The existing epilepsy seizure detection algorithms have problems such as overfitting and poor generalization ability due to high reliance on manual labeling of electroencephalogram's data and data imbalance between seizure and interictal periods. An unsupervised learning detection method for epileptic seizure that jointed graph attention network (GAT) and Transformer framework (GAT-T) was proposed. In this method, channel correlations were adaptively learned by GAT encoder. Temporal information was captured by one-dimensional convolution decoder. Combining outputs of the two mentioned above, predicted values for electroencephalogram were generated. The collective anomaly score was calculated and the detection threshold was determined. The results demonstrated that GAT-T achieved the average performance exceeding 90% (or 99%) with a 0.25 s (or 2 s) time segment length, which could effectively detect epileptic seizures. Moreover, the channel association probability matrix was expected to assist clinicians in the initial screening of the epileptogenic zone, and ablation experiments also reflected the significance of each module in GAT-T. This study may assist clinicians in making more accurate diagnostic and therapeutic decisions for epilepsy patients.
Humans ; Electroencephalography/methods* ; Epilepsy/physiopathology* ; Algorithms ; Seizures/physiopathology* ; Neural Networks, Computer ; Signal Processing, Computer-Assisted

Current epilepsy prediction methods are not effective in characterizing the multi-domain features of complex long-term electroencephalogram (EEG) data, leading to suboptimal prediction performance. Therefore, this paper proposes a novel multi-scale sparse adaptive convolutional network based on multi-head attention mechanism (MS-SACN-MM) model to effectively characterize the multi-domain features. The model first preprocesses the EEG data, constructs multiple convolutional layers to effectively avoid information overload, and uses a multi-layer perceptron and multi-head attention mechanism to focus the network on critical pre-seizure features. Then, it adopts a focal loss training strategy to alleviate class imbalance and enhance the model's robustness. Experimental results show that on the publicly created dataset (CHB-MIT) by MIT and Boston Children's Hospital, the MS-SACN-MM model achieves a maximum accuracy of 0.999 for seizure prediction 10 ~ 15 minutes in advance. This demonstrates good predictive performance and holds significant importance for early intervention and intelligent clinical management of epilepsy patients.
Humans ; Electroencephalography/methods* ; Epilepsy/physiopathology* ; Neural Networks, Computer ; Seizures/physiopathology* ; Signal Processing, Computer-Assisted ; Algorithms

Epilepsy is a chronic neurological disorder affecting ~65 million individuals worldwide. Abnormal synaptic plasticity is one of the most important pathological features of this condition. We investigated how ubiquitin-specific peptidase 47 (USP47) influences synaptic plasticity and its link to epilepsy. We found that USP47 enhanced excitatory postsynaptic transmission and increased the density of total dendritic spines and the proportion of mature dendritic spines. Furthermore, USP47 inhibited the degradation of the ubiquitinated α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) subunit glutamate receptor 1 (GluR1), which is associated with synaptic plasticity. In addition, elevated levels of USP47 were found in epileptic mice, and USP47 knockdown reduced the frequency and duration of seizure-like events and alleviated epileptic seizures. To summarize, we present a new mechanism whereby USP47 regulates excitatory postsynaptic plasticity through the inhibition of ubiquitinated GluR1 degradation. Modulating USP47 may offer a potential approach for controlling seizures and modifying disease progression in future therapeutic strategies.
Animals ; Receptors, AMPA/metabolism* ; Neuronal Plasticity/physiology* ; Seizures/physiopathology* ; Disease Models, Animal ; Mice, Inbred C57BL ; Mice ; Ubiquitin Thiolesterase/genetics* ; Male ; Excitatory Postsynaptic Potentials/physiology* ; Ubiquitination ; Dendritic Spines/metabolism* ; Hippocampus/metabolism*

