An approach to identify synchronization clusters within the epileptic network

Introduction: Epileptiform activity is usually assumed to be associated with an increased synchronization of neuronal assemblies involved in the epileptic process. Indeed recent studies indicate that the EEG dynamics of the epileptic focus exhibits a higher degree of synchronization than the dynamics of remote brain areas, even during the interictal state. This would imply the existence of spatially extended synchronization clusters within the epileptic network, the most pronounced cluster describing the spatial-extent of the epileptic focus. We here investigate whether such synchronization clusters can be reliably identified in intracranial EEG recordings in patients suffering from neocortical lesional epilepsies (NLE). For this purpose we utilize our recently proposed approach to detect synchronization clusters in multivariate time series [1].

Methods: We retrospectively analyzed long-term (up to several days) multichannel intracranial EEG data recorded during the interictal state from up to now eight patients who underwent presurgical evaluation. Extended lesionectomy led to post-operative complete seizure control in all cases. We identified synchronization clusters in temporally evolving networks constructed out of the EEG data using a moving window technique. Network links were estimated by a well established bivariate measure of synchronization strength (mean phase coherence) between all pairs of electrodes.

Results: Since we observed multiple synchronization clusters that varied in space and time during the recordings, we averaged the most dominant synchronization clusters over time. These clusters either coinceded with the extent of the epileptic focus as determined by the presurgical workup or reflected spatio-temporal activity of higher association cortices (e.g. Broca's area or motor strip).

Discussion: Our findings indicate that multivariate EEG analysis techniques that capture synchronization dynamics within complex network structures can provide relevant information for interictal localization of the epileptic focus and its delineation from functionally relevant areas. These techniques can thus contribute to a further improvement of the presurgical evaluation of epilepsy patients.

[1] Allefeld C, Bialonski S: Detecting synchronization clusters in multivariate time series via coarse-graining of Markov chains. Phys Rev E 76, 066207, 2007;

This work was supported by the Deutsche Forschungsgemeinschaft (Grant No. SFB-TR3 subproject A2). S.B. was supported by the German National Academic Foundation.