Engineering Nonlinear Epileptic Biomarkers Using Deep Learning and Benford’s Law

In this study we designed two deep neural networks to encode 16 feature latent spaces for early seizure detection in intracranial EEG and compared them to 16 widely used engineered metrics: Epileptogenicity Index (EI), Phase Locked High Gamma (PLHG), Time and Frequency Domain Cho Gaines Distance (TDCG, FDCG), relative band powers, and log absolute band powers (from alpha, beta, theta, delta, gamma, and high gamma bands. The deep learning models were pretrained for seizure identification on the time and frequency domains of one second single channel clips of 127 seizures (from 25 different subjects) using “leave-one-out” (LOO) cross validation. Each neural network extracted unique feature spaces that were used to train a Random Forest Classifier (RFC) for seizure identification and latency tasks. The Gini Importance of each feature was calculated from the pretrained RFC, enabling the most significant features (MSFs) for each task to be identified. The MSFs were extracted from the UPenn and Mayo Clinic's Seizure Detection Challenge to train another RFC for the contest. They obtained an AUC score of 0.93, demonstrating a transferable method to identify interpretable biomarkers for seizure detection.

Risk Factors for Early Seizure After Revascularization in Patients With Moyamoya Disease

Background To investigate the risk factors of early seizure after revascularization in patients with moyamoya disease (MMD). Methods A total of 298 patients with MMD diagnosed in our hospital from 2015 to 2018 were analyzed retrospectively. We summarized the characteristics of seizure after revascularization in patients with MMD and analyzed the predictors of early postoperative seizure. Results We identified 15 patients with MMD who developed seizures within 1 week after revascularization. According to logistic regression analysis, age (OR:1.04, 95% CI 0.998–1.086; P = 0.060), and infarct side (OR:1.92, 95% CI 0.856–4.290; P = 0.113) were not significantly associated with incident early seizure. Postoperative infarction (OR:12.89, 95% CI 4.198–39.525; P = 0.000) and preoperative cerebral infarction (OR:4.08, 95% CI 1.267–13.119; P = 0.018) were confirmed as risk factors for early seizure. Conclusions We believe that history of preoperative infarction and new infarction are independent risk factors of early seizure in patients with MMD after revascularization.

Discontinuation of preventive antiepileptic drugs in patients with intracerebral hemorrhage

Background The risk factors for seizures in patients with intracerebral hemorrhage (ICH) stroke and the effect of seizure prevention by anticonvulsant are not well understood. Limited studies have investigated the risk of seizure after discontinuing antiepileptic drugs in patients with ICH. This study aimed to investigate the role of valproic acid (VA) for seizure prevention and to access the risk of seizure after anticonvulsant withdrawal in patients with spontaneous ICH. Methods Between 2013 and 2015, 177 patients with ICH were enrolled in this 3-year retrospective study. Seizures were classified as early seizure (first seizure within 1 week of ICH), delayed seizure (first seizure after 1 week), and late seizure (any seizure after 1 week). Binary logistic regression was used to evaluate the relationship between baseline clinical factors and late seizures between study periods. VA was prescribed or discontinued based on the decision of the physician in charge. Results Seizures occurred in 24 patients, including early seizure in 6.78% (12/177) of the patients, delayed seizure in 7.27% (12/165) of the patients without early seizure, and late seizure in 9.60% (17/177) of the patients. Most seizures occurred within the first year. Binary logistic regression analysis showed ICH with cortex involvement as the independent risk factor for seizures. VA did not decrease the risk of seizures. Patients with ICH with cortical involvement using anticonvulsants for longer than 3 months did not have a decreased risk of seizures (odds ratio 1.86, 95% CI: 0.43–8.05). Conclusions Spontaneous ICH with cortex involvement is the risk factor for seizure. Most seizures occurred within 1 year after stroke onset over a 3-year follow up. Discontinuation of antiepileptic drug within 3 months in patients does not increase the risk of seizure.

