Wearable technology will become available and allow prolonged electroencephalography (EEG) monitoring in the home environment of patients with epilepsy. Neurologists analyse the EEG visually and annotate all seizures, which patients often under report. Visual analysis of a 24 hour EEG recording typically takes one to two hours. Reliable automated seizure detection algorithms will be crucial to reduce this analysis. We study a dataset of behind-the-ear EEG measurements. Our first aim was to develop a methodology to reduce the EEG dataset by classifying part of the data automatically, while retaining 100% detection sensitivity (DS). Prediction confidences are determined by temperature scaling of the classification model outputs and trust scores. A DS of approximately 90% (99%) can be achieved when automatically classifying around 90% (60%) of the data. Perfect DS can be achieved when automatically classifying 50% of the data. Our second contribution demonstrates that a common modelling strategy, where predictions from several short EEG segments are used to obtain a final prediction, can be improved by filtering out untrustworthy segments with low trust scores. The false detection rate shows a relative decrease between 21% and 43%, and the DS shows a small increase or decrease.

We describe a novel system for online prediction of lead seizures from long-term intracranial electroencephalogram (iEEG) recordings for canines with naturally occurring epilepsy. This study adopts new specification of lead seizures, reflecting strong clustering of seizures in observed data. This clustering results in fewer lead seizures (~7 lead seizures per dog), and hence new challenges for online seizure prediction, that are addressed in the proposed system. In particular, machine learning part of the system is implemented using the Group Learning method suitable for modeling sparse and noisy seizure data. In addition, several modifications for the proposed system are introduced to cope with non-stationarity of noisy iEEG signal. They include: (1) periodic re-training of SVM classifier using most recent training data; (2) removing samples with noisy labels from training data; (3) introducing new adaptive post-processing technique for combining many predictions made for 20-second windows into a single prediction for 4 hr segment. Application of the proposed system requires only 2 lead seizures for training the initial model, and results in high prediction performance for all four dogs (with mean 0.84 sensitivity, 0.27 time-in-warning, and 0.78 false-positive rate per day). Proposed system achieves accurate prediction of lead seizures during long-term test periods, 3–16 lead seizures during 169–364 days test period, whereas earlier studies did not differentiate between lead vs. non-lead seizures and used much shorter test periods (~few days long).

The modulation of Nrf2 activity has been reported to be implicated in the pathology of various neurological disorders, including epilepsy. Previous studies have demonstrated that Nrf2 is activated in the post-status epilepticus rat model, however, the spatio-temporal, as well as cell type-specific expression of Nrf2 following brief epileptic seizures remains unclear. Here, we evaluated how an acute epileptic seizure affected the expression of Nrf2 and its downstream genes in the cortex and the hippocampus up to 1-week following the induced seizure. We found that after a pentylenetetrazol-induced seizure, Nrf2 significantly increased at 24 h at the mRNA level and 3 to 6 h at the protein level in the cortex. In the hippocampus, the Nrf2 mRNA level peaked at 3 h after the seizure, and no significant changes were observed in the protein level. Interestingly, the mRNA level of Nrf2 downstream genes peaked at 3-6 h after seizure in both the cortex and the hippocampus. A significant increase in the expression of Nrf2 was observed in the neuronal population of CA1 and CA3 regions of the hippocampus, as well as in the cortex. Moreover, we observed no change in the co-localization of Nrf2 with astrocytes neither in the cortex nor in CA1 and CA3. Our results revealed that following a brief acute epileptic seizure, the expression of Nrf2 and its downstream genes is transiently increased and peaked at early timepoints after seizure predominantly in the hippocampus, and this expression is restricted to the neuronal population.

Key features of glycogen metabolism in excitable tissues are not well-explained by current concepts. Glycogen stores in brain and skeletal muscle are generally considered to function as local glucose reserves, to be utilized during transient mismatches between glucose supply and demand; however, quantitative measures show that blood glucose supply is likely never rate-limiting for energy metabolism in either brain or muscle under physiological conditions. These tissues nevertheless do normally utilize glycogen during intervals of increased energy demand, despite the availability of free glucose, and despite the ATP cost of cycling glucose through glycogen polymer. This seemingly wasteful shunt can be explained by considering the effect of glycogenolysis on the amount of energy derived from ATP (ΔG’ATP). ΔG’ATP is diminished by elevations in Pi, such as occur at sites of rapid ATP hydrolysis and net phosphocreatine consumption. Glycogen utilization counters this effect by sequestering Pi in glycolytic metabolites (glycogen_n + Pi → glycogen_n-1 + glucose-1-phosphate → phosphorylated glycolytic intermediates), and thereby maintains the amount of energy obtained from ATP at sites of rapid ATP consumption. This thermodynamic effect may be particularly important in the narrow, spatially constricted astrocyte processes that ensheath neuronal synapses. This effect can also explain the co-localization of glycogen and cytosolic phosphocreatine in brain astrocytes, glycolytic super-compensation in brain when glycogen is not available, and aspects of exercise physiology in muscle glycogen phosphorylase deficiency (McArdle’s disease).

Glucose is the primary energy fuel used by the brain and is transported across the blood-brain barrier (BBB) by the glucose transporter type 1 and 2.[1] A GLUT1 genetic defect is responsible for glucose transporter type 1 deficiency syndrome (GLUT1DS). Patients with GLUT1DS may present with pharmaco-resistant epilepsy, developmental delay, microcephaly, and/or abnormal movements, with tremendous phenotypic variability. Diagnosis is made by the presence of specific clinical features, hypoglycorrhachia and an SLC2A1 gene mutation. Treatment with a ketogenic diet therapy (KDT) is the standard of care as it results in production of ketone bodies which can readily cross the BBB and provide an alternate energy source to the brain in the absence of glucose. KDTs have been shown to reduce seizures and abnormal movements in children diagnosed with GLUT1DS. However, little is known about the impact of KDT on cognitive function, seizures and movement disorders in adults newly diagnosed with GLUT1DS and started on a KDT in adulthood, or the appropriate ketogenic diet therapy to administer. This case report demonstrates the potential benefits of using a modified Atkins diet (MAD), a less restrictive ketogenic diet therapy on cognition, seizure control and motor function in an adult with newly-diagnosed GLUT1SD.