PROJECT SUMMARY/ABSTRACT Approximately 40% of patients with epilepsy are resistant to drug treatment. These patients may benefit from targeted surgical interventions, such as surgical resection or neurostimulation of the seizure onset zone (SOZ). Patients being considered for surgical intervention undergo a series of noninvasive diagnostic tests to localize their SOZ. If these noninvasive tests cannot delineate the SOZ clearly enough for surgical intervention, patients undergo intracranial electroencephalography (iEEG) to more accurately localize their SOZ. Yet, even in patients who undergo iEEG, SOZ localization and subsequent surgical treatment only leads to seizure freedom in 47- 68% of patients, which can be partially attributed to SOZ mis-localization. Thus, there is a need for more accurate noninvasive SOZ localization methods to guide iEEG implantation and surgical treatment for improvement of seizure outcomes after surgical intervention. Resting-state, interictal (between seizure) functional magnetic resonance imaging (fMRI) noninvasively measures fluctuations in blood oxygenation, an indirect measure of brain activity. In this proposal, we aim to develop two novel interictal fMRI SOZ localization methods that improve upon current noninvasive methods by leveraging interictal physiological abnormalities of the SOZ. We will develop these methods in patients with temporal lobe epilepsy as a well-characterized model of focal epilepsy. Interictal epileptic spikes, large amplitude electrical events, are detected clinically with noninvasive scalp EEG to localize the SOZ. However, clinical interictal spike localization methods are limited by low sensitivity to detect spikes and either poor spatial resolution or specialized hardware and extensive preprocessing. There is evidence that interictal epileptic spikes induce specific dynamic fMRI connectivity patterns, therefore, here we propose to develop a method to detect and localize interictal epileptic spikes with dynamic fMRI connectivity (Aim 1). We hypothesize that interictal epileptic spikes identified on scalp EEG induce dynamic fMRI connectivity patterns that could be used to detect and localize spikes with fMRI alone. Fluorodeoxyglucose (FDG) positron emission tomography (PET) is used clinically to image the glucose hypometabolism of epileptic regions, however, this method has limited utility for SOZ localization due to the moderate specificity of hypometabolism to the SOZ. We aim to improve the specificity of FDG-PET SOZ localization by combining it with fMRI (Aim 2). We hypothesize that the SOZ has atypical physiological uncoupling of interictal metabolic (FDG-PET) and hemodynamic (fMRI) activity that could be used for localization. If successful, these studies will provide more accurate and specific noninvasive SOZ localization methods that could be integrated into presurgical noninvasive testing to guide both iEEG implantation and surgical intervention, ultimately improving seizure outco...