EEG power spectra parameterization and adaptive channel selection towards semi-supervised seizure prediction.

BACKGROUND: The seizure prediction algorithms have demonstrated their potential in mitigating epilepsy risks by detecting the pre-ictal state using ongoing electroencephalogram (EEG) signals. However, most of them require high-density EEG, which is burdensome to the patients for daily monitoring. Moreover, prevailing seizure models require extensive training with significant labeled data which is very time-consuming and demanding for the epileptologists. METHOD: To address these challenges, here we propose an adaptive channel selection strategy and a semi-supervised deep learning model respectively to reduce the number of EEG channels and to limit the amount of labeled data required for accurate seizure prediction. Our channel selection module is centered on features from EEG power spectra parameterization that precisely characterize the epileptic activities to identify the seizure-associated channels for each patient. The semi-supervised model integrates generative adversarial networks and bidirectional long short-term memory networks to enhance seizure prediction. RESULTS: Our approach is evaluated on the CHB-MIT and Siena epilepsy datasets. With utilizing only 4 channels, the method demonstrates outstanding performance with an AUC of 93.15% on the CHB-MIT dataset and an AUC of 88.98% on the Siena dataset. Experimental results also demonstrate that our selection approach reduces the model parameters and training time. CONCLUSIONS: Adaptive channel selection coupled with semi-supervised learning can offer the possible bases for a light weight and computationally efficient seizure prediction system, making the daily monitoring practical to improve patients' quality of life.

Differentiating Epileptic and Psychogenic Non-Epileptic Seizures Using Machine Learning Analysis of EEG Plot Images.

The treatment of epilepsy, the second most common chronic neurological disorder, is often complicated by the failure of patients to respond to medication. Treatment failure with anti-seizure medications is often due to the presence of non-epileptic seizures. Distinguishing non-epileptic from epileptic seizures requires an expensive and time-consuming analysis of electroencephalograms (EEGs) recorded in an epilepsy monitoring unit. Machine learning algorithms have been used to detect seizures from EEG, typically using EEG waveform analysis. We employed an alternative approach, using a convolutional neural network (CNN) with transfer learning using MobileNetV2 to emulate the real-world visual analysis of EEG images by epileptologists. A total of 5359 EEG waveform plot images from 107 adult subjects across two epilepsy monitoring units in separate medical facilities were divided into epileptic and non-epileptic groups for training and cross-validation of the CNN. The model achieved an accuracy of 86.9% (Area Under the Curve, AUC 0.92) at the site where training data were extracted and an accuracy of 87.3% (AUC 0.94) at the other site whose data were only used for validation. This investigation demonstrates the high accuracy achievable with CNN analysis of EEG plot images and the robustness of this approach across EEG visualization software, laying the groundwork for further subclassification of seizures using similar approaches in a clinical setting.

Calibrating Deep Learning Classifiers for Patient-Independent Electroencephalogram Seizure Forecasting.

The recent scientific literature abounds in proposals of seizure forecasting methods that exploit machine learning to automatically analyze electroencephalogram (EEG) signals. Deep learning algorithms seem to achieve a particularly remarkable performance, suggesting that the implementation of clinical devices for seizure prediction might be within reach. However, most of the research evaluated the robustness of automatic forecasting methods through randomized cross-validation techniques, while clinical applications require much more stringent validation based on patient-independent testing. In this study, we show that automatic seizure forecasting can be performed, to some extent, even on independent patients who have never been seen during the training phase, thanks to the implementation of a simple calibration pipeline that can fine-tune deep learning models, even on a single epileptic event recorded from a new patient. We evaluate our calibration procedure using two datasets containing EEG signals recorded from a large cohort of epileptic subjects, demonstrating that the forecast accuracy of deep learning methods can increase on average by more than 20%, and that performance improves systematically in all independent patients. We further show that our calibration procedure works best for deep learning models, but can also be successfully applied to machine learning algorithms based on engineered signal features. Although our method still requires at least one epileptic event per patient to calibrate the forecasting model, we conclude that focusing on realistic validation methods allows to more reliably compare different machine learning approaches for seizure prediction, enabling the implementation of robust and effective forecasting systems that can be used in daily healthcare practice.

