EEG Source Localization: Spike Averaging in Pediatric Patients as a Predictor of the Cortical Epileptic Source

Intractable focal epilepsy is a chronic neurological disorder with recurrent seizures that are not controlled by medication.  Surgical intervention at an identified epileptic focus may control such seizures.  A novel method to determine the epileptic area, source localization, records the brain’s electrical signals from various locations on the scalp to determine the most likely source of epileptic spikes.  The contribution of this technique is limited by a low signal to noise ratio (SNR).  We hypothesized that averaging multiple spikes would increase the SNR and reduce the variation in the solutions typically caused by noise pollution.  Seventeen pediatric patients were in this pilot study of spike averaging and source localization.  At least one hundred spikes with an 80% correlation amplitude and morphology to a prototype spike were automatically detected using Curry software.  The first 25, 50, 75, and 100 detected spikes were averaged; of the 100, three single spikes were randomly chosen to compare results.  The time interval was measured at the tangent half-way of the rising slope of the spike to just past the peak.   The SNR, strength of the dipole, the volume of the ellipsoid of confidence, the variance and residual deviation were measured.   There were no statistical significances between the averaged spike and the single spike analyses.  Averaging spikes failed to significantly decrease the noise contamination in EEG source localization.  Thus, expansion of this study with a larger sample size and restriction to single-source focal epilepsy is required to obtain more consistent source localizations with the resection area.