RESUMO
PURPOSE: To develop a fast and robust seizure anticipation program and test its usefulness in seizure prediction and localization of epileptic focus. METHODS: The video-EEG monitoring was performed in 20 epilepsy patients. The NPISI (Nonlinear Periodicity of Inter-Spike Intervals) of each channel and averaged NPISI of all channels were calculated in pre-ictal and ictal periods of 21 epileptic seizures. Seizure prediction times of NPISI in each channel and all channels were determined. RESULTS: There were eight mesial TLE, twelve neocortical epilepsies. The significantly reduced NPISI prior to seizure onset was observed in 20 seizures (95.2%) by NPISI of single EEG channel and in 14 seizures (66.7%) by the averaged NPISI of all channels. Time of seizure anticipation was 17.3+/-7.3 min by single channel NPISI and 14.0+/-6.0 min by all channel NPISI prior to the seizure onset. The earliest NPISI reduction was detected at electrodes concordant to the seizure focus in 6/20 (30.0%) seizures, electrodes adjacent to the focus in 5/20 (25.5%), mirror electrodes contralateral to the seizure focus in 4/20 (20.0%), and irrelevant electrodes to the focus in 5/20 (25.5%). CONCLUSIONS: This study demonstrated that the reduction of NPISI could predict epileptic seizures long before clinical seizure onset. Although the earliest NPISI reduction was observed at or near to epileptic focus in most cases, pre-ictal changes of NPISI occurred far beyond the epileptic focus. These findings suggest that pre-ictal NPISI reduction may reflect a widespread facilitation of brain before the seizure onset.