Multiple back fractures as a result of the first seizure.

Non-traumatic fractures due to seizures are an overlooked diagnostic group. It is well known that patients with generalized tonic-clonic seizures have an increased trauma risk. However, the cause of fracture is rarely due to the violent forces of muscle contractions. Usually, the primary patient examination focuses on the aetiology of the seizure, which sometimes delays the diagnosis of fractures. This is a case report of a 19-year-old woman who sustained three compression fractures of the thoracic spine due to a generalized tonic-clonic seizure, and a discussion of the diagnostic challenges in such a rare case.

Unrecognized Focal Nonmotor Seizures in Adolescents Presenting to Emergency Departments.

BACKGROUND AND OBJECTIVES: Many adolescents with undiagnosed focal epilepsy seek evaluation in emergency departments (EDs). Accurate history-taking is essential to prompt diagnosis and treatment. In this study, we investigated ED recognition of motor vs nonmotor seizures and its effect on management and treatment of focal epilepsy in adolescents. METHODS: This was a retrospective analysis of enrollment data from the Human Epilepsy Project (HEP), an international multi-institutional study that collected data from 34 sites between 2012 and 2017. Participants were 12 years or older, neurotypical, and within 4 months of treatment initiation for focal epilepsy. We used HEP enrollment medical records to review participants' initial diagnosis and management. RESULTS: A total of 83 adolescents were enrolled between 12 and 18 years. Fifty-eight (70%) presented to an ED before diagnosis of epilepsy. Although most ED presentations were for motor seizures (n = 52; 90%), many patients had a history of nonmotor seizures (20/52 or 38%). Adolescents with initial nonmotor seizures were less likely to present to EDs (26/44 or 59% vs 32/39 or 82%, p = 0.02), and nonmotor seizures were less likely to be correctly identified (2/6 or 33% vs 42/52 or 81%, p = 0.008). A history of initial nonmotor seizures was not recognized in any adolescent who presented for a first-lifetime motor seizure. As a result, initiation of treatment and admission from the ED was not more likely for these adolescents who met the definition of epilepsy compared with those with no seizure history. This lack of nonmotor seizure history recognition in the ED was greater than that observed in the adult group (0% vs 23%, p = 0.03) and occurred in both pediatric and nonpediatric ED settings. DISCUSSION: Our study supports growing evidence that nonmotor seizures are often undiagnosed, with many individuals coming to attention only after conversion to motor seizures. We found this treatment gap is exacerbated in the adolescent population. Our study highlights a critical need for physicians to inquire about the symptoms of nonmotor seizures, even when the presenting seizure is motor. Future interventions should focus on improving nonmotor seizure recognition for this population in EDs.

Approach to Patients with Seizures and Epilepsy: A Guide for Primary Care Physicians.

Seizures and epilepsy are common neurologic conditions that are frequently encountered in the outpatient primary care setting. An accurate diagnosis relies on a thorough clinical history and evaluation. Understanding seizure semiology and classification is crucial in conducting the initial assessment. Knowledge of common seizure triggers and provoking factors can further guide diagnostic testing and initial management. The pharmacodynamic characteristics and side effect profiles of anti-seizure medications are important considerations when deciding treatment and counseling patients, particularly those with comorbidities and in special populations such as patient of childbearing potential.

Identification of neurological complications in childhood influenza: a random forest model.

