Online Seizure Research Articles

Detection of focal to bilateral tonic-clonic seizures using a connected shirt.

This study was undertaken to develop and evaluate a machine learning-based algorithm for the detection of focal to bilateral tonic-clonic seizures (FBTCS) using a novel multimodal connected shirt. We prospectively recruited patients with epilepsy admitted to our epilepsy monitoring unit and asked them to wear the connected shirt while under simultaneous video-electroencephalographic monitoring. Electrocardiographic (ECG) and accelerometric (ACC) signals recorded with the connected shirt were used for the development of the seizure detection algorithm. First, we used a sliding window to extract linear and nonlinear features from both ECG and ACC signals. Then, we trained an extreme gradient boosting algorithm (XGBoost) to detect FBTCS according to seizure onset and offset annotated by three board-certified epileptologists. Finally, we applied a postprocessing step to regularize the classification output. A patientwise nested cross-validation was implemented to evaluate the performances in terms of sensitivity, false alarm rate (FAR), time in false warning (TiW), detection latency, and receiver operating characteristic area under the curve (ROC-AUC). We recorded 66 FBTCS from 42 patients who wore the connected shirt for a total of 8067 continuous hours. The XGBoost algorithm reached a sensitivity of 84.8% (56/66 seizures), with a median FAR of .55/24 h and a median TiW of 10 s/alarm. ROC-AUC was .90 (95% confidence interval = .88-.91). Median detection latency from the time of progression to the bilateral tonic-clonic phase was 25.5 s. The novel connected shirt allowed accurate detection of FBTCS with a low false alarm rate in a hospital setting. Prospective studies in a residential setting with a real-time and online seizure detection algorithm are required to validate the performance and usability of this device.

Online seizure first aid certification program is an effective means of improving student knowledge and self efficacy surrounding epilepsy

IntroductionEpilepsy is a neurological disorder, characterized by recurring seizures, affecting more than 3.4 million adults and children throughout the United States. Still, there are sizable gaps in awareness and knowledge of this disorder, and persistent misunderstandings and stigmas surrounding epilepsy and seizure first aid (SFA) pose a risk to those living with this condition. The Epilepsy Foundation, with support from the Centers for Disease Control and Prevention (CDC), has developed programs to address the issue of insufficient public education and awareness surrounding epilepsy. One of these programs is a free, accessible online Seizure Recognition and First Aid Certification program, established in 2021. We aimed to evaluate the effectiveness of the Epilepsy Foundation’s online Seizure Recognition and First Aid Certification program in improving student knowledge of epilepsy and appropriate bystander-intervention methods to assist a person experiencing a seizure. MethodsThe Epilepsy Foundation’s online Seizure Recognition and First Aid Certification course evaluates student performance via a 16-question knowledge assessment and six question self-efficacy assessment provided both before and after completion of the course. Pre- and post-course scores of students who enrolled between December of 2021 and September of 2022 were collected. Average score improvement was evaluated via the difference in pre-course and post-course first-attempt scores. Statistical significance was evaluated using paired sample, two-tailed t-tests of pre-course and post-course scores. ResultsAverage pre-course knowledge score was 74.33% (n = 10,371, σ2 = 3.04%), post-course score was 88.04% (n = 10,371, σ2 = 0.83%), and score difference was 13.71% (p < 0.001). Average pre-course self-efficacy score was 63.44% (n = 8,046, σ2 = 4.71%), post-course score was 87.08% (n = 8,046, σ2 = 1.68%), and score difference was 23.64% (p < 0.001). ConclusionWe found a significant increase in knowledge and self-efficacy assessment scores after students completed the online Seizure Recognition and First Aid Certification course, suggesting that the program is an effective method of improving the understanding of epilepsy and bystander interventions to assist a person who is experiencing a seizure. In the future, awareness should continue to be promoted through SFA training programs and improving accessibility to such programs so that the risks associated with experiencing a seizure without receiving assistance are reduced for people living with epilepsy.

