EEG Seizure Research Articles

Predictors of seizure detection and EEG clinical impact in an italian tertiary emergency department.

Epileptic seizures pose challenges in emergency departments(ED), affecting up to 10% of admitted patients. This study aimed to assess emergency electroencephalogram (EmEEG) utilization, identifying factors predicting seizure detection and its influence on clinical decisions. A retrospective review of 1135 EmEEGs on 1017 patients at a tertiary teaching hospital between June 2022 and June 2023 was conducted. Data included demographics, medical history, EmEEG indications, neuroimaging findings, and clinical outcomes. Statistical analyses utilized Fisher's exact tests and logistic regression models. EmEEG detected status epilepticus-related seizures in 5.40% of cases, seizures without status epilepticus in 3.05%, and status epilepticus without discrete seizures in 3.74%. Epileptiform abnormalities were noted in 22.12% of EmEEGs. EmEEG influenced initial diagnoses (21.24%), antiseizure medication changes (20.85%), and discharge decisions (39.04%). Predictors for seizures/status epilepticus included previous neurosurgery, seizures in the ED, and cognitive/behavioral impairment (p < 0.001). EmEEG significantly altered initial diagnoses based on witnessed seizures, involuntary movements, epileptiform abnormalities, and 1-2Hz generalized periodic discharges (p < 0.001). Changes in antiseizure medications correlated with seizure occurrence, neuroimaging results, epileptiform abnormalities, and EEG background slowing (p < 0.001). Factors influencing discharge decisions included previous neurosurgery, consciousness impairment, acute neuroimaging pathology, EEG focal slowing, and EEG background slowing (p < 0.001). The study clarifies EmEEG's role in modifying initial diagnoses, treatment approaches, and discharge decisions. The study provides insights into the nuanced impact of EmEEG in different clinical scenarios, offering valuable guidance for clinicians in selecting patients for EmEEG, particularly in conditions of limited EEG availability.

The Role of Routine Electroencephalography in the Diagnosis of Seizures in Medical Intensive Care Units.

Seizures should be diagnosed and treated to ensure optimal health outcomes in critically ill patients admitted in the medical intensive care unit (MICU). Continuous electroencephalography is still infrequently used in the MICU. We investigated the effectiveness of routine EEG (rEEG) in detecting seizures in the MICU. A total of 560 patients admitted to the MICU between October 2018 and March 2023 and who underwent rEEG were reviewed. Seizure-related rEEG constituted 47% of all rEEG studies. Totally, 39% of the patients experienced clinical seizures during hospitalization; among them, 48% experienced the seizure, and 13% experienced their first seizure after undergoing an rEEG study. Seventy-seven percent of the patients had unfavorable short-term outcomes. Patients with cardiovascular diseases were the most likely to have the suppression/burst suppression (SBS) EEG pattern and the highest mortality rate. The rhythmic and periodic patterns (RPPs) and electrographic seizure (ESz) EEG pattern were associated with seizures within 24 h after rEEG, which was also related to unfavorable outcomes. Significant predictors of death were age > 59 years, the male gender, the presence of cardiovascular disease, a Glasgow Coma Scale score ≤ 5, and the SBS EEG pattern, with a predictive performance of 0.737 for death. rEEG can help identify patients at higher risk of seizures. We recommend repeated rEEG in patients with ESz or RPP EEG patterns to enable a more effective monitoring of seizure activities.

