RNS) therapy, to determine significant changes in white matter diffusion properties. Conclusions: pspi-DTI is a new technique developed in our laboratory to test transient changes in diffusion properties of white matter tracts caused by an ictal state. As discussed, our pspi-DTI technique detected tracts with a statistically significant transient decrease (p < 0.01) in FA and increase (p < 0.01) in MD following stereotypic complex partial seizures without generalization secondarily for each patient. The identified tracts convey valuable information about the ictal propagation pathways involved in the epileptogenic network that can be later used to predict optimal depth placement implant sites for direct neurostimulation therapy. Funding: 1. RUSH University Medical Center