Abstracts

Rationale: Interictal high frequency oscillations (HFOs) have emerged as promising biomarkers of the epileptogenic zone (EZ) in patients with medically refractory epilepsy (MRE). HFOs are distinguished in ripples (>80 Hz) and fast ripples (> 250 Hz). Fast ripples are considered more pathologic than ripples, but they are hard to be detected in many patients or have low occurrence rates. Ripples are observed more frequently, but are less specific to the EZ, since the zone that generates ripples can be widespread and not entirely epileptogenic. In this study, we hypothesize that: (i) the large extent of the area that generates interictal ripples (entire ripple-zone) is the result of propagation from an epileptogenic onset (ripple onset-zone) to other areas with less epileptogenic potential; and (ii) the ripple onset-zone is a more precise estimator of the EZ compared to the entire ripple-zone. Our primary goal is to delineate the interictal ripple onset-zone and correlate it with the seizure onset zone (SOZ), the surgical resection, and the patient’s surgical outcome. Methods: We retrospectively analyzed intracranial electroencephalography (iEEG) data from children with MRE who underwent extended extraoperative iEEG. Ripples (80-250 Hz) were detected using our automated algorithm developed and validated by our group. The entire ripple-zone was defined as the area covered by iEEG electrodes with high rate of ripples. Then, we identified ripples propagating across different electrodes and mapped their propagation on the patient’s cortical surface (Fig. 1). We identified the onset of each propagation, calculated the onset rate per electrode, and defined the ripple onset-zone. We also estimated the mean ripple latency with respect to the onset for each electrode. Finally, we calculated sensitivity, specificity and accuracy of the ripple onset-zone to the clinically evaluated SOZ, and compare them with the ones of the entire ripple-zone. The resected electrodes were identified from surgeon’s notes and post-operative MRI if these were available, and the mean ripple latency of resected and non-resected areas were calculated for each patient. Results: We analyzed 1,818 iEEG electrodes (116 min of interictal activity) from 18 children with MRE (age: 12.3 ± 5.7 yrs). We detected 51,575 ripples grouped in 6,123 propagation events. We observed that interictal ripples occurred in multiple iEEG electrodes with different discernible latency (Fig. 1). The mean onset rate was significantly higher in the SOZ than outside (n=18, p < 0.001) and the ripple latency was significantly shorter (p=0.02). The ripple onset-zone was more specific and accurate to the SOZ than the entire ripple-zone (p < 0.05). Finally, in seizure-free patients (Engel=1), the resected area showed a shorter latency than the non-resected area (Fig. 2; p < 0.01, n=8); whereas, no differences were observed in patients with worse outcome (Engel>1; n=5, Fig.2). Contrarily, the ripple rate, which do not include information about propagation, did not show differences between resected and non-resected area independently from the seizure outcome. Conclusions: We present for the first time evidence that interictal ripples propagate rapidly across the cortex. Such propagation may reflect underlying epileptic neural networks and may provide critical information about the specific location of the EZ. Our findings suggest that the ripple onset-zone is a more precise estimator of the EZ than the entire ripple-zone. Investigating the propagation of ripples may thus be critical to understand the pathophysiology of epilepsy and improve the presurgical evaluation and surgical outcome of children with MRE. Funding: AES research grant (PI: Christos Papadelis)