Seizure pathways change on circadian and slower timescales in individual patients with focal epilepsy

Significance Epilepsy is an episodic disease characterized by brief periods of abnormal brain activity, known as seizures, that often have clinical correlates. In many patients, seizures preferentially happen during certain stages of daily and multiday cycles. However, it is unclear whether and how seizures themselves change over time, even though such variability may have clinical implications. To address this knowledge gap, we quantitatively analyze the nature of within-patient variability in seizure network evolutions. Contrary to common expectations, we find seizure variability throughout our cohort. Moreover, we demonstrate that seizures do not change randomly; instead, they also appear to fluctuate over daily and slower timescales. Ultimately, we may improve treatments by tailoring interventions to the full repertoire of seizures in each patient.
Personalized medicine requires that treatments adapt to not only the patient but also changing factors within each individual. Although epilepsy is a dynamic disorder characterized by pathological fluctuations in brain state, surprisingly little is known about whether and how seizures vary in the same patient. We quantitatively compared within-patient seizure network evolutions using intracranial electroencephalographic (iEEG) recordings of over 500 seizures from 31 patients with focal epilepsy (mean 16.5 seizures per patient). In all patients, we found variability in seizure paths through the space of possible network dynamics. Seizures with similar pathways tended to occur closer together in time, and a simple model suggested that seizure pathways change on circadian and/or slower timescales in the majority of patients. These temporal relationships occurred independent of whether the patient underwent antiepileptic medication reduction. Our results suggest that various modulatory processes, operating at different timescales, shape within-patient seizure evolutions, leading to variable seizure pathways that may require tailored treatment approaches.