Seizure suppression and neuroprotection in soman-exposed rats following delayed intramuscular treatment of adenosine A1 receptor agonist as an adjunct to standard medical treatment

Soman produces excitotoxic effects by inhibiting acetylcholinesterase in the cholinergic synapses and neuromuscular junctions, resulting in soman-induced sustained status epilepticus (SSE). Our previous work showed delayed intramuscular (i.m.) treatment with A1 adenosine receptor agonist N-bicyclo-[2.2.1]-hept-2-yl-5′-chloro-5′-deoxyadenosine (ENBA) alone suppressed soman-induced SSE and prevented neuropathology. Using this same rat soman seizure model, we tested if delayed therapy with ENBA (60 mg/kg, i.m.) would terminate seizure, protect neuropathology, and aid in survival when given in conjunction with current standard medical countermeasures (MCMs): atropine sulfate, 2-PAM, and midazolam (MDZ). Either 15- or 30-min following soman-induced SSE onset, male rats received atropine and 2-PAM plus either MDZ or MDZ + ENBA. Electroencephalographic (EEG) activity, physiologic parameters, and motor function were recorded. Either 2- or 14-days following exposure surviving rats were euthanized and perfused for histology. All animals treated with MDZ + ENBA at both time points had 100% EEG seizure termination and reduced total neuropathology compared to animals treated with MDZ (2-day, p = 0.015 for 15-min, p = 0.002 for 30-min; 14-day, p < 0.001 for 15-min, p = 0.006 for 30-min), showing ENBA enhanced MDZ's anticonvulsant and neuroprotectant efficacy. However, combined MDZ + ENBA treatment, when compared to MDZ treatment groups, had a reduction in the 14-day survival rate regardless of treatment time, indicating possible enhancement of MDZ's neuronal inhibitory effects by ENBA. Based on our findings, ENBA shows promise as an anticonvulsant and neuroprotectant in a combined treatment regimen following soman exposure; when given as an adjunct to standard MCMs, the dose of ENBA needs to be adjusted.

Clinical analysis of developmental and/or epileptic encephalopathy with spike-and-wave activation in sleep: A single tertiary care center experience in China

PurposeTo analyze the electroclinical features of patients with developmental and/or epileptic encephalopathy with spike-and-wave activation in sleep (DEE/EE-SWAS) and study the efficacy of different therapies on seizure control, electroencephalogram (EEG) improvements of electrical status epilepticus during sleep (ESES), and cognition outcomes. MethodsPatients with DEE/EE-SWAS who underwent at least one follow-up EEG 3 months after therapy were retrospectively enrolled. The demographic and clinical characteristics of the patients were analyzed. Variables that influenced the outcomes were evaluated using logistic regression models. ResultsIn total, 87 patients (47 males) were included. The median age at ESES recognition was 81.0 months (IQR 64.0, 96.0). Forty-six patients were diagnosed with self-limited focal epilepsies (SeLFEs) before ESES recognition, 24 with developmental and epileptic encephalopathies with spike-and-wave activation in sleep (DEE-SWAS), and 17 with other epilepsies. Steroids, benzodiazepines, and antiseizure medications (ASMs) were the initial treatment options for ESES. Patients with structural etiologies or slow EEG backgrounds at the time of ESES recognition were less likely to respond to treatment than other patients. However, only children with slow EEG backgrounds had lower odds of response in logistic regression models. Children with clinical or EEG response showed improvements in cognition. ConclusionSteroids, benzodiazepines, and ASMs are effective treatments for patients with DEE/EE-SWAS. Children with structural etiologies or slow EEG backgrounds at the time of ESES recognition may have a poor long-term prognosis. The efficacy of seizure reduction and EEG improvement is associated with cognitive improvement.

Gershgorin circle theorem-based feature extraction for biomedical signal analysis.

