Seizure Detection by Fractal Dimensions Measurement in Long-Term Automated Video-EEG Monitoring: A Preliminary Report

Abstract

ABSTRACT.A simple introduction to the concept of fractal dimensions and their use in EEG long-term video-EEG monitoring is discussed. This method detects primary and secondary generalized seizures and distinguishes them from pseudoseizures and complex partial seizures without secondary generalization and purely artifactual activity. The fractal values of physiological artifacts, e.g., chewing, laughing, are different from those obtained from raw EEG but similar to seizure activity. However, when electrodes became dislodged during a seizure a high fractal dimension in excess of that used to trigger the seizure alarm occurred. Inspection of the video allows ready separation of artifactual activity in a nonseizure situation from that created by generalized seizures.