Dynamic Interplay Between Wake Slow Waves and Epileptiform Discharges in the Epileptogenic Zone
Laurent Sheybani, Umesh Vivekananda, Nishant Sinha, and
5 more authors
Background and Objectives. Outcome of epilepsy surgery remains suboptimal, calling for the identification of new, complementary biomarkers of the epileptogenic zone (EZ). Recently, we identified local wake slow waves (LoWS) as a potential regulator of network excitability that interacts with interictal epileptiform discharges (IEDs). In this study, we tested whether this interaction is associated with surgical outcome. Methods. In this retrospective study, we analyzed intracranial recordings from patients with intractable focal epilepsy who underwent surgery at the Hospital of the University of Pennsylvania. We used surgical success as an indicator that most or all of the EZ had been resected. We used linear mixed models to test whether the incidence of IEDs and LoWS, as well as their interaction, can accurately delineate the EZ in patients with successful vs poor outcome. Results. Across 55 patients (30 women, mean age ±SD: 34 ± 10 years), we found that, although IEDs were more frequent in the seizure-onset zone, their rate in resected and nonresected areas was not associated with surgical success. Indeed, neither the rate of IEDs (F [1, 52.57] = 0.070, p = 0.793) nor that of LoWS (F [1, 48.81] = 1.1032, p = 0.299) in resected vs nonresected areas differed across surgical outcomes. Next, we examined their interaction, validating our previous findings in this larger, independent cohort, by confirming that the closer the LoWS are to an IED, the lower the network excitability during the IED. Furthermore, we found that the delay from IED to the subsequent LoWS—but not the reverse—is associated with surgical outcomes (IED to LoWS: F [1, 52.17] = 5.344, p = 0.025; LoWS to IED: F [1, 52.56] = 1.038, p = 0.313), with shorter delays observed within the EZ. We confirm this using classification analyses that yielded a significant accuracy of 63% (interquartile range: 57%–69%, p < 0.0001), underscoring its potential utility as an additional biomarker of the EZ. Discussion. The temporal proximity of LoWS to a preceding IED in the resected cortex is associated with surgical outcome. This may reflect changes in the regulation of network excitability in the EZ as a form of homeostatic regulation. It raises the possibility to use this index as an additional prognostic biomarker in epilepsy surgery.