Seizure Onset Patterns Research Articles

Time domain based seizure onset analysis of brain signatures in pediatric EEG

A comprehensive insight into the epileptiform discharges at the time of seizure onset can aid neurophysiologists in the diagnosis and treatment of epileptic seizures. Visual analysis of seizure onset patterns is often a complex and tedious task. These problems suggest the development of automated seizure onset detection systems. The present research work is oriented for automatic detection of epileptic seizures at the onset using statistical measures. A quadratic classifier with fourfold cross-validation is used to demarcate the seizure and non-seizure activity. The algorithm is evaluated for 24 patients from the CHB MIT scalp EEG database. Classifier performance is assessed in terms of sensitivity, specificity, accuracy, and latency.

Multimodal in vivo recording using transparent graphene microelectrodes illuminates spatiotemporal seizure dynamics at the microscale

Neurological disorders such as epilepsy arise from disrupted brain networks. Our capacity to treat these disorders is limited by our inability to map these networks at sufficient temporal and spatial scales to target interventions. Current best techniques either sample broad areas at low temporal resolution (e.g. calcium imaging) or record from discrete regions at high temporal resolution (e.g. electrophysiology). This limitation hampers our ability to understand and intervene in aberrations of network dynamics. Here we present a technique to map the onset and spatiotemporal spread of acute epileptic seizures in vivo by simultaneously recording high bandwidth microelectrocorticography and calcium fluorescence using transparent graphene microelectrode arrays. We integrate dynamic data features from both modalities using non-negative matrix factorization to identify sequential spatiotemporal patterns of seizure onset and evolution, revealing how the temporal progression of ictal electrophysiology is linked to the spatial evolution of the recruited seizure core. This integrated analysis of multimodal data reveals otherwise hidden state transitions in the spatial and temporal progression of acute seizures. The techniques demonstrated here may enable future targeted therapeutic interventions and novel spatially embedded models of local circuit dynamics during seizure onset and evolution.

Extracting seizure onset from surface EEG with independent component analysis: Insights from simultaneous scalp and intracerebral EEG

The success of stereoelectroencephalographic (SEEG) investigations depends crucially on the hypotheses on the putative location of the seizure onset zone. This information is derived from non-invasive data either based on visual analysis or advanced source localization algorithms. While source localization applied to interictal spikes recorded on scalp is the classical method, it does not provide unequivocal information regarding the seizure onset zone. Raw ictal activity contains a mixture of signals originating from several regions of the brain as well as EMG artifacts, hampering direct input to the source localization algorithms. We therefore introduce a methodology that disentangles the various sources contributing to the scalp ictal activity using independent component analysis and uses equivalent current dipole localization as putative locus of ictal sources. We validated the results of our analysis pipeline by performing long-term simultaneous scalp – intracerebral (SEEG) recordings in 14 patients and analyzing the wavelet coherence between the independent component encoding the ictal discharge and the SEEG signals in 8 patients passing the inclusion criteria. Our results show that invasively recorded ictal onset patterns, including low-voltage fast activity, can be captured by the independent component analysis of scalp EEG. The visibility of the ictal activity strongly depends on the depth of the sources. The equivalent current dipole localization can point to the seizure onset zone (SOZ) with an accuracy that can be as high as 10 mm for superficially located sources, that gradually decreases for deeper seizure generators, averaging at 47 mm in the 8 analyzed patients. Independent component analysis is therefore shown to have a promising SOZ localizing value, indicating whether the seizure onset zone is neocortical, and its approximate location, or located in mesial structures. That may contribute to a better crafting of the hypotheses used as basis of the stereo-EEG implantations.

Detecting High Frequency Oscillations for Stereoelectroencephalography in Epilepsy via Hypergraph Learning.

Successful epilepsy surgeries depend highly on pre-operative localization of epileptogenic zones. Stereoelectroencephalography (SEEG) records interictal and ictal activities of the epilepsy in order to precisely find and localize epileptogenic zones in clinical practice. While it is difficult to find distinct ictal onset patterns generated the seizure onset zone from SEEG recordings in a confined region, high frequency oscillations are commonly considered as putative biomarkers for the identification of epileptogenic zones. Therefore, automatic and accurate detection of high frequency oscillations in SEEG signals is crucial for timely clinical evaluation. This work formulates the detection of high frequency oscillations as a signal segment classification problem and develops a hypergraph-based detector to automatically detect high frequency oscillations such that human experts can visually review SEEG signals. We evaluated our method on 4,000 signal segments from clinical SEEG recordings that contain both ictal and interictal data obtained from 19 patients who suffer from refractory focal epilepsy. The experimental results demonstrate the effectiveness of the proposed detector that can successfully localize interictal high frequency oscillations and outperforms multiple peer machine learning methods. In particular, the proposed detector achieved 90.7% in accuracy, 80.9% in sensitivity, and 96.9% in specificity.