BACKGROUND: Among the neurological complications of influenza in children, the most severe is acute necrotizing encephalopathy (ANE), with a high mortality rate and neurological sequelae. ANE is characterized by rapid progression to death within 1-2 days from onset. However, the knowledge about the early diagnosis of ANE is limited, which is often misdiagnosed as simple seizures/convulsions or mild acute influenza-associated encephalopathy (IAE). OBJECTIVE: To develop and validate an early prediction model to discriminate the ANE from two common neurological complications, seizures/convulsions and mild IAE in children with influenza. METHODS: This retrospective case-control study included patients with ANE (median age 3.8 (2.3,5.4) years), seizures/convulsions alone (median age 2.6 (1.7,4.3) years), or mild IAE (median age 2.8 (1.5,6.1) years) at a tertiary pediatric medical center in China between November 2012 to January 2020. The random forest algorithm was used to screen the characteristics and construct a prediction model. RESULTS: Of the 433 patients, 278 (64.2%) had seizures/convulsions alone, 106 (24.5%) had mild IAE, and 49 (11.3%) had ANE. The discrimination performance of the model was satisfactory, with an accuracy above 0.80 from both model development (84.2%) and internal validation (88.2%). Seizures/convulsions were less likely to be wrongly classified (3.7%, 2/54), but mild IAE (22.7%, 5/22) was prone to be misdiagnosed as seizures/convulsions, and a small proportion (4.5%, 1/22) of them was prone to be misdiagnosed as ANE. Of the children with ANE, 22.2% (2/9) were misdiagnosed as mild IAE, and none were misdiagnosed as seizures/convulsions. CONCLUSION: This model can distinguish the ANE from seizures/convulsions with high accuracy and from mild IAE close to 80% accuracy, providing valuable information for the early management of children with influenza.

Characterizing Frontal Lobe Seizure Semiology in Children.

OBJECTIVE: The objective was to analyze seizure semiology in pediatric frontal lobe epilepsy patients, considering age, to localize the seizure onset zone for surgical resection in focal epilepsy. METHODS: Fifty patients were identified retrospectively, who achieved seizure freedom after frontal lobe resective surgery at Great Ormond Street Hospital. Video-electroencephalography recordings of preoperative ictal seizure semiology were analyzed, stratifying the data based on resection region (mesial or lateral frontal lobe) and age at surgery (≤4 vs >4). RESULTS: Pediatric frontal lobe epilepsy is characterized by frequent, short, complex seizures, similar to adult cohorts. Children with mesial onset had higher occurrence of head deviation (either direction: 55.6% vs 17.4%; p = 0.02) and contralateral head deviation (22.2% vs 0.0%; p = 0.03), ictal body-turning (55.6% vs 13.0%; p = 0.006; ipsilateral: 55.6% vs 4.3%; p = 0.0003), and complex motor signs (88.9% vs 56.5%; p = 0.037). Both age groups (≤4 and >4 years) showed hyperkinetic features (21.1% vs 32.1%), contrary to previous reports. The very young group showed more myoclonic (36.8% vs 3.6%; p = 0.005) and hypomotor features (31.6% vs 0.0%; p = 0.003), and fewer behavioral features (36.8% vs 71.4%; p = 0.03) and reduced responsiveness (31.6% vs 78.6%; p = 0.002). INTERPRETATION: This study presents the most extensive semiological analysis of children with confirmed frontal lobe epilepsy. It identifies semiological features that aid in differentiating between mesial and lateral onset. Despite age-dependent differences, typical frontal lobe features, including hyperkinetic seizures, are observed even in very young children. A better understanding of pediatric seizure semiology may enhance the accuracy of onset identification, and enable earlier presurgical evaluation, improving postsurgical outcomes. ANN NEUROL 2024;95:1138-1148.

Genotype and phenotype features and prognostic factors of neonatal-onset pyridoxine-dependent epilepsy: A systematic review.

