Platform Session – Electroencephalography/Epilepsy: Onset of interictal ripple high-frequency oscillations in intracranial electroencephalography predicts outcome in children undergoing epilepsy surgery

Introduction Approximately 20–30% of children with epilepsy develop medically refractory epilepsy (MRE), and require surgical resection of the epileptogenic zone for seizure-freedom. High-frequency oscillations (HFOs), recorded with intracranial electroencephalography (iEEG) and separated into ripples (>80 Hz) and fast-ripples (>250 Hz) are new promising interictal biomarkers of the epileptogenic zone. Yet, ripple HFOs (80–250 Hz) are seen in large brain areas whose resection may be unnecessary for seizure-freedom, limiting their utility as biomarkers for surgical planning. The goal of this study was to: (i) assess the spatiotemporal propagation of ripples on iEEG, (ii) identify the ripples that initiate interictal propagation ( onset-ripples ), and (iii) evaluate the clinical value of onset-ripples as new epilepsy biomarkers that estimate the epileptogenic zone better than the entirety of ripples observed on iEEG. Methods We analyzed interictal iEEG data from children with MRE who underwent surgical resection. Postsurgical outcome was evaluated with Engel scale at more than 12 months from surgery, and was classified as good (Engel class = 1) or poor (Engel class ⩾ 2) outcome. We developed an automated method to detect ripples and to evaluate their spatiotemporal propagation across iEEG contacts. We identified sequences of ripples propagating across neighboring contacts, and determined latency and distance of each ripple from the propagation onset. We defined three ripple categories based on their spatiotemporal properties, i.e. onset-ripples , spread-ripples , and isolated-ripples , and correlated their mean rates inside and outside resection with postsurgical outcome. We determined the areas that most frequently generated these ripples as onset-ripple-zone , spread-ripple-zone and isolated-ripple-zone , and evaluated the positive and negative predictive value (PPV and NPV) of their resection. Results Twenty-seven patients were included in the study (16 had good postsurgical outcome). We observed propagation sequences of ripples in all patients. The mean rate of onset-ripples inside resection correlated with Engel class (Spearman rho = −0.55, p  = 0.003) and predicted good vs. poor outcome (odds ratio = 5.37, confidence interval = 1.34–21.56, p  = 0.02). Resection of onset-ripple-zone was associated with good-outcome (p = 0.047) with a PPV of 80% and NPV of 60%. No association was found for spread-ripple-zone or isolated-ripple-zone . Conclusion This study presented for the first time propagation of interictal ripples across multiple iEEG contacts in children with MRE. Resecting the areas that initiate ripple propagation predicts good postsurgical outcome. Onset-ripples are potential biomarkers of epileptogenicity and may help surgical treatment of patients with MRE.