Epilepsy is the fourth most common neurological disease globally. In the U.S., 1.2% of the national population reported active epilepsy in 2015. Due to bursts in electrical discharge in the brain, epilepsy patients suffer from evidently random seizures which may cause temporary loss of consciousness, sensation, and control of body movement.
Studies have identified subject-specific circadian and multi-day periodicity in seizure timing by monitoring chronic interictal epileptiform discharges (IED). Also, various types of epilepsy have distinctive timing patterns with respect to sleep. However, researches are inconclusive about how sleep quantitatively affects seizure episodes. More importantly, sleeping is in fact a behavioral state modulated by underlying endogenous circadian rhythms, and we currently lack a pathological understanding of how circadian rhythms and epilepsy are correlated.
We first investigate how circadian rhythms entrain and sync with epilepsy cycles. We hypothesize that instead of timing with respect to sleep, seizure likelihood is modulated by the internal circadian phase of patients. In the propo… View full description
Epilepsy is the fourth most common neurological disease globally. In the U.S., 1.2% of the national population reported active epilepsy in 2015. Due to bursts in electrical discharge in the brain, epilepsy patients suffer from evidently random seizures which may cause temporary loss of consciousness, sensation, and control of body movement.
Studies have identified subject-specific circadian and multi-day periodicity in seizure timing by monitoring chronic interictal epileptiform discharges (IED). Also, various types of epilepsy have distinctive timing patterns with respect to sleep. However, researches are inconclusive about how sleep quantitatively affects seizure episodes. More importantly, sleeping is in fact a behavioral state modulated by underlying endogenous circadian rhythms, and we currently lack a pathological understanding of how circadian rhythms and epilepsy are correlated.
We first investigate how circadian rhythms entrain and sync with epilepsy cycles. We hypothesize that instead of timing with respect to sleep, seizure likelihood is modulated by the internal circadian phase of patients. In the proposed clinical trial, we will control the sleep-wake schedule of patients to be distributed at different circadian phases. Circadian phases will be quantified by passive digital biomarkers from non-invasive wearable sensors. Seizure episodes will be objectively recorded by sub-scalp implants. The hypothesis will be evaluated with conservative statistical tests.
We then focus on ultradian and multi-day personalized probabilistic seizure forecasting in an ambulatory and unobtrusive manner. We will apply machine learning to model the multi-modal sleep and IED time series as a dynamical system. Fitted differential equations may parameterize the periodicity in epilepsy, utilizing circadian rhythms as prognostic biomarkers. Forecast performance will be evaluated by area under the curve, Brier skill score, and Improvement over Chance compared to surrogate time series.