Recently, graph theory has become a promising tool for biomedical signal analysis, wherein the signals are transformed into a graph network and represented as either adjacency or Laplacian matrices. However, as the size of the time series increases, the dimensions of transformed matrices also expand, leading to a significant rise in computational demand for analysis. Therefore, there is a critical need for efficient feature extraction methods demanding low computational time. This paper introduces a new feature extraction technique based on the Gershgorin Circle theorem applied to biomedical signals, termed Gershgorin Circle Feature Extraction (GCFE). The study makes use of two publicly available datasets: one including synthetic neural recordings, and the other consisting of EEG seizure data. In addition, the efficacy of GCFE is compared with two distinct visibility graphs and tested against seven other feature extraction methods. In the GCFE method, the features are extracted from a special modified weighted Laplacian matrix from the visibility graphs. This method was applied to classify three different types of neural spikes from one dataset, and to distinguish between seizure and non-seizure events in another. The application of GCFE resulted in superior performance when compared to seven other algorithms, achieving a positive average accuracy difference of 2.67% across all experimental datasets. This indicates that GCFE consistently outperformed the other methods in terms of accuracy. Furthermore, the GCFE method was more computationally-efficient than the other feature extraction techniques. The GCFE method can also be employed in real-time biomedical signal classification where the visibility graphs are utilized such as EKG signal classification.

Distinct profiles of cerebral oxygenation in focal vs. secondarily generalized EEG seizures in children undergoing cardiac surgery.

Seizures are common in children undergoing cardiopulmonary bypass (CPB). Cerebral oxygen saturation (ScO2) by near-infrared spectroscopy is routinely monitored in many centers, but the relations between the levels and changes of ScO2 and brain injuries remain incompletely understood. We aimed to analyze the postoperative profiles of ScO2 and cerebral blood flow velocity in different types of EEG seizures in relation to brain injuries on MRI. We monitored continuous EEG and ScO2 in 337 children during the first 48 h after CPB, which were analyzed in 3 h periods. Cerebral blood flow peak systolic velocity (PSV) in the middle cerebral artery was measured daily by transcranial Doppler. Postoperative cerebral MRI was performed before hospital discharge. Based on the occurrence and spreading types of seizures, patients were divided into three groups as patients without seizures (Group N; n = 309), those with focal seizures (Group F; n = 13), or with secondarily generalized seizures (Group G; n = 15). There were no significant differences in the onset time and duration of seizures and incidence of status epilepticus between the two seizures groups (Ps ≥ 0.27). ScO2 increased significantly faster across Group N, Group G, and Group F during the 48 h (p < 0.0001) but its overall levels were not significantly different among the three groups (p = 0.30). PSV was significantly lower (p = 0.003) but increased significantly faster (p = 0.0003) across Group N, Group G, and Group F. Group F had the most severe brain injuries and the highest incidence of white matter injuries on MRI among the three groups (Ps ≤ 0.002). Postoperative cerebral oxygenation showed distinct profiles in secondarily generalized and particularly focal types of EEG seizures in children after CPB. A state of 'overshooting' ScO2 with persistently low PSV was more frequently seen in those with focal seizures and more severe brain injury. Information from this study may have important clinical implications in detecting brain injuries when monitoring cerebral oxygenation in this vulnerable group of children after CPB.

Recovery From Homonymous Hemianopia in Hyperglycemia-Induced Occipital Lobe Electroclinical Seizures Following Glycemic Control.

Visual changes due to hyperglycemia in diabetes are not uncommon. While blurred vision is a well-established sequela of chronic hyperglycemia, homonymous hemianopia with or without electroclinical seizures is much rarer and can be mistaken for migraine, temporal arteritis, or ischemia of the central nervous system. This article analyzed case studies for 3 patients (67M, 68M, 52F) presenting with complex visual phenomena, from 3 to 42 days duration, including pathogenesis, clinical findings, management, and follow-up. Examinations demonstrated dense left homonymous hemianopias in 2 patients and a left inferior homonymous quadrantanopia in one, with no other abnormalities. Patients described vivid, nonstereotyped intermittent hallucinations in the affected fields. Blood glucose levels ranged from 13.5 to 35.0 mmol/L (243-630 mg/dL) without ketosis and HbA1c from 14.6% to 16.8%. Computed tomography of the brain showed no acute intracranial pathology. MRI of the brain either detected no abnormalities or demonstrated changes consistent with seizure activity. Electroencephalogram (EEG) demonstrated seizures over the right occipital region in each patient. EEG seizures coincided with patients' hallucinations, while they remained otherwise conscious. Oral hypoglycemic and antiepileptic medications were commenced with rapid and complete reversal of the seizures and visual field deficits, confirmed by repeat Automated 30-2 and MRI. Hyperglycemia-induced occipital lobe seizures with visual hallucinations and interictal homonymous visual field defects represent a rare but clinically important diagnosis. This article highlights the importance of prompt recognition and treatment to facilitate recovery.

