Prediction of epileptic seizures and validation of prediction performances as part of the European "EPILEPSIAE" project

Abstract

Event Abstract Back to Event Prediction of epileptic seizures and validation of prediction performances as part of the European "EPILEPSIAE" project Hinnerk Feldwisch1*, Matthias Ihle2, Michael Jachan1, Jens Timmer1, Andreas Schulze-Bonhage2 and Björn Schelter1 1 University of Freiburg, Germany 2 University Hospital, Germany In recent years, several methods for the prediction of epileptic seizures have been proposed, which are based on recordings of surface and invasive electroencephalogram (EEG). By application of linear or nonlinear measures, pre-seizure changes have been reported. Here, we present a prediction scheme which is based on the mean phase coherence and the dynamical similarity index. Additionally, we compare different approaches for statistical validation of prediction performances. Concerning the development and assessment of the predictability of epileptic seizures, the evaluation of continuous and long-term neurophysiological recordings of epilepsy patients is necessary. The project “EPILEPSIAE” funded by the European Union was set up to compile an extensive database of EEG recordings of 250 epileptic patients in total within the project duration of three years. Based on software infrastructure for both raw data and metadata containing information about the recordings and patients, seizure prediction methods will be developed and assessed. We applied and evaluated two predictive measures to the data of eight patients with invasive EEG recordings. The univariate dynamical similarity index measures the similarity of the dynamics of a reference window of EEG to the dynamics of a sliding window of EEG, whereas the bivariate mean phase coherence is a measure for phase synchronization of two channels of the EEG. For two/one of the patients, significant prediction performances can be observed by the dynamical similarity index/mean phase coherence. As a means to increase seizure prediction performance we tested whether or not two different kinds of combination of seizure prediction methods yield superior prediction results, which is the case for about half of the patients. On average for all patients, prediction sensitivities of 52.8 % can be observed for a maximum false prediction rate of 3.6 false alarms per day. Conclusions: The presented prediction methods are applied to a database of continuous long-term recordings of the invasive EEG of epilepsy patients. For two different measures, significant predictive performances can be observed for a part of the patients, which can be improved by a combination of both measures. Further necessary studies of seizure predictability will be enabled by the currently evolving, extensive database “EPILEPSIAE” of neurophysiological recordings and associated meta-information. Acknowledgements: This work was supported by the European Union (Grant 211713), the German Federal Ministry of Education and Research (BMBF grant 01GQ0420) and the German Science Foundation (Ti 315/2-2; Sonderforschungsbereich-TR3; He1949/1-1). Conference: Neuroinformatics 2008, Stockholm, Sweden, 7 Sep - 9 Sep, 2008. Presentation Type: Poster Presentation Topic: Clinical Neuroinformatics Citation: Feldwisch H, Ihle M, Jachan M, Timmer J, Schulze-Bonhage A and Schelter B (2008). Prediction of epileptic seizures and validation of prediction performances as part of the European "EPILEPSIAE" project. Front. Neuroinform. Conference Abstract: Neuroinformatics 2008. doi: 10.3389/conf.neuro.11.2008.01.113 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 28 Jul 2008; Published Online: 28 Jul 2008. * Correspondence: Hinnerk Feldwisch, University of Freiburg, Freiburg, Germany, hinnerk.feldwisch@bccn.uni-freiburg.de Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Hinnerk Feldwisch Matthias Ihle Michael Jachan Jens Timmer Andreas Schulze-Bonhage Björn Schelter Google Hinnerk Feldwisch Matthias Ihle Michael Jachan Jens Timmer Andreas Schulze-Bonhage Björn Schelter Google Scholar Hinnerk Feldwisch Matthias Ihle Michael Jachan Jens Timmer Andreas Schulze-Bonhage Björn Schelter PubMed Hinnerk Feldwisch Matthias Ihle Michael Jachan Jens Timmer Andreas Schulze-Bonhage Björn Schelter Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.