Identifying the seizure onset zone (SOZ) in the epileptic brain is crucial, especially for drug-resistant epilepsy patients who may need to undergo resection surgery for treatment. Interictal epileptiform discharges (IEDs), known as a biomarker for epilepsy, are not always consistent in the recorded intracranial data. Delayed responses (DRs) to single pulse electrical stimulation are believed to be associated with epileptogenic regions. If DRs and IEDs are found to originate from the same brain location, and DRs can be accurately localized, it would mark a significant advancement in identifying SOZ.
This study investigates whether DRs and IEDs share the same origin. It introduces an adaptive Bayesian beamforming pipeline applied to an array of cortical EEG electrodes and utilizes the morphology of DRs and IEDs as a template for the desired source. The adaptive Bayesian beamformer with multiple constraints (ABMC) is robust to the temporal variability of these spike types while being sensitive to the morphology of target activity due to employing the cross-correlation between the desired template and the beamformer output, able to differentiate between highly correlated sources in close proximity to each other, and surpasses conventional and recently proposed beamformers.
The ABMC beamformer was tested on 300 data segments selected from five DRE cases, representing a larger clinical population eligible for surgical treatment. The results strongly suggest that IEDs and DRs associated with seizures are likely to originate from the same location in the brain. This finding significantly aids in the rapid identification of seizure sources, potentially improving surgical outcomes for epilepsy patients. This is a significant step towards improving clinical assessment for DRE patients.