Leveraging electronic patient diaries in SUDEP risk evaluation

ObjectiveOur aim was to describe the risk factors known to be related to sudden unexpected death in epilepsy (SUDEP) that can be extracted from patients that utilizes an online seizure diary tool (SeizureTracker™). MethodWe conducted a descriptive analysis of SeizureTracker™ users across factors relevant to SUDEP risk. We also compared our app-using cohort to published SUDEP case-control studies. ResultsWe report across seven risk factors from 30,813 users of SeizureTracker™ who had a median length of time using the app of 5.69 years (range from 1 month to 15 years). We found that they are at greater risk for SUDEP than groups from published studies (p < .00001) based on the risk factor of generalized tonic-clonic seizures. SignificanceWe demonstrated that the population using the SeizureTracker™ tool can be a valuable population for expanding investigation of SUDEP risk factors and is a first step towards establishing a large sample with a method to ascertain data prospectively that might be critical to developing a SUDEP risk algorithm.

Online Prediction of Lead Seizures from iEEG Data.

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, the 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 the non-stationarity of a noisy iEEG signal. They include: (1) periodic retraining of the SVM classifier using most recent training data; (2) removing samples with noisy labels from training data; and (3) introducing a new adaptive post-processing technique for combining many predictions made for 20 s windows into a single prediction for a 4 h segment. Application of the proposed system requires only two 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). The proposed system achieves accurate prediction of lead seizures during long-term test periods, 3–16 lead seizures during a 169–364 day test period, whereas earlier studies did not differentiate between lead vs. non-lead seizures and used much shorter test periods (~few days long).

Linear and non-linear feature extraction from rat electrocorticograms for seizure detection by support vector machine.

Seizures, the main symptom of epilepsy, are provoked due to a neurological disorder that underlies the disease. The accurate detection of seizures is a crucial step in any procedure of treatment. In the present study, electrocorticogram (ECoG) signals were recorded from awake and freely moving animals implanted with cortical electrodes before and after pentylenetetrazol, the chemo-convulsant injection. ECoG signals were segmented into 4-s epochs and labeled. Twenty-four linear and non-linear features were extracted from the time and frequency domains of the ECoG signals. The extracted features either individually or in combinations were fed to an automatic support vector machine (SVM) classification system. SVM classifier was trained with 5min of ictal and non-ictal labeled ECoG signals to build the hyperplane that separates two sets of training signals. Sensitivity, specificity, and accuracy were determined for the testing dataset using the different feature combinations. It has been found that some linear features either individually or in combinations outperform non-linear features in terms of the accuracy for seizure detection. The maximum accuracy achieved by the system was 95.3% and has been obtained only after linear and non-linear features were combined. ECoG signals were classified without pre-processing or removal of artifacts to reduce the required computational time to be suitable for online implementation purposes. This may prove the detection system's robustness and supports its use in online seizure detection protocols.

Scale of the issue: Mapping the impact of the COVID-19 lockdown on pangolin trade across India

Recent studies have linked COVID-19 induced restrictions to an increase in wildlife crime, with severe yet unknown implications for severely threatened taxa like pangolins. We analyze publicly available online seizure reports involving pangolins across India before (2018–2019) and during the pandemic (March–August 2020), using a longitudinal study design to estimate how lockdowns have impacted pangolin trade. Our analysis indicates a significant increase in seizures reported during the lockdown months of March to August 2020, in comparison to the same period in 2018 and 2019. We discuss the drivers behind this spike in pangolin trade and offer potential conservation measures.

Deep learning for robust detection of interictal epileptiform discharges

Objective. Automatic detection of interictal epileptiform discharges (IEDs, short as ‘spikes’) from an epileptic brain can help predict seizure recurrence and support the diagnosis of epilepsy. Developing fast, reliable and robust detection methods for IEDs based on scalp or intracranial electroencephalogram (iEEG) may facilitate online seizure monitoring and closed-loop neurostimulation. Approach. We developed a new deep learning approach, which employs a long short-term memory network architecture (‘IEDnet’) and an auxiliary classifier generative adversarial network (AC-GAN), to train on both expert-annotated and augmented spike events from iEEG recordings of epilepsy patients. We validated our IEDnet with two real-world iEEG datasets, and compared IEDnet with the support vector machine (SVM) and random forest (RF) classifiers on their detection performances. Main results. IEDnet achieved excellent cross-validated detection performances in terms of both sensitivity and specificity, and outperformed SVM and RF. Synthetic spike samples augmented by AC-GAN further improved the detection performance. In addition, the performance of IEDnet was robust with respect to the sampling frequency and noise. Furthermore, we demonstrated the cross-institutional generalization ability of IEDnet while testing between two datasets. Significance. IEDnet achieves excellent detection performances in identifying interictal spikes. AC-GAN can produce augmented iEEG samples to improve supervised deep learning.

