HIGH-RESOLUTION ELECTROENCEPHALOGRAPHIC FORWARD MODELING IN TRAUMATIC BRAIN INJURY USING THE FINITE ELEMENT METHOD S.Y. Matthew Goh 1 , Andrei Irimia 1 , Carinna M. Torgerson 1 , Ron Kikinis 2 , Paul M. Vespa 3 , John D. Van Horn 1 1 Laboratory of Neuro Imaging, University of California, Los Angeles 2 Surgical Planning Laboratory, Harvard Medical School, Boston, Massachusetts 3 Brain Injury Research Center, University of California, Los Angeles ABSTRACT Localization of electrical brain activity via electroencephalography (EEG) remains a challenging task in traumatic brain injury (TBI) patients, partly due to the complexity of structural brain changes resulting from neurological insult. When localizing EEG-recorded brain activity, the failure to account for pathology-related changes in tissue conductivities may cause forward model inaccuracies which translate into large localization errors. Here, the effects of TBI-related pathology upon the accuracy of the EEG forward matrix are explored in the context of a realistic finite element method (FEM) model of the head with 25 tissue types. It is found that the omission of TBI pathology from the anatomical model can lead to substantial inaccuracies in the calculation of the forward matrix, with EEG lead field focality being underestimated by as much as ~90% if TBI-related conductivity changes are ignored. Our study is the first to rigorously quantify the extent to which TBI-related pathology can affect forward EEG calculations. Index Terms— traumatic brain injury, epilepsy, electroencephalography, finite element method 1. INTRODUCTION The development of epilepsy following traumatic brain injury (TBI) accounts for 20% of symptomatic epilepsy cases and 5% of all epilepsy cases [1]. Studies in civilian patients and military personnel with post traumatic epilepsy (PTE) have estimated the incidence of seizures at 2-25% and 33%, respectively [2, 3]. Pharmacological intervention remains the preferred treatment solution, although elimination of seizures is successful in only about 35% of PTE cases [4]. In cases of failure, surgical resection of the epileptogenic region is often necessary, a process which requires localization of the epileptic foci through methods such as electroencephalography (EEG). However, the complexity of structural changes as a consequence of TBI complicates the task of source localization which may lead to inaccurate estimations of the sensitivity of EEG sensors. As evident in the case of PTE, the localization of electrical activity via EEG has important applications for the monitoring and therapeutic evaluation of TBI patients. Despite its clinical usefulness, however, the topic of EEG source localization in TBI has been insufficiently explored, likely due to the difficulty of capturing TBI-related structural pathology within EEG forward models. Such pathology may include (1) the absence of skin and skull parts due to open head injuries or craniotomies, and (2) local conductivity profile alterations due to brain hemorrhage and edema. Accurately localizing cortical activity depends on a number of factors, one of which is the realism of the head model used in the forward calculation of electric potentials [5]. Realistic head models account for head anatomy and tissue conductivity, both of which are of particular significance in TBI forward modeling due to the presence of pathology. Studies involving models which account for the presence of lesions and cavities have shown that the latter can have significant qualitative and quantitative effects upon the computed electric potentials [6]. Thus, because EEG source localization accuracy depends highly upon forward model accuracy, the ability to generate EEG forward solutions with high fidelity is crucial to the task of epileptic foci localization. Here, we present a case study in which we explore the effects of including TBI-related changes in tissue conductivities upon the EEG forward model as computed using the finite element method (FEM), which requires the generation of a 3D mesh to account for the geometric and electric properties of the head [5]. In contrast to previous FEM studies which have modeled the head using 10 or fewer tissue types [5], we model the TBI- affected head using 25 tissue types, thereby including the effects of brain hemorrhage and edema upon the forward solution. 2. METHOD 2.1. Participants and image acquisition The representative case selected for presentation is that of a 45-year old male who suffered closed head trauma. The