Directed connectivity analysis of resting-state SEEG and DWI to improve lateralization and localization in focal epilepsy

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT Focal epilepsy is the most common form of epilepsy, a debilitating disorder that affects 50 million people worldwide. Approximately 30-40% of patients with focal epilepsy continue to have debilitating seizures despite maximal medical therapy. Epilepsy surgery can eliminate or reduce seizures using resection, ablation, or neurostimulation of regions that generate seizures (“Epileptogenic Zones”, EZs). However, 33-50% of patients that undergo surgery continue to have seizures post-operatively. An important determinate of post-operative outcome is accurate pre-surgical lateralization and localization of EZs. In 50% of patients, lateralization and localization requires invasive intracranial monitoring with stereo-electroencephalography (SEEG) in the hospital for days to weeks to record multiple seizures. This invasive diagnostic process causes significant morbidity to the patient, and interpretation of ictal (seizure) activity from SEEG may sometimes be challenging, inaccurate, and incapable of capturing all the patient’s seizure types. Resting-state (between seizures) SEEG analysis may supplement clinical interpretation by identifying EZs without requiring ictal recordings. Beyond SEEG, diffusion MRI (DWI) and neurostimulation have also been used to attempt EZ lateralization and localization. These studies rely on generating connectivity networks of brain regions and extracting features that predict EZ locations, but EZ lateralization and localization accuracy with these data has been suboptimal. However, few studies have evaluated the directionality of connectivity patterns involving EZs. Therefore, building from previous neurophysiological work that shows tonic inhibition of EZs in focal epilepsy, we hypothesize that electrophysiological resting-state inhibitory inward directed connectivity of EZs will be markedly increased vs. that of Non-EZs, and thus key to predicting epileptogenicity of brain regions. Further, integrating previous work done across the fields of neuroscience and neuropsychology, we also hypothesize specific DWI-derived structural network alterations that are important to lateralize EZs and predict surgical outcome. Our first goal is to develop directed connectivity measures to reliably identify EZs using brief resting-state SEEG recordings and neurostimulation sessions (Aim 1). We then seek to identify noninvasive structural connectivity measures to lateralize EZs and predict surgical outcome using DWI to ultimately reduce the need for invasive intracranial monitoring. We will do this through advanced network analysis of DWI-generated structural connectivity maps (Aim 2). This proposed fellowship will provide research training in a collaborative research atmosphere with expert mentors in translational neuroscience and engineering research. Research training will be conducted in an environment that combines an academic medical center with a level 4 epilepsy center, world class imaging institute, and engineering all on one campus, ensuring an environment uniquely suited to excellent training in all aspects of this proposed work. Studying multiple modalities to characterize epileptic networks and localize EZs has the potential to drastically improve the lives of patients living with this devastating neurological disorder.