Integrating old and new complexity measures toward automated seizure detection from long-term video EEG recordings

SummaryAutomated seizure detection in long-term video-EEG recordings is far from being integrated into common clinical practice. Here, we leverage classical and state-of-the-art complexity measures to robustly and automatically detect seizures from scalp recordings. Brain activity is scored through eight features, encompassing traditional time domain and novel measures of recurrence. A binary classification algorithm tailored to treat unbalanced dataset is used to determine whether a time window is ictal or non-ictal from its features. The application of the algorithm on a cohort of ten adult patients with focal refractory epilepsy indicates sensitivity, specificity, and accuracy of 90%, along with a true alarm rate of 95% and less than four false alarms per day. The proposed approach emphasizes ictal patterns against noisy background without the need of data preprocessing. Finally, we benchmark our approach against previous studies on two publicly available datasets, demonstrating the good performance of our algorithm.

Optimal Selection of Customized Features for Implementing Seizure Detection in Wearable Electroencephalography Sensor

Wearable electroencephalography sensor integrated with seizure detection algorithm has become an emerging research topic in health management for patients with refractory epilepsy. To enhance seizure detection accuracy and reduce power consumption, this paper has proposed a selection method of optimal features customized for each patient. A feature pool consisting of several popular features is constructed. Then, the customized multi-domain features for each patient are selected from feature pool based on minimum redundancy maximum relevance. By removing redundant features, the computational complexity of on-line seizure detection can be greatly reduced, which is desired in wearable device where low-power and real-time operation are required. Four test datasets are used to evaluate the performance of proposed method, with detection accuracy calculated. Experiment results show that different datasets vary in optimal features. By using the proposed method, the detection accuracy and the value of area under receiver operating characteristic curve can be improved by 11.49% and 0.13 respectively, while feature dimension being reduced to 1.

Performance metrics for online seizure prediction

Many recent studies on online seizure prediction from iEEG signal describe various prediction algorithms and their prediction performance. In contrast, this paper focuses on proper specification of system parameters, such as prediction period, prediction horizon and data-driven characterization of lead seizures. Whereas prediction performance clearly depends on these system parameters many researchers simply set the values of these parameters in an ad hoc manner. Our paper investigates the effect of these system parameters on online prediction performance, using both synthetic and real-life data sets. Therefore, meaningful comparison of methods/algorithms (for online seizure prediction) should consider proper specification of system parameters.