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.

Do Spike Domain Analysis Interictally Correlate With the Ictal Patterns in Temporal Lobe Epilepsy?

To study if one can conceptualize the scalp ictal onset pattern through analysis of interictal spike domain analysis in temporal lobe epilepsy (TLE). Seventy-four patients with unilateral mesial temporal sclerosis (MTS) were categorized into "type A" interictal epileptiform discharges (IEDs) with negativity over infero-lateral scalp electrodes over temporal region and contralateral central region showing positivity; all IEDs other than type A were categorized as type B. The ictal electrographic patterns was termed "focal" when confined to side of MTS, was "regional" when lateralized to the ipsilateral hemisphere; "diffuse" if nonlateralized/localized; and ictal onset contralateral to MTS termed as "discordant." A total of 377 seizures and 5,476 spikes were studied. These were divided into four types: (1) type A IEDs ipsilateral to MTS (44 patients), (2) type A IEDs bitemporally (16 patients), (3) type A IEDs contralaterally (7 patients) and type B IEDs ipsilaterally, and (4) bilateral type B IEDs (7 patients). The ictal pattern was either focal or regional in 51 of 60 patients (85%) with type A IEDs; it was "diffuse" in 9 patients (15%). Diffuse ictal onset was seen in 12 of 14 (86%) with either ipsilateral/bitemporal type B IEDs. Ictal onset on the opposite hemisphere was noted in 2 (14%). Type A IEDs signify a focal ictal onset and type B IEDs suggest a diffuse ictal onset in patients with MTS on one side. Interictal spike domain analysis helps predicting ictal patterns in temporal lobe epilepsy.

The Ictal Signature of Thalamus and Basal Ganglia in Focal Epilepsy: A SEEG Study.

To determine the involvement of subcortical regions in human epilepsy by analyzing direct recordings from these regions during epileptic seizures using stereo-EEG (SEEG). We studied the SEEG recordings of a large series of patients (74 patients, 157 seizures) with an electrode sampling the thalamus and in some cases also the basal ganglia (caudate nucleus, 22 patients; and putamen, 4 patients). We applied visual analysis and signal quantification methods (Epileptogenicity Index [EI]) to their ictal recordings and compared electrophysiologic with clinical data. We found that in 86% of patients, thalamus was involved during seizures (visual analysis) and 20% showed high values of epileptogenicity (EI >0.3). Basal ganglia may also disclose high values of epileptogenicity (9% in caudate nucleus) but to a lesser degree than thalamus (p < 0.01). We observed different seizure onset patterns including low voltage high frequency activities. We found high values of thalamic epileptogenicity in different epilepsy localizations, including opercular and motor epilepsies. We found no difference between epilepsy etiologies (cryptogenic vs malformation of cortical development, p = 0.77). Thalamic epileptogenicity was correlated with the extension of epileptogenic networks (p = 0.02, ρ 0.32). We found a significant effect (p < 0.05) of thalamic epileptogenicity regarding the postsurgical outcome (higher thalamic EI corresponding to higher probability of surgical failure). Thalamic involvement during seizures is common in different seizure types. The degree of thalamic epileptogenicity is a possible marker of the epileptogenic network extension and of postsurgical prognosis.

Domino-like transient dynamics at seizure onset in epilepsy.

The International League Against Epilepsy (ILAE) groups seizures into “focal”, “generalized” and “unknown” based on whether the seizure onset is confined to a brain region in one hemisphere, arises in several brain region simultaneously, or is not known, respectively. This separation fails to account for the rich diversity of clinically and experimentally observed spatiotemporal patterns of seizure onset and even less so for the properties of the brain networks generating them. We consider three different patterns of domino-like seizure onset in Idiopathic Generalized Epilepsy (IGE) and present a novel approach to classification of seizures. To understand how these patterns are generated on networks requires understanding of the relationship between intrinsic node dynamics and coupling between nodes in the presence of noise, which currently is unknown. We investigate this interplay here in the framework of domino-like recruitment across a network. In particular, we use a phenomenological model of seizure onset with heterogeneous coupling and node properties, and show that in combination they generate a range of domino-like onset patterns observed in the IGE seizures. We further explore the individual contribution of heterogeneous node dynamics and coupling by interpreting in-vitro experimental data in which the speed of onset can be chemically modulated. This work contributes to a better understanding of possible drivers for the spatiotemporal patterns observed at seizure onset and may ultimately contribute to a more personalized approach to classification of seizure types in clinical practice.

