Multi-Scale Cortical Dynamics and Seizure Prediction in Human Focal Epilepsy

Providence Va Medical Center, Providence RI

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): The care of Veterans with epilepsy is a priority for the Department of Veterans Affairs as attested by the creation of several Epilepsy Centers of Excellence (ECoE) across the country. The constant but seemingly unpredictable re-occurrence of seizures dramatically impacts the autonomy and quality of life of Veterans and civilians suffering from epilepsy. Unfortunately, 25-35% of patients with epilepsy continue to have seizures despite maximal medical therapy (Annegers 2001). Regional surgical resection, besides carrying significant risk, is curative in only ~ 40-70% of cases (Engel et al., 2003). Even when pharmacological or surgical methods work, side effects can be severe. Alternative seizure prevention and warning systems would be highly desirable to restore quality of life and autonomy of Veterans and civilians with intractable epilepsy. Seizure prevention and warning systems based on intracranial EEG (iEEG) recordings have been intensively examined in the past 30 years, yet reliable seizure prediction and early detection remain elusive. Overall, progress has been hampered by the difficulty of monitoring the activity of ensembles of single neurons in humans. We will use a novel technology, intracortical 96-microelectrode arrays (MEAs), to examine human epilepsy at a much higher spatial resolution than in iEEG recordings. We will record the activity of ensembles of single neurons (single units, SUs), and multichannel multi-unit (MU) and high-density field potentials (LFPs) in patients with pharmacologically intractable focal epilepsy undergoing pre-resection surgery monitoring. AIM 1 will study how seizures start, spread and terminate at multiple scales, from the microphysiological level of ensembles of SUs, MUs and LFPs to the macroscopic dynamics reflected in iEEGs signals. We will test the hypothesis that neuronal ensemble dynamics preceding seizure onset change gradually and that these changes can be detected several minutes before the seizure onset. We further hypothesize that this transition should manifest in changes in spatiotemporal correlations among SUs, MUs and LFPs. AIM 2a will develop and test a new framework for seizure prediction and early warning based on multi-scale MEA neural signals in people with focal epilepsy. AIM 2b will compare this new framework for seizure prediction with state of the art prediction approaches based exclusively on iEEG. This comparison will reveal the advantages and disadvantages of seizure prediction approaches that include MEA neural signals versus approaches that do not.