Abstract
Childhood epilepsy with centrotemporal spikes, previously known as Benign Epilepsy with Centro-temporal Spikes (BECTS) or Rolandic Epilepsy, is one of the most common forms of focal childhood epilepsy. Despite its prevalence, BECTS is often misdiagnosed or missed entirely. This is in part due to the nocturnal and brief nature of the seizures, making it difficult to identify during a routine electroencephalogram (EEG). Detecting brain activity that is highly associated with BECTS on a brief, awake EEG has the potential to improve diagnostic screening for BECTS and predict clinical outcomes. For this study, 31 patients with BECTS were retrospectively selected from the BCH Epilepsy Center database along with a contrast group of 31 patients in the database who had no form of epilepsy and a normal EEG based on a clinical chart review. Nonlinear features, including multiscale entropy and recurrence quantitative analysis, were computed from 30-second segments of awake EEG signals. Differences were found between these multiscale nonlinear measures in the two groups at all sensor locations, while visual EEG inspection by a board-certified child neurologist did not reveal any distinguishing features. Moreover, a quantitative difference in the nonlinear measures (sample entropy, trapping time and the Lyapunov exponents) was found in the centrotemporal region of the brain, the area associated with a greater tendency to have unprovoked seizures, versus the rest of the brain in the BECTS patients. This difference was not present in the contrast group. As a result, the epileptic zone in the BECTS patients appears to exhibit lower complexity, and these nonlinear measures may potentially serve as a clinical screening tool for BECTS, if replicated in a larger study population.
Highlights
Seizures associated with Benign Epilepsy with Centrotemporal Spikes (BECTS) can go unrecognized due to their nocturnal, brief and sometimes subtle nature
We found that there was a difference between several multiscale nonlinear measures of awake BECTS patients and awake patients without BECTS in our contrast group, while visual EEG inspection did not reveal any distinguishing features
Several nonlinear measures derived from EEGs are clearly different in the centrotemporal region of BECTS patients when compared to the same region in patient without BECTS
Summary
Seizures associated with BECTS can go unrecognized due to their nocturnal, brief and sometimes subtle nature. Various measures of nonlinear dynamics have been computed from EEG time series in order to detect changes immediately prior to the onset of seizures or epileptiform discharges[18,19]. Machine learning was applied to nonlinear signal features derived from EEG measurements taken as early as three months of age to predict infants who later developed autism spectrum disorder from those that did not[13]. Recurrence quantitative analysis has been used for early seizure detection by distinguishing ictal and interictal entropy states and recently for differentiating children with autism spectrum disorder from typically developing children[29]. Our work builds on these results by developing a general approach to computing nonlinear dynamical features from EEG signals from a clinically well characterized epilepsy patient population, and evaluate these results to determine sensor location and measures that are distinctly different in BECTS patients even while awake
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call