Pyridoxine-dependent epilepsy (PDE-ALDH7A1) is a rare autosomal recessive disorder due to a deficiency of α-aminoadipic semialdehyde dehydrogenase. This study aimed to systematically explore genotypic and phenotypic features and prognostic factors of neonatal-onset PDE. A literature search covering PubMed, Elsevier, and Web of Science was conducted from January 2006 to August 2023. We identified 56 eligible studies involving 169 patients and 334 alleles. The c.1279 G>C variant was the most common variant of neonatal-onset PDE (25.7 %). All patients were treated with pyridoxine; forty patients received dietary intervention therapy. 63.9 % of the patients were completely seizure-free; however, 68.6 % of the patients had neurodevelopmental delays. Additionally, homozygous c.1279 G>C variants were significantly associated with ventriculomegaly, abnormal white matter signal, and cysts (P<0.05). In contrast, homozygous c.1364 T>C was associated with clonic seizure (P=0.031). Pyridoxine used immediately at seizure onset was an independent protective factor for developmental delay (P=0.035; odds ratio [OR]: 3.14). Besides, pyridoxine used early in the neonatal period was a protective factor for language delay (P=0.044; OR: 4.59). In contrast, neonatal respiratory distress (P=0.001; OR: 127.44) and abnormal brain magnetic resonance imaging (P=0.049; OR: 3.64) were risk factors. Prenatal movement abnormality (P=0.041; OR: 20.56) and abnormal white matter signal (P=0.012; OR: 24.30) were risk factors for motor delay. Myoclonic seizure (P=0.023; OR: 7.13) and status epilepticus (P=0.000; OR: 9.93) were risk factors for breakthrough seizures. In conclusion, our study indicated that pyridoxine should be started immediately when unexplained neonatal seizures occur and not later than the neonatal period to prevent poor neurodevelopmental outcomes.

Designing next-generation subscalp devices for seizure monitoring: A systematic review and meta-analysis of established extracranial hardware.

Implantable brain recording and stimulation devices apply to a broad spectrum of conditions, such as epilepsy, movement disorders and depression. For long-term monitoring and neuromodulation in epilepsy patients, future extracranial subscalp implants may offer a promising, less-invasive alternative to intracranial neurotechnologies. To inform the design and assess the safety profile of such next-generation devices, we estimated extracranial complication rates of deep brain stimulation (DBS), cranial peripheral nerve stimulation (PNS), responsive neurostimulation (RNS) and existing subscalp EEG devices (sqEEG), as proxy for future implants. Pubmed was searched systematically for DBS, PNS, RNS and sqEEG studies from 2000 to February 2024 (48 publications, 7329 patients). We identified seven categories of extracranial adverse events: infection, non-infectious cutaneous complications, lead migration, lead fracture, hardware malfunction, pain and hemato-seroma. We used cohort sizes, demographics and industry funding as metrics to assess risks of bias. An inverse variance heterogeneity model was used for pooled and subgroup meta-analysis. The pooled incidence of extracranial complications reached 14.0%, with infections (4.6%, CI 95% [3.2 - 6.2]), surgical site pain (3.2%, [0.6 - 6.4]) and lead migration (2.6%, [1.0 - 4.4]) as leading causes. Subgroup analysis showed a particularly high incidence of persisting pain following PNS (12.0%, [6.8 - 17.9]) and sqEEG (23.9%, [12.7 - 37.2]) implantation. High rates of lead migration (12.4%, [6.4 - 19.3]) were also identified in the PNS subgroup. Complication analysis of DBS, PNS, RNS and sqEEG studies provides a significant opportunity to optimize the safety profile of future implantable subscalp devices for chronic EEG monitoring. Developing such promising technologies must address the risks of infection, surgical site pain, lead migration and skin erosion. A thin and robust design, coupled to a lead-anchoring system, shall enhance the durability and utility of next-generation subscalp implants for long-term EEG monitoring and neuromodulation.

Predictors of follow-up care for critically-ill patients with seizures and epileptiform abnormalities on EEG monitoring.

