Interictal high frequency oscillations as non-invasive biomarkers of epileptogenicity in pediatric patients

点击翻译按钮获取中文摘要

Project Summary Crucial to the success of epilepsy surgery is the availability of a robust presurgical biomarker to identify the epileptogenic zone (EZ). Complete resection of the EZ may lead to medication and seizure freedom. Since the EZ cannot be measured directly, its location is estimated indirectly based on concordant data from a multitude of noninvasive tests. Yet, the results of these tests are often insufficiently concordant or inconclusive. Intracranial electroencephalography (iEEG) serves as the gold standard for the delineation of the seizure onset zone (SOZ). However, the SOZ does not always predict the surgical outcome, and its delineation requires many days of recordings to capture clinical seizures. High-frequency oscillations (HFOs), recorded with iEEG are promising interictal biomarkers of the EZ. Yet, their clinical value for epilepsy surgery is still debated since the HFO-generating area is often relatively large and its complete resection may overlap with eloquent areas. This is often attributed to the presence of physiological HFOs in non-epileptogenic areas. In our recent iEEG study, we showed that interictal HFOs are initiated by an onset generator and spread to other brain areas over time. This generator constitutes a promising interictal biomarker of the EZ since its resection is associated with good surgical outcome. Despite the copious literature on HFOs, the clinical value of HFOs for surgery has been only investigated using iEEG, which presents serious limitations due to its invasiveness and its limited spatial sampling. This application aims to noninvasively localize interictal HFOs with high-density electroencephalography (HD-EEG) and magnetoencephalography (MEG) in children with medically refractory epilepsy (MRE), distinguish pathological from physiological HFOs, and assess the noninvasive localization of the HFO-onset generator with respect to the surgical resection and patients’ outcome. Our hypothesis is that HD-EEG and MEG can distinguish pathological from physiological HFOs non-invasively and can localize the HFO-onset generator whose removal leads to better surgical outcome than the removal of the area of secondary spread. To test our hypothesis, we specifically aim to: (i) assess the ability of HD-EEG and MEG to localize HFOs; (ii) differentiate physiological (nHFOs) from pathological (pHFOs) HFOs using unsupervised machine learning; (iii) localize noninvasively the HFO-onset generator and compare it with the clinical gold standard for resection tailoring, i.e. the iEEG-defined SOZ; and (iv) assess the predictive value of the HFO- onset generator in terms of surgical outcome. To pursue these aims, we will record HD-EEG and MEG data from 50 children (0-18 years old) with MRE and 50 typically developing (TD) children. This application combines the use of cutting-edge pediatric neuroimaging instruments, 3D printing technology, and innovative signal processing tools together with extensive neuroimaging experience with children. Our research will have a direct impact on the life of children with MRE since it will provide a new noninvasive biomarker of the EZ that would limit invasive long-term monitoring, augment presurgical planning, and improve the surgical outcome.