Dynamic Training of a Novelty Classifier Algorithm for Real-Time Early Seizure Onset Detection

The objective of this study was to develop an adaptive framework for seizure detection in real-time that is practical to use in the Epilepsy Monitoring Unit (EMU) as a warning signal, and whose output helps characterize epileptiform activity. Our framework uses a one-class Support Vector Machine (SVM) that is being trained dynamically according to past activity in all available channels. This is done to evaluate the novelty of the current instance according to previous activity. Our algorithm was tested on intracranial EEG from human epilepsy patients that are admitted to the EMU for presurgical evaluation. In this study, we compared multiple configurations for using a one-class SVM to assess if there is significance over specific neural features or electrode locations. Our results show our algorithm is capable of running in real-time and achieving a high performance for early seizure-onset detection with a low false-positive rate and robustness to different types of seizure-onset patterns as well as to the number of channels used. This algorithm offers a solution to warning systems in the EMU as well as a tool for seizure characterization during post-hoc analysis of intracranial EEG data for surgical resection of the epileptogenic network.HighlightsThis study proposes a dynamic training algorithm that efficiently detects sudden novel changes in intracranial electroencephalographic activity, creating a reliable seizure onset detection algorithm that does not need prior training.The algorithm described has the capability to be implemented in real-time, independently of the number of channels that are being analyzed.The presented detector shows high performance and reliability to be easily implemented in the Epilepsy Monitoring Unit to quickly alert clinical staff of seizure events.

Discontinuation of Prophylactic Antiepileptic Drugs in Patients with Intracerebral Hemorrhage

Background: The risk factors for seizures in patients with intracerebral hemorrhage (ICH) stroke and the effect of prophylactic anticonvulsant are not well understood. Limited studies investigated the risk of seizure after discontinuing prophylactic antiepileptic drugs in patients with ICH. This study aimed to investigate the role of valproic acid (VA) for seizure prevention and to access the risk of seizure after anticonvulsant withdrawal in patients with spontaneous ICH.Methods: Between 2013 and 2015, 177 patients with ICH were enrolled in this 3-year retrospective study. Seizure was classified as early seizure (first seizure within 1 week of ICH), delayed seizure (first seizure after 1 week), and late seizure (any seizure after 1 week). Binary logistic regression was used to evaluate the relationship between baseline clinical factors and late seizure between study VA was prescribed or discontinued based on the decision of the physician in charge.Results: Seizures occurred in 24 patients, including early seizure in 6.78% (12/177) of the patients, delayed seizure in 7.27% (12/165) of the patients without early seizure, and late seizure in 9.60% (17/177) of the patients. Most seizures occurred within the first year. Binary logistic regression analysis showed ICH with cortex involvement as the independent risk factor for seizures. VA did not decrease the risk of seizure. Patients with ICH with cortical involvement using prophylactic anticonvulsant for longer than 3 months did not decrease risk of seizure (Odds ratio 1.86, 95% CI: 0.43-8.05).Conclusion: Spontaneous ICH with cortex involvement is the risk factor for seizure. Most seizures occurred within 1 year after stroke onset over a 3-year follow up. Discontinuation Prophylactic antiepileptic drug within 3 months in patients do not increase risk of seizure. VA cannot prevent seizure in patients with ICH.

Predictors of long length of hospital stay among elders admitted with seizures in a tertiary centre: a prospective study

ABSTRACT Population ageing is a global phenomenon, and life expectancy in Brazil is growing fast. Epilepsy is the third most important chronic neurological disorder, and its incidence is higher among elderly patients than in any other segment of the population. The prevalence of epilepsy is greater among inpatients than in the general population and it is related to long length of hospital stay (LOS), which is associated with hospital mortality and higher healthcare costs. Despite these facts, reports of elderly inpatients admitted with seizures and associated outcomes are scarce. Objective: To identify predictors of long LOS among elderly inpatients admitted with seizures. Methods: We prospectively enrolled elders admitted with epileptic seizures or who experienced seizures throughout hospitalization between November 2015 and August 2019. We analysed demographic data, neurological disorders, clinical comorbidities, and seizure features to identify risk factors. Results: The median LOS was 11 days, with an interquartile range (IQR) of 5-21 days. The frequency of long LOS (defined as a period of hospitalization ≥12 days) was 47%. Multivariate analysis showed there was an exponential increase in long LOS if a patient showed any of the following conditions: intensive care unit (ICU) admission (OR=4.562), urinary tract infection (OR=3.402), movement disorder (OR=5.656), early seizure recurrence (OR=2.090), and sepsis (OR=4.014). Conclusion: Long LOS was common among elderly patients admitted with seizures, and most predictors of long LOS found in this cohort might be avoidable; these findings should be confirmed with further research.