OBJECTIVE: Post-hospitalization follow-up visits are crucial for preventing long-term complications. Patients with electrographic epileptiform abnormalities (EA) including seizures and periodic and rhythmic patterns are especially in need of follow-up for long-term seizure risk stratification and medication management. We sought to identify predictors of follow-up. METHODS: This is a retrospective cohort study of all patients (age ≥ 18 years) admitted to intensive care units that underwent continuous EEG (cEEG) monitoring at a single center between 01/2016-12/2019. Patients with EAs were included. Clinical and demographic variables were recorded. Follow-up status was determined using visit records 6-month post discharge, and visits were stratified as outpatient follow-up, neurology follow-up, and inpatient readmission. Lasso feature selection analysis was performed. RESULTS: 723 patients (53 % female, mean (std) age of 62.3 (16.4) years) were identified from cEEG records with 575 (79 %) surviving to discharge. Of those discharged, 450 (78 %) had outpatient follow-up, 316 (55 %) had a neurology follow-up, and 288 (50 %) were readmitted during the 6-month period. Discharge on antiseizure medications (ASM), younger age, admission to neurosurgery, and proximity to the hospital were predictors of neurology follow-up visits. Discharge on ASMs, along with longer length of stay, younger age, emergency admissions, and higher illness severity were predictors of readmission. SIGNIFICANCE: ASMs at discharge, demographics (age, address), hospital care teams, and illness severity determine probability of follow-up. Parameters identified in this study may help healthcare systems develop interventions to improve care transitions for critically-ill patients with seizures and other EA.

A hybrid 1D CNN-BiLSTM model for epileptic seizure detection using multichannel EEG feature fusion.

Epilepsy, a chronic non-communicable disease is characterized by repeated unprovoked seizures, which are transient episodes of abnormal electrical activity in the brain. While Electroencephalography (EEG) is considered as the gold standard for diagnosis in current clinical practice, manual inspection of EEG is time consuming and biased. This paper presents a novel hybrid 1D CNN-Bi LSTM feature fusion model for automatically detecting seizures. The proposed model leverages spatial features extracted by one dimensional convolutional neural network and temporal features extracted by bi directional long short-term memory network. Ictal and inter ictal data is first acquired from the long multichannel EEG record. The acquired data is segmented and labelled using small fixed windows. Signal features are then extracted from the segments concurrently by the parallel combination of CNN and Bi-LSTM. The spatial and temporal features thus captured are then fused to enhance classification accuracy of model. The approach is validated using benchmark CHB-MIT dataset and 5-fold cross validation which resulted in an average accuracy of 95.90%, with precision 94.78%, F1 score 95.95%. Notably model achieved average sensitivity of 97.18% with false positivity rate at 0.05/hr. The significantly lower false positivity and false negativity rates indicate that the proposed model is a promising tool for detecting seizures in epilepsy patients. The employed parallel path network benefits from memory function of Bi-LSTM and strong feature extraction capabilities of CNN. Moreover, eliminating the need for any domain transformation or additional preprocessing steps, model effectively reduces complexity and enhances efficiency, making it suitable for use by clinicians during the epilepsy diagnostic process.

Concept-drifts adaptation for machine learning EEG epilepsy seizure prediction.

Seizure prediction remains a challenge, with approximately 30% of patients unresponsive to conventional treatments. Addressing this issue is crucial for improving patients' quality of life, as timely intervention can mitigate the impact of seizures. In this research field, it is critical to identify the preictal interval, the transition from regular brain activity to a seizure. While previous studies have explored various Electroencephalogram (EEG) based methodologies for prediction, few have been clinically applicable. Recent studies have underlined the dynamic nature of EEG data, characterised by data changes with time, known as concept drifts, highlighting the need for automated methods to detect and adapt to these changes. In this study, we investigate the effectiveness of automatic concept drift adaptation methods in seizure prediction. Three patient-specific seizure prediction approaches with a 10-minute prediction horizon are compared: a seizure prediction algorithm incorporating a window adjustment method by optimising performance with Support Vector Machines (Backwards-Landmark Window), a seizure prediction algorithm incorporating a data-batch (seizures) selection method using a logistic regression (Seizure-batch Regression), and a seizure prediction algorithm with a dynamic integration of classifiers (Dynamic Weighted Ensemble). These methods incorporate a retraining process after each seizure and use a combination of univariate linear features and SVM classifiers. The Firing Power was used as a post-processing technique to generate alarms before seizures. These methodologies were compared with a control approach based on the typical machine learning pipeline, considering a group of 37 patients with Temporal Lobe Epilepsy from the EPILEPSIAE database. The best-performing approach (Backwards-Landmark Window) achieved results of 0.75 ± 0.33 for sensitivity and 1.03 ± 1.00 for false positive rate per hour. This new strategy performed above chance for 89% of patients with the surrogate predictor, whereas the control approach only validated 46%.

