System-Identification-Based Automatic Brain Tissue Classification for Stereoelectroencephalography

Abstract

In the cases of drug-resistant epilepsy, patients might undergo resective surgery of the epileptic zone (EZ). The success of the surgery depends on the correct identification of the EZ and the eloquent cortex to be avoided. In both cases, the correct classification of the tissue where the measuring contacts are inserted is needed during the stereoelectroencephalography (SEEG). Most of the tissue classification procedures rely on imaging. In this paper a system identification based automatic classifier is proposed using previously proposed non-parametric and parametric methods for single contact tissue classification. By combining both identification methods, poorly classified contacts can be eliminated, and overall contact classification can be improved, especially for the parametric classifier. The proposed method can be either used in combination with imaging methods, or it could be used to help select contacts to be recorded during SEEG Investigation.