Search

Objective. In this work, we propose a new computational framework to assist neurophysiologists in Stereoelectroencephalography (SEEG) analysis, with the final aim of improving the definition of the Epileptogenic Zone (EZ) in patients with drug-resistant epilepsy. Method. We consider the Phase Transfer Entropy (PTE) model to estimate the effective connectivity between SEEG channels. Hence, we design an algorithm, named the Desynchronization Index (DI), that classifies as epileptogenic the channels that show independent behavior with respect to the rest of the network during the seconds preceding the seizure propagation. Results. We test the proposed DI algorithm against Epileptogenic Index (EI) on a clinical dataset of 11 patients, considering the neurophysiological evaluation of the EZ as the clinical ground truth. Our results denote that DI overcomes EI in terms of area under the ROC curve (AUC=0.80 vs AUC=0.74), while combining the two algorithms as a unique tool leads to the best performance (AUC=0.87). Conclusions. The DI algorithm underscores connectivity dynamics that can hardly be identified with a pure visual analysis, increasing the accuracy in the EZ definition compared to traditional methods. Significance. The integration of connectivity- and energy-based metrics, as the one we propose, can lead to the definition of a new effective biomarker of the EZ, reducing the burden required by the SEEG review in the case of extensive implants and improving our understanding of the dynamics leading to the generation of seizures.