Abstracts

Rationale: The number and density of electrode contacts needed to adequately sample the epileptic brain is unknown, but is likely far higher than currently-available technology can support. Furthermore, emerging evidence supports the presence of widespread, often well-defined epileptogenic networks that extend far beyond the focal abnormalities seen on non-invasive tests of localization-related epilepsy (LRE). Together, these observations suggest that intracranial EEG (icEEG) recordings that are focused only on areas of structural or metabolic abnormality as commonly designed at most epilepsy centers may under-sample a given patient s epileptogenic network and provide false localization of seizure onset. Accordingly, our practice in the Yale Epilepsy Surgery Program (YESP) has been one of increasingly broad icEEG coverage in the evaluation of most forms of LRE. Here we quantify and evaluate that practice over the past 20 years.Methods: We retrospectively reviewed our prospectively-collected database of patients monitored with icEEG in the YESP since 1990. The number of icEEG contacts placed per patient, results of intracranial monitoring and surgical outcomes were abstracted and analyzed with descriptive statistics.Results: We observed a steady increase in the number of icEEG contacts over the study period, with average numbers of contacts rising from 80-100 per patient in the early 1990s to 200 per patient in the mid-2000s. This trend has continued in recent years as our technical capabilities have increased. Although overall rates of surgical resection and outcomes have remained stable, we have frequently observed network phenomena that would not have been detected by icEEG studies directed only to the areas of abnormality seen on noninvasive testing. A small number of patients developed neurologic symptoms possibly due to electrode mass effect, leading us to adopt a strategy of broad survey icEEG studies followed by more focused pre-resection studies.Conclusions: Intuition that LRE often represents a network process, in which the visible abnormality is only the tip of the iceberg of functionally-connected epileptogenic tissue, led our program to use broader arrays of subdural electrodes over the past two decades. These arrays have in turn produced more evidence of network phenomena, including the characterization of a few well-defined epileptogenic networks. We have been limited by technical constraints, including the number of available recording channels, and the contact density and mass effect of currently-available electrode arrays. These constraints highlight the need for new materials for electrode fabrication, increased transmission capabilities, and new approaches to the temporospatial analysis of large numbers of icEEG channels. We believe that broad icEEG monitoring is needed to fully understand most cases of LRE. Ultimately, understanding epileptogenic networks in patients with LRE may lead to treatment strategies aimed at disconnection or disruption of epileptogenic networks in cases not amenable to focal resection.