Epileptic seizures and the interictal epileptiform discharges both have similar waveforms. And a method to effectively extract features that can be used to distinguish seizures is of crucial importance both in theory and clinical practice. We constructed state transfer networks by using visibility graphlet at multiple sampling intervals and analyzed network features. We found that the characteristics waveforms in ictal periods were more robust with various sampling intervals, and those feature network structures did not change easily in the range of the smaller sampling intervals. Inversely, the feature network structures of interictal epileptiform discharges were stable in range of relatively larger sampling intervals. Furthermore, the feature nodes in networks during ictal periods showed long-term correlation along the process, and played an important role in regulating system behavior. For stereo-electroencephalography at around 500 Hz, the greatest difference between ictal and the interictal epileptiform occurred at the sampling interval around 0.032 s. In conclusion, this study effectively reveals the correlation between the features of pathological changes in brain system and the multiple sampling intervals, which holds potential application value in clinical diagnosis for identifying, classifying, and predicting epilepsy.
Humans ; Electroencephalography/methods* ; Epilepsy/diagnosis* ; Signal Processing, Computer-Assisted ; Brain/physiopathology* ; Seizures/diagnosis* ; Algorithms

OBJECTIVE: To determine the correlation of phosphorylated ribosomal S6 protein (P-S6) content in blood and brain tissue in mice and rats with seizure. METHODS: Seizure models were induced by intraperitoric injection of kainic acid (KA) in C57BL/mice and SD rats. Flow cytometry was used to detect the content of P-S6 in blood; Western blot was used to detect the expression of P-S6 in brain tissues. The correlation between P-S6 expression in blood and in brain tissue was examine by Pearson analysis, and the correlation between P-S6 expression in blood and the severity of seizure was also observed. RESULTS: Western blotting analysis showed that the expression of P-S6 was significantly increased in peripheral blood and brain tissue in mice 1 h after KA-induced seizure,and the expression levels increased to (1.49±0.45) times (<0.05) and (2.55±0.66) times ( <0.01) of the control group, respectively. Flow cytometry showed that the positive percentage and average fluorescence intensity of P-S6 in the blood of mice increased significantly 1 h after KA-induced seizures (<0.01), which was consistent with the expression of P-S6 in brain tissue (=0.8474, <0.01). Flow cytometry showed that the average fluorescence intensity of P-S6 in blood increased from 14.89±9.75 to 52.35±21.72 (<0.01) in rats with seizure, which was consistent with the change of P-S6 in brain tissue (=0.9385, <0.01). Rats with higher levels of seizure were of higher levels of P-S6 in peripheral blood. CONCLUSIONS Consistent correlation of P-S6 expression is demonstrated in peripheral blood and in brain tissue after KA-induced seizure, suggesting that the expression of P-S6 in blood can accurately reflect the changes of mTOR signaling pathway in brain tissue.
Animals ; Brain ; drug effects ; physiopathology ; Gene Expression Regulation ; drug effects ; Kainic Acid ; Mice ; Mice, Inbred C57BL ; Phosphorylation ; Rats ; Rats, Sprague-Dawley ; Seizures ; blood ; chemically induced ; physiopathology

OBJECTIVE: To investigate the clinical significance of high-frequency oscillations (HFOs) on scalp electroencephalography (EEG) in patients with epileptic encephalopathy with continuous spike-and-wave during sleep (CSWS). METHODS: Twenty-one CSWS patients treated for epilepsy from January 2006 to December 2016 in Pediatric Department of Peking University First Hospital were enrolled into the study. Selected clinical variables including gender, age parameters, seizure frequencies and antiepileptic drugs were compared between (a). HFO-positive group and HFO-negative group before methylprednisolone treatment and (b). excellent seizure outcome group and not-excellent seizure outcome group after methylprednisolone treatment. Interictal HFOs and spikes in pre- and post-methylprednisolone scalp EEG were measured and analyzed. RESULTS: Before methylprednisolone treatment, there were 12 of 21 (57%) CSWS patients had HFOs, with a mean value 43.17 per 60 s per patient. The 12 patients with HFOs tended to have more frequent epileptic negative myoclonus/atonic/myoclonus/atypical absences than those without HFOs in a month before methylprednisolone treatment. A total of 518 HFOs and 22 592 spikes were found in the pre-methylprednisolone EEG data of 21 patients, and 441 HFOs (86%) were associated with spikes. The highest amplitudes of HFOs were significantly positively correlated with that of spikes (r=0.279, P<0.001). Rates reduced by methylprednisolone treatment were statistically significant for both HFOs (P=0.002) and spikes (P=0.006). The percentage of reduction was 91% (473/518) and 39% (8 905/22 592) for spikes and HFOs, respectively. The percentage of spike and HFOs changes was respectively 100% decrease and 47% decrease in the excellent seizure outcome group, and they were 79% decrease and 18% increase in the not-excellent seizure outcome group. CONCLUSION Prevalence of HFOs might reflect some aspect of epileptic activity. HFOs were more sensitive to methylprednisolone treatment than spikes and had a good correlation with the prognosis of seizures, and HFOs could be applied to assess epilepsy severity and antiepileptic therapy.
Anticonvulsants/therapeutic use* ; Child ; Electroencephalography/methods* ; Epilepsy/physiopathology* ; Epilepsy, Absence ; Humans ; Methylprednisolone ; Scalp ; Seizures ; Sleep