Criminal procedure legalism and Online seizure・search for servers

이 글에서는 역외 원격지 서버의 압수·수색과 관련한 대법원의 판결에 대하여 비판적으로 검토하고 이와 관련한 국제사회의 노력에 대하여 연구하였다. 국제사회에서도 최근 급증하고 있는 테러범죄나 초국가적 조직범죄에 대응하기 위하여 역외 원격지 서버에 대한 압수·수색의 필요성을 인정하고 있는 실정이다. 그러나 한국과 같이 법적 근거 규정도 없이 법원의 해석으로 이를 실현하는 나라는 없다고 할 수 있다. 왜냐하면 수사상의 필요성에 의하여 이를 인정하더라도 국제사법공조의 문제, 피압수자의 기본권 및 개인정보의 침해문제, 각국 관할권과의 충돌문제 등이 아직도 해결되지 않은 상태로 남아 있기 때문이다. 또한 원격지 서버에 대한 압수 수색은 필연적으로 헌법에 의하여 보장되고 있는 기본권인 재산권 및 주거권을 침해할 수 밖에 없다. 특히 과학기술의 발달에 따른 새로운 유형으로 등장한 과학수사는 개인의 privacy를 침해하고 인간의 존엄과 가치를 부정할 위험을 가지고 있다. 그러므로 강체처분에 의한 개인의 자유의 제한은 기본권을 보장하고 있는 헌법규범의 정신에 비추어 엄격히 제한하지 않으면 안 된다. 따라서 우리가 지금 해야 할 일은 원격지 서버에 대한 압수·수색을 법원의 해석으로 인정할 것이 아니라, 최소한 일본의 ‘형사소송법’이나 유럽의 ‘사이버범죄방지협약’의 관련 규정을 검토하여 법률 규정을 제정하는 것이다. 이러한 법률에 담겨야 할 최소한의 내용으로는 ① 우선 최초 수색대상 컴퓨터로부터 필요한 데이터가 존재하는 원격시스템(저장매체)에 접속이 가능해야 하고, ② 최초 수색대상 컴퓨터가 원격시스템에 대한 합법적인 접근권한을 보유하고 있으며, ③ 원격시스템에 관련된 정보가 저장되어 있다고 믿을 만한 근거가 존재해야 한다. 또한 이러한 원격 수색의 경우에는 기업 등 ‘피의자 아닌 자’의 시스템을 수색하는데 활용될 가능성이 높고, 특히 앞서 원격 수색의 허용요건으로 “원격시스템에 관련된 정보가 저장되어 있다고 믿을 만한 근거의 존재”를 요구하고 있다는 점에서 ‘피의자 아닌 자’의 컴퓨터 시스템을 수색하기 위한 영장의 청구를 할 때에는 압수할 정보가 있다고 인정될 만한 자료를 제출하도록 할 필요가 있다. 나아가 수색의 일반적 요건으로 수사기관의 경우 “범죄수사에 필요한 때에는 피의자가 죄를 범하였다고 의심할 만한 정황이 있고 해당 사건과 관계가 있다고 인정할 수 있는 것에 한정하여” 수색할 수 있고, 법원의 경우 “필요한 때에는 피고사건과 관계가 있다고 인정할 수 있는 것에 한정하여” 수색할 수 있다는 점이 고려되어야 한다. 한편 원격 수색이 가능하더라도 원격 수색을 통해 발견한 전자적 증거를 압수할 수 있는 방법이 마련되어야 할 것이다.In this paper, I review critically the Supreme Court s judgment on online seizure·search for servers in national security cases and study the international efforts related to online seizure·search for servers. The international community has acknowledged the necessity of online seizure·search of servers in foreign countries in response to the recent surge in terrorism or transnational organized crime. However, there is no country that seeks online seizure·search by interpretation of the court without legal basis such as Korea. This is because, despite the recognition of the necessity of investigation, the problem of international legal cooperation, the fundamental rights of the suspects, the infringement of personal information, and the conflict with the jurisdiction of each country remain unresolved. Therefore, what we need to do now is not to accept the online seizure· search of the server as an interpretation of the court, but rather to review the relevant provisions of the Japanese ‘Criminal Procedure Act’ or the Council of Europe ‘Convention on Cybercrime ’. The minimum content that should be included in these laws is presented in this paper.

Adaptive Filtering for Epileptic Event Detection in the EEG

PurposeThe development of online seizure detection techniques as well as prediction methods are very critical. Patient quality of life could improve significantly if the beginning of a seizure could be predicted or detected early.MethodsThis paper proposes a method to automatically detect epileptic seizures based on adaptive filters and signal averaging. The process was applied to 425 h of epileptic EEG records from CHB-MIT EEG database. The developed algorithm does not require any training since it is simple and involves low processing time. Therefore, it can be implemented in real time as well as offline.ResultsThree thresholds were evaluated and calculated as 10, 20 and 30 times the median value of ST(n). The threshold of 20 showed the best relation between SEN and SPE. In this case, these indexes reached average values, across all the patients, of 90.3% and 73.7% respectively.ConclusionsThe proposed method has several strengths, for example: that no training is required due to the automatic adaptation to the threshold to each new EEG record. The algorithm could be implemented in real time. It is simple owing to its low processing time which makes it suitable for the analysis of long-term records and a large number of channels. The system could be implemented on electronic devices for warning purposes (of the seizure onset). It employs methods to process signals that were not used with epileptic seizure detection in EEG, such as in the case of adaptive predictive filters.

Towards an Online Seizure Advisory System-An Adaptive Seizure Prediction Framework Using Active Learning Heuristics.