Prediction of seizure freedom after epilepsy surgery – Critical reappraisal of significance of intracranial EEG parameters

ObjectiveTo analyze the significance of intracranial electroencephalography (iEEG) parameters such as seizure onset patterns (SOP) and size of seizure onset zone (SOZ) with respect to prediction of seizure freedom after resective epilepsy surgery. MethodsAll patients who underwent iEEG with subdural electrodes between January 2006 and December 2015 in our epilepsy-center were included. Various iEEG parameters were retrospectively analyzed regarding their predictive value to post-operative seizure freedom. Furthermore, associations of specific SOPs with underlying histopathology and brain regions of the SOZ were examined. ResultsEighty-one patients (34 female) with 324 seizures were assessed. Low-voltage fast activity (37%) and sharp activity <13 Hz (30%) were the most frequent SOPs. Focal SOZ (≤2 cm) was the only iEEG parameter independently associated with 1-year post-operative seizure freedom (OR 4.1, 95% CI 1.433–11.679). While no SOP was linked to specific histopathologies, some associations between SOPs and anatomical regions of SOZ were found. ConclusionsA circumscribed SOZ, but no specific SOP was predictive for seizure freedom after epilepsy surgery. SignificanceIntracranial EEG may be helpful to predict post-operative seizure freedom. Multicenter studies with larger numbers of patients are required to reliably assess the significance of specific SOPs for successful resective epilepsy surgery.

Brain-responsive corticothalamic stimulation in the centromedian nucleus for the treatment of regional neocortical epilepsy

ObjectiveThe aim of the study was to determine if corticothalamic responsive stimulation targeting the centromedian nucleus of the thalamus (CMT) is a potential treatment for neocortical epilepsies with regional onsets. MethodsWe assessed efficacy and safety of CMT and neocortical responsive stimulation, detection, and stimulation programming, methods for implantation, and location and patterns of electrographic seizure onset and spread in 7 patients with medically intractable focal seizures with a regional neocortical onset. ResultsThe median follow-up duration was 17 months (average: 17 months, range: 8–28 months). The median % reduction in disabling seizures (excludes auras) in the 7 patients was 88% (mean: 80%, range: 55–100%). The median % reduction in all seizure types (disabling + auras) was 73% (mean: 67%, range: 15–94%). There were no adverse events related to implantation of the responsive neurostimulator and leads or related to the delivery of responsive stimulation. Stimulation-related contralateral paresthesias were addressed by adjusting stimulation parameters in the clinic during stimulation testing.Electrographic seizures were detected in the CMT and neocortex in all seven patients. Four patients had simultaneous or near simultaneous seizure onsets in the neocortex and CMT and three had onsets in the neocortex with spread to the CMT. ConclusionIn this small series of patients with medically intractable focal seizures and regional neocortical onset, responsive neurostimulation to the neocortex and CMT improved seizure control and was well tolerated. SignificanceResponsive corticothalamic neurostimulation of the CMT and neocortex is a potential treatment for patients with regional neocortical epilepsies.

Stereoelectroencephalography (SEEG) and epilepsy surgery in posttraumatic epilepsy: A multicenter retrospective study

PurposePosttraumatic epilepsy (PTE) is a common cause of drug-resistant epilepsy, especially in young adults. Nevertheless, such patients are not common candidates for intracranial presurgical evaluation. We investigated the role of stereoelectroencephalography (SEEG) in defining epileptogenicity and surgical strategy in patients with PTE. MethodsWe analyzed ictal SEEG recordings from 18 patients. We determined the seizure onset zone (SOZ) by quantifying the epileptogenicity of the sampled structures, using the “epileptogenicity index” (EI). We also identified seizure onset patterns (SOPs) through visual and frequency analysis. Postsurgical outcome was assessed by Engel's classification. ResultsThe SOZ in PTE was most often located in temporal lobes, followed by frontal lobes. The SOZ was network-organized in the majority of the cases. Half of the SOP did not contain fast discharges. Half of the recordings showed SOZ that were less extensive than the posttraumatic lesions seen on brain magnetic resonance imaging (MRI). All but one operated patient benefited from tailored cortectomy. Only 3 patients were contraindicated for surgical resection due to bilateral epileptogenicity. The overall surgical outcome was good in majority of patients (67% Engel I). ConclusionDespite the potential risk of bilateral or multifocal epilepsy, patients with PTE may benefit from presurgical assessment in well-selected cases. In this context, SEEG allows guidance of tailored resections adapted to the SOZ.