The Surgical Method of Craniectomy Differentially Affects Acute Seizures, Brain Deformation, and Behavior in a Traumatic Brain Injury Animal Model.

Traumatic brain injury (TBI) is the leading cause of morbidity and mortality worldwide. Multiple injury models have been developed to study this neurological disorder. One such model is the lateral fluid percussion injury (LFPI) rodent model. The LFPI model can be generated with different surgical procedures that could affect the injury and be reflected in neurobehavioral dysfunction and acute electroencephalograph (EEG) changes. A craniectomy was performed either with a trephine hand drill or with a trephine electric drill that was centered over the left hemisphere of adult, male Sprague Dawley rats. Sham craniectomy groups were assessed by hand-drilled (ShamHMRI) and electric-drilled (ShamEMRI) to evaluate by magnetic resonance imaging (MRI). Then, TBI was induced in separate groups, (TBIH) and (TBIE), using a fluid-percussion device. Sham-injured rats (ShamH/ShamE) underwent the same surgical procedures as the TBI rats. During the same surgery session, rats were implanted with screw and microwire electrodes positioned in the neocortex and hippocampus and the EEG activity was recorded 24 h for the first 7 days after TBI for assessing the acute EEG seizure and gamma event coupling. The electric drilling craniectomy induced greater tissue damage and sensorimotor deficits compared with the hand drill. Analysis of the EEG revealed acute seizures in at least one animal from each group after the procedure. Both TBI and Sham rats from the electric drill groups had a significant greater total number of seizures than the animals that were craniectomized manually (p < 0.05). Similarly, EEG functional connectivity was lower in ShamE compared with ShamH rats. These results suggest that electrical versus hand-drilling craniectomies produce cortical injury in addition to the LFPI which increases the likelihood for acute post-traumatic seizures. Differences in the surgical approach could be one reason for the variability in the injury that makes it difficult to replicate results between preclinical TBI studies.

De novo variants in GABRA4 are associated with a neurological phenotype including developmental delay, behavioral abnormalities and epilepsy

Nine out of 19 genes encoding GABAA receptor subunits have been linked to monogenic syndromes characterized by seizures and developmental disorders. Previously, we reported the de novo variant p.(Thr300Ile) in GABRA4 in a patient with epilepsy and neurodevelopmental abnormalities. However, no new cases have been reported since then. Through an international collaboration, we collected molecular and phenotype data of individuals carrying de novo variants in GABRA4. Patients and their parents were investigated either by exome or genome sequencing, followed by targeted Sanger sequencing in some cases. All variants within the transmembrane domain, including the previously reported p.(Thr300Ile) variant, were characterized in silico and analyzed by molecular dynamics (MD) simulation studies. We identified three novel de novo missense variants in GABRA4 (NM_000809.4): c.797 C > T, p.(Pro266Leu), c.899 C > A, p.(Thr300Asn), and c.634 G > A, p.(Val212Ile). The p.(Thr300Asn) variant impacts the same codon as the previously reported variant p.(Thr300Ile) and likely arose post-zygotically as evidenced by sequencing oral mucosal cells. Overlapping phenotypes among affected individuals included developmental delay (4/4), epileptiform EEG abnormalities (3/4), attention deficits (3/4), seizures (2/4), autistic features (2/4) and structural brain abnormalities (2/4). MD simulations of the three variants within the transmembrane domain of the receptor indicate that sub-microsecond scale dynamics differ between wild-type and mutated subunits. Taken together, our findings further corroborate an association between GABRA4 and a neurological phenotype including variable neurodevelopmental, behavioral and epileptic abnormalities.