Development of a machine learning model and nomogram to predict seizures in children with COVID-19: a two-center study.

OBJECTIVE: This study aimed to use machine learning to evaluate the risk factors of seizures and develop a model and nomogram to predict seizures in children with coronavirus disease 2019 (COVID-19). MATERIAL AND METHODS: A total of 519 children with COVID-19 were assessed to develop predictive models using machine learning algorithms, including extreme gradient boosting (XGBoost), random forest (RF) and logistic regression (LR). The performance of the models was assessed using area under the receiver operating characteristic curve (AUC) values. Importance matrix plot and SHapley Additive exPlanations (SHAP) values were calculated to evaluate feature importance and to show the visualization results. The nomogram and clinical impact curve were used to validate the final model. RESULTS: Two hundred and seventeen children with COVID-19 had seizures. According to the AUC, the RF model performed the best. Based on the SHAP values, the top three most important variables in the RF model were neutrophil percentage, cough and fever duration. The nomogram and clinical impact curve also verified that the RF model possessed significant predictive value. CONCLUSIONS: Our research indicates that the RF model demonstrates excellent performance in predicting seizures, and our novel nomogram can facilitate clinical decision-making and potentially offer benefit for clinicians to prevent and treat seizures in children with COVID-19.

[Long-term follow-up of adult patients with serial and status course of epileptic seizures]. / Dolgosrochnyi katamnez vzroslykh patsientov s epilepsiei s epilepticheskim statusom ili seriinymi pristupami.

OBJECTIVE: To study the follow-up of adult patients with status epilepticus or a history of serial seizures, assessing the likelihood of achieving long-term remission and identifying predictors of treatment effectiveness. MATERIAL AND METHODS: The study included 280 patients divided into 137 patients with epilepsy with a series of seizures or a history of status epilepticus (group 1) and 143 patients, who had not previously received therapy and did not have a series of seizures or a history of status epilepticus (group 2). A clinical and neurological examination, analysis of medical documentation data, electroencephalography, and MRI were performed. RESULTS: After correction of therapy, remission in patients in group 1 was achieved in 21.9%, improvement in 30%, no effect was observed in 48.1%; in group 2 the indicators were 51%, 28.7%, 20.3%, respectively. The onset of epilepsy in childhood, frequent seizures, and regional epileptiform activity were associated with the lack of treatment effect. CONCLUSION: The results confirm the main role of the clinical examination in determining the prognosis of epilepsy in a particular patient. Currently available instrumental techniques have limited predictive value.

Method for cycle detection in sparse, irregularly sampled, long-term neuro-behavioral timeseries: Basis pursuit denoising with polynomial detrending of long-term, inter-ictal epileptiform activity.

Numerous physiological processes are cyclical, but sampling these processes densely enough to perform frequency decomposition and subsequent analyses can be challenging. Mathematical approaches for decomposition and reconstruction of sparsely and irregularly sampled signals are well established but have been under-utilized in physiological applications. We developed a basis pursuit denoising with polynomial detrending (BPWP) model that recovers oscillations and trends from sparse and irregularly sampled timeseries. We validated this model on a unique dataset of long-term inter-ictal epileptiform discharge (IED) rates from human hippocampus recorded with a novel investigational device with continuous local field potential sensing. IED rates have well established circadian and multiday cycles related to sleep, wakefulness, and seizure clusters. Given sparse and irregular samples of IED rates from multi-month intracranial EEG recordings from ambulatory humans, we used BPWP to compute narrowband spectral power and polynomial trend coefficients and identify IED rate cycles in three subjects. In select cases, we propose that random and irregular sampling may be leveraged for frequency decomposition of physiological signals. Trial Registration: NCT03946618.