INTRODUCTION In partial seizures, from a clinical point of view or even from electroencephalography characterization of post-ictal period can be difficult. The spectral and complex network analysis could lead to a more accurate definition of its limits, as well as to a great understanding of the seizures.

PATIENTS AND METHODS Digital EEG recordings from scalp and foramen oval electrodes were used, 32 seizures, from 15 patients with drug-resistant mesial temporal lobe epilepsy (Engel I). We analyzed numerically: the spectral entropy, the different frequency bands and several variables used to characterize the cortical network, density of links, modularity, cluster coefficient and average path length. Variations of for post-ictal versus pre-ictal periods were quantified.

RESULTS The cortical network density of links increased during the post-ictal period of complex seizures matching with an spectral entropy decrease, mainly due to an increase in Delta band activity. This variables reached extreme values around one minute after seizure end, defined by classical electroencephalography.

CONCLUSIONS Our results can be explained by the appearance of an ‘ending’ mechanism that starts in the ictal period, classically defined, and reach their maximum effect during the post-ictal period. These results could be useful to define the post-ictal period start, as the moment with maximum synchrony, which has a highest density of links and a lowest spectral entropy.