Semiology and Neurophysiology of Clonic Seizures: A Report of 39 Patients.

Clonic seizures are currently defined as repetitive and rhythmic myoclonic contractions of a specific body part, producing twitching movements at a frequency of 0.2-5 Hz. There are few studies in the literature that have reported a detailed analysis of the semiology, neurophysiology, and lateralizing value of clonic seizures. In this article, we aim to report our findings from a retrospective review of 39 patients. We identified 39 patients (48 seizures) from our center who had been admitted with clonic seizures between 2016 and 2022. We performed a retrospective review of their video-EEG recordings for semiology and ictal EEG findings. Seventeen patients also had simultaneous surface-EMG (sEMG) electrodes placed on affected body parts, which were analyzed as well. The most common initial affected body parts were face, arm, and hand. In most of the cases, seizures propagated from lower face to upper face and distal hand to proximal arm. The most common seizure-onset zone was the perirolandic region, and the most common EEG seizure pattern was paroxysmal rhythmic monomorphic activity. The lateralizing value for EEG seizure onset to contralateral hemisphere in unilateral clonic seizures (n = 39) was 100%. All seizures recorded with sEMG electrodes demonstrated synchronous brief tetanic contractions of agonists and antagonists, alternating with synchronous silent periods. Arrhythmic clonic seizures were associated with periodic epileptiform discharges on the EEG, whereas rhythmic clonic seizures were associated with paroxysmal rhythmic monomorphic activity. Overall, the most common etiology was cerebrovascular injuries, followed by tumors. Clonic seizures are characterized by synchronized brief tetanic contractions of agonist and antagonistic muscles alternating with synchronized silent periods, giving rise to the visible twitching. The most common seizure onset zone is in the perirolandic region, which is consistent with the symptomatogenic zone being in the primary motor area. The lateralizing value of unilateral clonic seizures for seizure onset in the contralateral hemisphere is 100%.

Unraveling the intricacies of EEG seizure detection: A comprehensive exploration of machine learning model performance, interpretability, and clinical insights

Unraveling the intricacies of EEG seizure detection: A comprehensive exploration of machine learning model performance, interpretability, and clinical insights

Automatic classification of seizure and seizure-free EEG signals based on phase space reconstruction features.

Epilepsy is a type of brain disorder triggered by an abrupt electrical imbalance of neuronal networks. An electroencephalogram (EEG) is a diagnostic tool to capture the underlying brain mechanisms and detect seizure onset in epileptic patients. To detect seizures, neurologists need to manually monitor EEG recordings for long periods, which is challenging and susceptible to errors depending on expertise and experience. Therefore, automatic identification of seizure and seizure-free EEG signals becomes essential. This study introduces a method based on the features extracted from the phase space reconstruction for classifying seizure and seizure-free EEG signals. The computed features are derived from the elliptical area and interquartile range of the Euclidean distance by varying percentage values of data points ranging from 50 to 100%. We consider two public datasets and evaluate these features in each EEG epoch that includes the healthy, interictal, preictal, and ictal stages of epileptic subjects, utilizing the K-nearest neighbor classifier for classification. Results show that the features have higher values during the seizure than the seizure-free EEG signals and healthy subjects. Furthermore, the proposed features can effectively discriminate seizure EEG signals from the seizure-free and normal subjects with 100% accuracy, sensitivity, and specificity in both datasets. Likewise, the classification between the preictal stage and seizure EEG signals attains 98% accuracy. Overall, the reconstructed phase space features significantly enhance the accuracy of detecting epileptic EEG signals compared with existing methods. This advancement holds great potential in assisting neurologists in swiftly and accurately diagnosing epileptic seizures from EEG signals.