In the last decade, seizure prediction systems have gained a lot of attention because of their enormous potential to largely improve the quality-of-life of the epileptic patients. The accuracy of the prediction algorithms to detect seizure in real-world applications is largely limited because the brain signals are inherently uncertain and affected by various factors, such as environment, age, drug intake, etc., in addition to the internal artefacts that occur during the process of recording the brain signals. To deal with such ambiguity, researchers transitionally use active learning, which selects the ambiguous data to be annotated by an expert and updates the classification model dynamically. However, selecting the particular data from a pool of large ambiguous datasets to be labelled by an expert is still a challenging problem. In this paper, we propose an active learning-based prediction framework that aims to improve the accuracy of the prediction with a minimum number of labelled data. The core technique of our framework is employing the Bernoulli-Gaussian Mixture model (BGMM) to determine the feature samples that have the most ambiguity to be annotated by an expert. By doing so, our approach facilitates expert intervention as well as increasing medical reliability. We evaluate seven different classifiers in terms of the classification time and memory required. An active learning framework built on top of the best performing classifier is evaluated in terms of required annotation effort to achieve a high level of prediction accuracy. The results show that our approach can achieve the same accuracy as a Support Vector Machine (SVM) classifier using only of the labelled data and also improve the prediction accuracy even under the noisy condition.

Illegal pangolin trade in northernmost Myanmar and its links to India and China

The northern Myanmar region has been identified as a potential transit and source place for the illegal trade of pangolins and their scales. In this study, we surveyed the trade links between Kachin State (northern Myanmar) and China and Kachin and India based on interviews, market surveys and online seizure data. From our results we cannot extrapolate that there is a link between Myanmar and India. Based on the results from interviews (17 of 38), we found that around 140–168 pangolins/year are smuggled into China via three different routes from Kachin to China. Scales are the most traded parts of pangolins in this part of Myanmar. Based on the online sources, 30 seizures of pangolin and their products were made on the Kachin–China route during 2010–2016, with all seizures made on the Chinese side of the border. We thus, recommend an increase in law enforcement on the Myanmar side, with focused effort at identifying trade hubs and deterring wholesalers. We further suggest investigating possible trade links between Kachin and other source areas. We recommend, a reclassification of the pangolins’ protection status in China from a Class II to a Class I Key Protected Species, and the prohibition of the use of pangolins’ scales for Traditional Chinese Medicine.

Erratum to: Optimal features for online seizure detection.

Erratum to: Optimal features for online seizure detection.

Tracking generalized tonic-clonic seizures with a wrist accelerometer linked to an online database

PurposeClinical management of epilepsy and current epilepsy therapy trials rely on paper or electronic diaries often with inaccurate self-reported seizure frequency as the primary outcome. This is the first study addressing the feasibility of detecting and recording generalized tonic-clonic seizures (GTCS) through a biosensor linked to an online seizure database. MethodA prospective trial was conducted with video-EEG (vEEG) in an epilepsy monitoring unit. Patients wore a wristwatch accelerometer that detected shaking and transmitted events via Bluetooth® to a bedside electronic tablet and then via Wi-Fi to an online portal. The watch recorded the date, time, audio, duration, frequency and amplitude of events. Events logged by the watch and recorded in a bedside paper diary were measured against vEEG, the “gold standard.” ResultsThirty patients were enrolled and 62 seizures were recorded on vEEG: 31 convulsive and 31 non-convulsive. Twelve patients had a total of 31 convulsive seizures, and of those, 10 patients had 13 GTCS. The watch captured 12/13 (92.3%) GTCS. Watch audio recordings were consistent with seizures in 11/12 (91.6%). Data were successfully transferred to the bedside tablet in 11/12 (91.6%), and to the online database in 10/12 (83.3%) GTCS. The watch recorded 81 false positives, of which 42/81 (51%) were cancelled by the patients. Patients and caregivers verbally reported 15/62 seizures (24.2% sensitivity) but no seizures were recorded on paper logs. ConclusionAutomatic detection and recording of GTCS to an online database is feasible and may be more informative than seizure logging in a paper diary.

Online detection of tonic-clonic seizures in pediatric patients using ECG and low-complexity incremental novelty detection.

