Synchronization analysis of epileptic ECOG data using SOM-based SI measure

The exact spatio-temporal changes leading to epileptic seizures, although widely studied, are not well understood yet. We propose to investigate the mechanisms leading to epileptic seizures by using a self-organising map (SOM) based similarity index (SI) measure. While it is shown that this measure is statistically as accurate as the original SI measure, it is also computationally faster and therefore applicable for real-time analyses. Application of SOM-based SI measure on epileptic seizure data reveals interesting aspects of synchronization and de-synchronization at various spatio-temporal levels.