Pediatric and Adolescent Seizure Detection: A Machine Learning Approach Exploring the Influence of Age and Sex in Electroencephalogram Analysis

Background: Epilepsy, a prevalent neurological disorder characterized by recurrent seizures affecting an estimated 70 million people worldwide, poses a significant diagnostic challenge. EEG serves as an important tool in identifying these seizures, but the manual examination of EEGs by experts is time-consuming. To expedite this process, automated seizure detection methods have emerged as powerful aids for expert EEG analysis. It is worth noting that while such methods are well-established for adult EEGs, they have been underdeveloped for pediatric and adolescent EEGs. This study sought to address this gap by devising an automatic seizure detection system tailored for pediatric and adolescent EEG data. Methods: Leveraging publicly available datasets, the TUH pediatric and adolescent EEG and CHB-MIT EEG datasets, the machine learning-based models were constructed. The TUH pediatric and adolescent EEG dataset was divided into training (n = 118), validation (n = 19), and testing (n = 37) subsets, with special attention to ensure a clear demarcation between the individuals in the training and test sets to preserve the test set’s independence. The CHB-MIT EEG dataset was used as an external test set. Age and sex were incorporated as features in the models to investigate their potential influence on seizure detection. Results: By leveraging 20 features extracted from both time and frequency domains, along with age as an additional feature, the method achieved an accuracy of 98.95% on the TUH test set and 64.82% on the CHB-MIT external test set. Our investigation revealed that age is a crucial factor for accurate seizure detection in pediatric and adolescent EEGs. Conclusion: The outcomes of this study hold substantial promise in supporting researchers and clinicians engaged in the automated analysis of seizures in pediatric and adolescent EEGs.

Deep Learning for Generalized EEG Seizure Detection after Hypoxia-Ischemia-Preclinical Validation.

Brain maturity and many clinical treatments such as therapeutic hypothermia (TH) can significantly influence the morphology of neonatal EEG seizures after hypoxia-ischemia (HI), and so there is a need for generalized automatic seizure identification. This study validates efficacy of advanced deep-learning pattern classifiers based on a convolutional neural network (CNN) for seizure detection after HI in fetal sheep and determines the effects of maturation and brain cooling on their accuracy. The cohorts included HI-normothermia term (n = 7), HI-hypothermia term (n = 14), sham-normothermia term (n = 5), and HI-normothermia preterm (n = 14) groups, with a total of >17,300 h of recordings. Algorithms were trained and tested using leave-one-out cross-validation and k-fold cross-validation approaches. The accuracy of the term-trained seizure detectors was consistently excellent for HI-normothermia preterm data (accuracy = 99.5%, area under curve (AUC) = 99.2%). Conversely, when the HI-normothermia preterm data were used in training, the performance on HI-normothermia term and HI-hypothermia term data fell (accuracy = 98.6%, AUC = 96.5% and accuracy = 96.9%, AUC = 89.6%, respectively). Findings suggest that HI-normothermia preterm seizures do not contain all the spectral features seen at term. Nevertheless, an average 5-fold cross-validated accuracy of 99.7% (AUC = 99.4%) was achieved from all seizure detectors. This significant advancement highlights the reliability of the proposed deep-learning algorithms in identifying clinically translatable post-HI stereotypic seizures in 256Hz recordings, regardless of maturity and with minimal impact from hypothermia.

Measurable transitions during seizures in intracranial EEG: A stereoelectroencephalography and SPECT study