Home monitoring of refractory epilepsy patients has become of more interest the last couple of decades. A biomedical signal that can be used for online seizure detection at home is the electrocardiogram. Previous studies have shown that tonic-clonic seizures are most often accompanied with a strong heart rate increase. The main issue however is the strong patient-specific behavior of the ictal heart rate features, which makes it hard to make a patient-independent seizure detection algorithm. A patient-specific algorithm might be a solution, but existing methods require the availability of data of several seizures, which would make them inefficient in case the first seizure only occurs after a couple of days. Therefore an online method is described here that automatically converts from a patient-independent towards a patient-specific algorithm as more patient-specific data become available. This is done by using online feedback from the users to previously given alarms. By using a simplified one-class classifier, no seizure training data needs to be available for a good performance. The method is already able to adapt to the patient-specific characteristics after a couple of hours, and is able to detect 23 of 24 seizures longer than 10s, with an average of 0.38 false alarms per hour. Due to its low-complexity, it can be easily used for wearable seizure detection at home.

Cerebellar Directed Optogenetic Intervention Inhibits Spontaneous Hippocampal Seizures in a Mouse Model of Temporal Lobe Epilepsy.

Temporal lobe epilepsy is often medically refractory and new targets for intervention are needed. We used a mouse model of temporal lobe epilepsy, on-line seizure detection, and responsive optogenetic intervention to investigate the potential for cerebellar control of spontaneous temporal lobe seizures. Cerebellar targeted intervention inhibited spontaneous temporal lobe seizures during the chronic phase of the disorder. We further report that the direction of modulation as well as the location of intervention within the cerebellum can affect the outcome of intervention. Specifically, on-demand optogenetic excitation or inhibition of parvalbumin-expressing neurons, including Purkinje cells, in the lateral or midline cerebellum results in a decrease in seizure duration. In contrast, a consistent reduction in spontaneous seizure frequency occurs uniquely with on-demand optogenetic excitation of the midline cerebellum, and was not seen with intervention directly targeting the hippocampal formation. These findings demonstrate that the cerebellum is a powerful modulator of temporal lobe epilepsy, and that intervention targeting the cerebellum as a potential therapy for epilepsy should be revisited.

Prospective multi-center study of an automatic online seizure detection system for epilepsy monitoring units

ObjectiveA method for automatic detection of epileptic seizures in long-term scalp-EEG recordings called EpiScan will be presented. EpiScan is used as alarm device to notify medical staff of epilepsy monitoring units (EMUs) in case of a seizure. MethodsA prospective multi-center study was performed in three EMUs including 205 patients. A comparison between EpiScan and the Persyst seizure detector on the prospective data will be presented. In addition, the detection results of EpiScan on retrospective EEG data of 310 patients and the public available CHB–MIT dataset will be shown. ResultsA detection sensitivity of 81% was reached for unequivocal electrographic seizures with false alarm rate of only 7perday. No statistical significant differences in the detection sensitivities could be found between the centers. The comparison to the Persyst seizure detector showed a lower false alarm rate of EpiScan but the difference was not of statistical significance. ConclusionsThe automatic seizure detection method EpiScan showed high sensitivity and low false alarm rate in a prospective multi-center study on a large number of patients. SignificanceThe application as seizure alarm device in EMUs becomes feasible and will raise the efficiency of video-EEG monitoring and the safety levels of patients.

Online Seizure Prediction Using an Adaptive Learning Approach

Epilepsy is one of the most common neurological disorders, characterized by recurrent seizures. Being able to predict impending seizures could greatly improve the lives of patients with epilepsy. In this study, we propose a new adaptive learning approach for online seizure prediction based on analysis of electroencephalogram (EEG) recordings. For each individual patient, we construct baseline patterns of normal and preseizure EEG samples, continuously monitor sliding windows of EEG recordings, and classify each window to normal or preseizure using a $(K)$-nearest-neighbor (KNN) method. A new reinforcement learning algorithm is proposed to continuously update both normal and preseizure baseline patterns based on the feedback from prediction result of each window. The proposed approach was evaluated on EEG data from 10 patients with epilepsy. For each one of the 10 patients, the adaptive approach was trained using the recordings containing the first half of seizure occurrences, and tested prospectively on the subsequent recordings. Using a 150-minute prediction horizon, our approach achieved 73 percent sensitivity and 67 percent specificity on average over 10 patients. This result is shown to be far better than those of a nonupdate prediction scheme and two native prediction schemes.