Precise localization of epileptogenic zones(EZ)is essential for the successful surgical treatment of medically intractable epilepsy.Despite the wide range of non-invasive investigations,intracranial stereo-EEG(SEEG)recordings represent the gold standard for the EZ identification.In clinical practice,the EZ localization is generally based on visual analysis by epileptologists.However,the rather high rate of failure in epilepsy surgery highlights that the precise identification of the EZ is still an unsolved problem that requires more sophisticated methods of investigation.Recent studies have proposed that epilepsy is a network-level disorder.Therefore we used graph theory,a formalism to quantify topological properties of complex systems,to study the network dynamics and to identify critical network nodes in epileptic networks.SEEG recordings were obtained from four patients undergoing presurgical monitoring for the treatment of intractable epilepsy at Renji Hospital.Partial directed coherence(PDC)analysis was applied to construct functional connectivity networks during ictal events,and then graph theoretical indices including degree and betweenness centrality(BC)were evaluated to find the primary hubs of the ictal epileptic network.We found that the SEEG channels with high degree or high BC values coincided well with the surgical resection areas in three patients who became seizure-free after surgery.In the other patient with unsuccessful surgical outcomes,the SEEG channels with high degree and high BC values located outside the surgical resection areas.These findings suggest the potential application of graph theory in localizing EZ and the presurgical planning of medically intractable epilepsy patients.