Temporal lobe epilepsy (TLE) is a common and severe neurological disorder which is often intractable. It can not only damage the normal structure and function of hippocampus, but also affect the neurogenesis in dentate gyrus (DG). It is well documented from researches on the animal models of TLE that after a latent period of several days, prolonged seizure activity leads to a dramatic increase in mitotic activity in the hippocampal DG. However, cell proliferation returns to baseline levels within 3-4 weeks after status epilepticus (SE). Meanwhile, there are two major abnormalities of DG neurogenesis, including the formation of hilar basal dendrites and the ectopic migration of newborn granule cells into the polymorphic cell layer, which may affect epileptogenesis and seizure onset. However, the specific contribution of these abnormalities to seizures is still unknown. In other words, whether they are anti-epileptic or pro-epileptic is still under heated discussion. This article systematically reviews current knowledge on neurogenesis and epilepsy based on the results of studies in recent years and discusses the possible roles of neurogenesis in epileptogenesis and pathologic mechanisms, so as to provide information for the potential application of neurogenesis as a new clinical therapeutic target for temporal lobe epilepsy.
Animals ; Brain ; Cell Movement ; physiology ; Cell Proliferation ; physiology ; Dendrites ; pathology ; Dentate Gyrus ; growth & development ; pathology ; Epilepsy, Temporal Lobe ; etiology ; pathology ; physiopathology ; Hippocampus ; growth & development ; pathology ; Humans ; Mitosis ; physiology ; Neurogenesis ; physiology ; Neurons ; pathology ; Seizures ; etiology ; physiopathology ; Status Epilepticus ; physiopathology

To investigate the effect of crocin on the progression and generalized seizure of temporal lobe epilepsy in mice.Hippocampus rapid kindling model was established in C57BL/6J mice. The effects of crocin on seizure stage, afterdischarge duration (ADD), number of stimulation in each stage and final state, the incidence of generalized seizure (GS), average seizure stage and ADD were observed.Crocin (20 mg/kg) significantly retarded behavioral seizure stages (<0.05) and shortened cumulative ADD (<0.01) during hippocampus rapid kindling acquisition in mice compared with vehicle group. Meanwhile, number of stimulations in stage 1-2 was significantly increased (<0.05) and the incidence of fully kindled animals was significantly decreased (<0.01). However, 10 or 50 mg/kg crocin showed no significant effect on the above indexes (all>0.05). Crocin (100 or 200 mg/kg) significantly decreased the incidence of GS (all<0.01) and reduced average seizure stages (all<0.01) in fully-kindled mice compared with vehicle group; Fifty mg/kg crocin only reduced average seizure stages (<0.05).Low-dose crocin can retard the progression in hippocampus rapid kindling acquisition in mice, while high-dose crocin relieves the GS in fully-kindled mice, which suggests that crocin may be a potential anti-epileptic compound.
Animals ; Anticonvulsants ; pharmacology ; Carotenoids ; pharmacology ; therapeutic use ; Dose-Response Relationship, Drug ; Electric Stimulation ; Epilepsy, Temporal Lobe ; chemically induced ; drug therapy ; Hippocampus ; drug effects ; physiopathology ; Kindling, Neurologic ; drug effects ; physiology ; Mice ; Mice, Inbred C57BL ; Seizures ; classification ; drug therapy