ObjectiveIctal Single Photon Emission Computed Tomography (SPECT) and stereo-electroencephalography (SEEG) are diagnostic techniques used for the management of patients with drug-resistant focal epilepsies. While hyperperfusion patterns in ictal SPECT studies reveal seizure onset and propagation pathways, the role of ictal hypoperfusion remains poorly understood. The goal of this study was to systematically characterize the spatio-temporal information flow dynamics between differently perfused brain regions using stereo-EEG recordings. MethodsWe identified seizure-free patients after resective epilepsy surgery who had prior ictal SPECT and SEEG investigations. We estimated directional connectivity between the epileptogenic-zone (EZ), non-resected areas of hyperperfusion, hypoperfusion, and baseline perfusion during the interictal, preictal, ictal, and postictal periods. ResultsCompared to the background, we noted significant information flow (1) during the preictal period from the EZ to the baseline and hyperperfused regions, (2) during the ictal onset from the EZ to all three regions, and (3) during the period of seizure evolution from the area of hypoperfusion to all three regions. ConclusionsHypoperfused brain regions were found to indirectly interact with the EZ during the ictal period. SignificanceOur unique study, combining intracranial electrophysiology and perfusion imaging, presents compelling evidence of dynamic changes in directional connectivity between brain regions during the transition from interictal to ictal states.

Falls in oxygen saturations accompany electrographic seizures in term neonates: an observational study

BackgroundEffective seizure detection is important however, clinical signs of seizure activity may be subtle in neonates. This study aimed to systematically investigate SpO2 and respiratory pattern changes associated with EEG seizures in term-born neonates.MethodAn observational study in term neonates at risk of seizures admitted to a single tertiary level neonatal intensive care unit. Synchronised high-resolution physiological data (ECG, pulse oximetry, respiration) and EEG/amplitude-integrated EEG (aEEG) monitoring were recorded. Sections of traces with evidence of clear EEG seizure activity were compared with physiological data recorded at the same time.Results22/44 (50%) neonates who had aEEG monitoring were noted to have electrographic seizures. Physiologic download measurements were available for 11 of these neonates. In nine of these, an acute drop in oxygen saturation (SpO2) of at least 5% was noted in at least one seizure. Accompanying apnoeas were noted in three neonates.ConclusionAcute decreases in SpO2 were seen in term neonates associated with seizures and these were not always accompanied by an apnoeic episode. Physiologic download in association with EEG monitoring may assist in improving seizure detection. Unexplained drops in SpO2 could indicate further investigation for possible seizures in at-risk neonates.ImpactA decrease in blood oxygen saturation (SpO2) associated with EEG seizures can occur in term infants with HIE or perinatal stroke.Drops in SpO2 associated with EEG seizures in term infants with HIE or stroke may occur in the absence of apnoeas.Unexplained acute falls in SpO2 in sick neonates may suggest possible seizures.Drops in SpO2 associated with seizures in term infants can occur over less than 3 minutes.Physiological monitoring alongside EEG monitoring could help to improve seizure detection.

Supervised and Unsupervised Deep Learning Approaches for EEG Seizure Prediction.

Epilepsy affects more than 50 million people worldwide, making it one of the world's most prevalent neurological diseases. The main symptom of epilepsy is seizures, which occur abruptly and can cause serious injury or death. The ability to predict the occurrence of an epileptic seizure could alleviate many risks and stresses people with epilepsy face. We formulate the problem of detecting preictal (or pre-seizure) with reference to normal EEG as a precursor to incoming seizure. To this end, we developed several supervised deep learning approaches model to identify preictal EEG from normal EEG. We further develop novel unsupervised deep learning approaches to train the models on only normal EEG, and detecting pre-seizure EEG as an anomalous event. These deep learning models were trained and evaluated on two large EEG seizure datasets in a person-specific manner. We found that both supervised and unsupervised approaches are feasible; however, their performance varies depending on the patient, approach and architecture. This new line of research has the potential to develop therapeutic interventions and save human lives.

