Granger causality analysis of intraoperative interictal sEEG data predicts the seizure focus
MetadataShow full item record
The purpose of the study was to investigate whether interictal stereotactic EEG data recorded during surgical implantation of depth electrodes can be used to predict the seizure focus in patients with epilepsy. We retrospectively studied 21 patients who underwent invasive stereotactic EEG monitoring and surgical treatment at Boston Children’s Hospital. Interictal data were analyzed using the method of Granger Causality (GC) statistics to calculate causal interactions between brain regions. In 10 cases, intraoperative GC analysis accurately identified the seizure focus, as supported by their statistically significant rank order sum P values. The remaining 11 cases failed to achieve statistical significance for intraoperative GC analysis. When we examined the visual representations of the causality network described by GC, we observed that despite insignificant rank order P values, GC analysis could still be valuable in revealing the seizure network. It is apparent that there needs to be additional, more in-depth investigations regarding the value of using intraoperative interictal EEG data for defining the seizure focus. Future studies should focus on identifying clinical factors that may affect the utility of intraoperative GC analysis and also on experimenting how GC analysis could aid in decision-making during electrode implantation.