PURPOSE: Glufosinate poisoning can cause neurologic complications that may be difficult to treat due to delayed manifestation. Studies assessing possible predictors of complications are lacking. Although serum ammonia level is a potential predictor of severe neurotoxicity, it has only been assessed via case reports. Therefore, we investigated factors that predict neurologic complications in acute glufosinate-poisoned patients. MATERIALS AND METHODS: We conducted a retrospective review of 45 consecutive glufosinate-poisoning cases that were diagnosed in the emergency department (ED) of Wonju Severance Christian Hospital between May 2007 and July 2014. Patients with a Glasgow Coma Scale (GCS) score of <8, seizure, and/or amnesia were defined to a neurologic complication group. RESULTS: The neurologic complication group (29 patients, 64.4%) comprised patients with GCS<8 (27 patients, 60.0%), seizure (23 patients, 51.1%), and amnesia (5 patients, 11.1%). Non-neurologic complications included respiratory failure (14 patients, 31.1%), intubation and ventilator care (23 patients, 51.1%), shock (2 patients, 4.4%), pneumonia (16 patients, 35.6%), acute kidney injury (10 patients, 22.2%), and death (4 patients, 8.9%). Complications of GCS<8, seizure, respiratory failure, and intubation and ventilator care appeared during latent periods within 11 hrs, 34 hrs, 14 hrs, and 48 hrs, respectively. Initial serum ammonia was a predictor of neurologic complications [odds ratio 1.039, 95% confidence interval (1.001-1.078), p=0.046 and area under the curve 0.742]. CONCLUSION: Neurologic complications developed in 64.4% of patients with acute glufosinate poisoning. The most common complication was GCS<8. Initial serum ammonia level, which can be readily assessed in the ED, was a predictor of neurologic complications.
Adult ; Aged ; Aged, 80 and over ; Aminobutyrates/blood/*poisoning ; Ammonia/*blood ; *Emergency Service, Hospital ; Female ; Glasgow Coma Scale ; Humans ; Male ; Middle Aged ; Nausea/etiology ; Neurotoxicity Syndromes/blood/immunology/*physiopathology ; Respiratory Insufficiency/etiology ; Retrospective Studies ; Seizures/etiology ; Severity of Illness Index ; Vomiting/etiology

BACKGROUNDMitochondrial diseases are a group of energy metabolic disorders with multisystem involvements. Variable clinical features present a major challenge in pediatric diagnoses. We summarized the clinical spectrum of m.3243A>G mutation in Chinese pediatric patients, to define the common clinical manifestations and study the correlation between heteroplasmic degree of the mutation and clinical severity of the disease.

METHODSClinical data of one-hundred pediatric patients with symptomatic mitochondrial disease harboring m.3243A>G mutation from 2007 to 2013 were retrospectively reviewed. Detection of m.3243A>G mutation ratio was performed by polymerase chain reaction (PCR)-restriction fragment length polymorphism. Correlation between m.3243A>G mutation ratio and age was evaluated. The differences in clinical symptom frequency of patients with low, middle, and high levels of mutation ratio were analyzed by Chi-square test.

RESULTSSixty-six patients (66%) had suffered a delayed diagnosis for an average of 2 years. The most frequent symptoms were seizures (76%), short stature (73%), elevated plasma lactate (70%), abnormal magnetic resonance imaging/computed tomography (MRI/CT) changes (68%), vomiting (55%), decreased vision (52%), headache (50%), and muscle weakness (48%). The mutation ratio was correlated negatively with onset age (r = -0.470, P < 0.001). Myopathy was more frequent in patients with a high level of mutation ratio. However, patients with a low or middle level of m.3243A>G mutation ratio were more likely to suffer hearing loss, decreased vision, and gastrointestinal disturbance than patients with a high level of mutation ratio.

CONCLUSIONSOur study showed that half of Chinese pediatric patients with m.3243A>G mutation presented seizures, short stature, abnormal MRI/CT changes, elevated plasma lactate, vomiting, and headache. Pediatric patients with these recurrent symptoms should be considered for screening m.3243A>G mutation. Clinical manifestations and laboratory abnormalities should be carefully monitored in patients with this point mutation.

Adolescent ; Age of Onset ; Asian Continental Ancestry Group ; Chi-Square Distribution ; Child ; Child, Preschool ; DNA, Mitochondrial ; genetics ; Female ; Hearing Loss ; pathology ; physiopathology ; Humans ; Infant ; Lactic Acid ; blood ; Magnetic Resonance Imaging ; Male ; Mitochondrial Diseases ; blood ; genetics ; pathology ; physiopathology ; Mutation ; Point Mutation ; genetics ; Polymerase Chain Reaction ; Retrospective Studies ; Seizures ; pathology ; physiopathology