Place of the Ketogenic Diet in Refractory Infantile Spasms

Introduction: The ketogenic diet is an alternative nutritional regimen for refractory epilepsies, including infantile spasms. The objective of our study is the primary and secondary evaluation at 12 months of the effectiveness and tolerance of the ketogenic diet on seizure frequency, EEG (electroencephalogram) patterns, and side effects. Patients and Methods: We evaluated the ketogenic diet as a second- or third-line treatment, following corticosteroid therapy and/or Vigabatrin. We assessed the effectiveness and tolerance of the ketogenic diet based on patients being seizure-free, experiencing a reduction in seizures of more than 50%, improvement in EEG background pattern (suppression of hypsarrhythmia), psychomotor development, side effects, and reduction in antiepileptic drugs during clinical, electrical, and neurocognitive follow-ups at 1, 3, and 12 months. Results: Thirty-eight children, twentytwo boys, and sixteen girls, with a gender ratio of 1.37, presenting with infantile spasms visualized directly clinically or through videos and electric evaluation, with or without hypsarrhythmia on EEG, all symptomatic: genetic etiologies (14, 36.9%), structural in 18 patients (47.3%), and metabolic in six patients (15.8%). At 3 months, the number of patients who were seizure-free with a normal EEG was 22 patients (57.9%), and those with a reduction in seizures of more than 50%, 15 patients (39.5%). At 12 months on the ketogenic diet, the number of patients who were seizure-free was 24 (63.2%), and those with a reduction in seizures of more than 50% 14 patients(36.8%). Good tolerance and efficiency of the diet were confirmed by treatment compliance, maintained high response rate, and a reduction in the number of antiepileptic drugs to 1.97. The children became more vigilant, attentive, and responsive to their parents. Conclusion: Implementing the ketogenic diet within a short period of time, less than a year, would yield even better results in terms of seizure frequency and behavior in a developing infant brain.

Minimizing artifact-induced false-alarms for seizure detection in wearable EEG devices with gradient-boosted tree classifiers

Electroencephalography (EEG) is widely used to monitor epileptic seizures, and standard clinical practice consists of monitoring patients in dedicated epilepsy monitoring units via video surveillance and cumbersome EEG caps. Such a setting is not compatible with long-term tracking under typical living conditions, thereby motivating the development of unobtrusive wearable solutions. However, wearable EEG devices present the challenges of fewer channels, restricted computational capabilities, and lower signal-to-noise ratio. Moreover, artifacts presenting morphological similarities to seizures act as major noise sources and can be misinterpreted as seizures. This paper presents a combined seizure and artifacts detection framework targeting wearable EEG devices based on Gradient Boosted Trees. The seizure detector achieves nearly zero false alarms with average sensitivity values of 65.27%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$65.27\\%$$\\end{document} for 182 seizures from the CHB-MIT dataset and 57.26%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$57.26\\%$$\\end{document} for 25 seizures from the private dataset with no preliminary artifact detection or removal. The artifact detector achieves a state-of-the-art accuracy of 93.95%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$93.95\\%$$\\end{document} (on the TUH-EEG Artifact Corpus dataset). Integrating artifact and seizure detection significantly reduces false alarms—up to 96%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$96\\%$$\\end{document} compared to standalone seizure detection. Optimized for a Parallel Ultra-Low Power platform, these algorithms enable extended monitoring with a battery lifespan reaching 300 h. These findings highlight the benefits of integrating artifact detection in wearable epilepsy monitoring devices to limit the number of false positives.

An IoT-Based Novel Hybrid Seizure Detection Approach for Epileptic Monitoring

This research focuses on a new Electroencephalogram (EEG)-based system for the early detection of epileptic episodes that is made possible by the Internet of Things (IoT). The system is made up of two important units, namely, 1) the multichannel EEG recording unit; and 2) the seizure detection unit. Since epileptic information is more useful when included in multichannel EEG data from many brain regions, the primary goals of this work are: designing and developing the seizure detection unit, making use of Spike-Statistical (SS)- Flower Pollination Algorithm (FPA) based Critical Spectral Verge (CSV) derived features termed as SS-CSV; and employing the convolutional neural network (CNN) method in an IoT-enabled EEG monitoring system to detect EEG seizures. The presented system performed better, with an average accuracy of 98.48% with the CNN classifier. Neuroexperts will find this approach very useful to analyse seizure information, especially in wearable medical devices.