Objective: This work investigates the hypothesis that focal seizures can be predicted using scalp electroencephalogram (EEG) data. Our first aim is to learn features that distinguish between the interictal and preictal regions. The second aim is to define a prediction horizon in which the prediction is as accurate and as early as possible, clearly two competing objectives. Methods: Convolutional filters on the wavelet transformation of the EEG signal are used to define and learn quantitative signatures for each period: interictal, preictal, and ictal. The optimal seizure prediction horizon is also learned from the data as opposed to making an a priori assumption. Results: Computational solutions to the optimization problem indicate a ten-minute seizure prediction horizon. This result is verified by measuring Kullback-Leibler divergence on the distributions of the automatically extracted features. Conclusion: The results on the EEG database of 204 recordings demonstrate that (i) the preictal phase transition occurs approximately ten minutes before seizure onset, and (ii) the prediction results on the test set are promising, with a sensitivity of 87.8% and a low false prediction rate of 0.142 FP/h. Our results significantly outperform a random predictor and other seizure prediction algorithms. Significance: We demonstrate that a robust set of features can be learned from scalp EEG that characterize the preictal state of focal seizures.

ObjectiveTo compare by 7 Tesla (7T) magnetic resonance imaging (MRI) in patients with focal epilepsy who have non-lesional clinical MRI scans with healthy controls.Methods37 patients with focal epilepsy, based on clinical and electroencephalogram (EEG) data, with non-lesional MRIs at clinical field strengths and 21 healthy controls were recruited for the 7T imaging study. The MRI protocol consisted of high resolution T1-weighted, T2-weighted and susceptibility weighted imaging sequences of the entire cortex. The images were read by two neuroradiologists, who were initially blind to clinical data, and then reviewed a second time with knowledge of the seizure onset zone.ResultsA total of 25 patients had findings with epileptogenic potential. In five patients these were definitely related to their epilepsy, confirmed through surgical intervention, in three they co-localized to the suspected seizure onset zone and likely caused the seizures. In seven patients the imaging findings co-localized to the suspected seizure onset zone but were not the definitive cause, and ten had cortical lesions with epileptogenic potential that did not localize to the suspected seizure onset zone. There were multiple other findings of uncertain significance found in both epilepsy patients and healthy controls. The susceptibility weighted imaging sequence was instrumental in guiding more targeted inspection of the other structural images and aiding in the identification of cortical lesions.SignificanceInformation revealed by the improved resolution and enhanced contrast provided by 7T imaging is valuable in noninvasive identification of lesions in epilepsy patients who are non-lesional at clinical field strengths.

These findings suggest that epilepsy may be associated with significantly asymmetric distribution of PVSs in the brain. Furthermore, the region of maximal asymmetry of the PVSs may help provide localization or confirmation of the seizure onset zone.

Objective Surgical resection of seizure‐producing brain tissue is a gold standard treatment for drug‐resistant focal epilepsy. However, several patient‐specific factors can preclude resective surgery, including a spatially extensive (“regional”) seizure‐onset zone (SOZ). For such patients, responsive neurostimulation (RNS) represents a potential treatment, but its efficacy has not been investigated in this population. Methods We performed a multicenter retrospective cohort study of patients (N = 30) with drug‐resistant focal epilepsy and a regional neocortical SOZ delineated by intracranial monitoring who were treated with the RNS System for at least 6 months. RNS System leads were placed at least 1‐cm apart over the SOZ, and most patients were treated with a lead‐to‐lead stimulation pathway. Five patients underwent partial resection of the SOZ concurrent with RNS System implantation. We assessed change in seizure frequency relative to preimplant baseline and evaluated correlation between clinical outcome and stimulation parameters. Results Median follow‐up duration was 21.5 months (range 6‐52). Median reduction in clinical seizure frequency was 75.5% (interquartile range [IQR] 40%‐93.9%). There was no significant difference in outcome between patients treated with and without concurrent partial resection. Most patients were treated with low charge densities (1‐2.5 µC/cm2), but charge density, interlead distance, and duration of treatment were not significantly correlated with outcome. Significance RNS is a feasible and effective treatment in patients with drug‐resistant regional neocortical seizures. Prospective studies in larger cohorts are necessary to determine optimal lead configuration and stimulation parameters, although our results suggest that lead‐to‐lead stimulation and low charge density may be effective in some patients.

The RNS System is not approved in patients under 18, although a critical need for novel treatment modalities in this vulnerable population persist. We present two pediatric patients with drug-resistant epilepsy secondary to Lennox-Gastaut Syndrome (LGS) and autism spectrum disorder (ASD) treated with the RNS System. Both patients have experienced 75-99% clinical seizure reductions in >1 year of follow-up. We illustrate that children with diffuse onset, multifocal epilepsy, including frontal and thalamic circuits thought to exist in the generation of LGS seizures, can be treated with responsive neurostimulation safely and effectively, targeting thalamic networks, and avoiding palliative disconnections and resections.