Mapping the Insula with Stereo-Electroencephalography: The Emergence of Semiology in Insula Lobe Seizures.

Insula epilepsy is rare and can be evaluated effectively by Stereotactic intracerebral EEG (SEEG). Many previous studies of insulo-opercular seizures have been unable to separate insular and opercular onset. With adequate sampling of the insula, this study shows this is possible. We analyzed intrainsular dynamics and extrainsular propagation in 12 patients with "pure" insula epilepsy (n = 9) or insular and only deepest opercular involvement (n = 3) at seizure onset. Review of semiology defined clinical groups, agglomerative cluster, and principal component analysis of semiological features were performed. Quantitative epileptogenicity, and intrainsular and extrainsular propagation were computed via time frequency analysis and epileptogenicity mapping. Seizure onset patterns were heterogeneous; the seizure onset zone was focal. Seizure onset and first ictal change within insula functional subdivision correlated with aura and reflex component. No paninsular spread occurred; contralateral insular spread was very early. While the discharge was intrainsular, clinical signs related to aura or vegetative signs. Extrainsular propagation was early and related to the emergence of the majority of clinical signs. Cluster analysis found an anterior, intermediate, and posterior insula seizure onset group. The largest principal component separated anterior insula manifestations, including early hypermotor signs, early recovery, and no aura from posterior insula features of early dystonia, early tonic motor features, and sensorimotor aura. Aura is vital to identifying seizure onset and relates to insula functional subdivision. Seizures are heterogenous; extrainsular propagation occurs early, accounting for most of the semiology. With adequate sampling, "pure" insula epilepsy can be identified and focal curative resection is possible. ANN NEUROL 2020;88:477-488.

Stereotactic MRI-guided laser interstitial thermal therapy for extratemporal lobe epilepsy.

Magnetic resonance imaging (MRI)-guided laser interstitial thermal therapy (MRg-LITT) is an alternative to open epilepsy surgery. We assess safety and effectiveness of MRg-LITT for extratemporal lobe epilepsy (ETLE) in patients who are considered less favorable for open resection. We retrospectively reviewed sequential cases of patients with focal ETLE who underwent MRg-LITT between 2012 and 2019. Epileptogenic zones were determined from standard clinical and imaging data±stereoelectroencephalography (SEEG). Standard stereotactic techniques, MRI thermometry, and a commercial laser thermal therapy system were used for ablations. Anatomic MRI was used to calculate ablation volumes. Clinical outcomes were determined longitudinally. Thirty-five patients with mean epilepsy duration of 21.3±12.2years underwent MRg-LITT for focal ETLE at a mean age 36.4±12.7years. A mean 2.59±1.45 trajectories per patient were used to obtain ablation volumes of 8.8±7.5cm3 . Mean follow-up was 27.3±19.5months. Of 32 patients with >12months of follow-up, 17 (53%) achieved good outcomes (Engel class I+II) of whom 14 (44%) were Engel class I. Subgroup analysis revealed better outcomes for patients with lesional ETLE than for those who were nonlesional, multifocal, or who had failed prior interventions (P=.02). Of 13 patients showing favorable seizure-onset patterns (localized low voltage fast activity or rhythmic spiking on SEEG) prior to ablation, 9 (69%) achieved good outcomes, whereas only 3 of 11 (27%) who show other slower onset patterns achieved good outcomes. Minor adverse events included six patients with transient sensorimotor neurologic deficits and four patients with asymptomatic hemorrhages along the fiber tract. Major adverse events included one patient with a brain abscess that required stereotactic drainage and one patient with persistent hypothalamic obesity. Three deaths-two seizure-associated and one suicide-were unrelated to surgical procedures. MRI-guided laser interstitial thermal therapy (or MRg-LITT) was well-tolerated and yielded good outcomes in a heterogeneous group of ETLE patients. Lesional epilepsy and favorable seizure-onset patterns on SEEG predicted higher likelihoods of success.

Ictal stereo-electroencephalography onset patterns of mesial temporal lobe epilepsy and their clinical implications

ObjectiveThe differences in mesial temporal epilepsy (MTE) stereo-electroencephalography (SEEG) seizure-onset patterns and their clinical implications remains unclear. MethodsWe analyzed consecutive patients with MTE undergoing non-invasive workup, SEEG evaluation and resective surgery. Cases were classified into either mesial temporal sclerosis (MTS) group or non-MTS group based on magnetic resonance imaging (MRI). Seizure-onset patterns of SEEG were classified to analyze their correlation with surgical outcome and clinical subtypes. ResultsTwenty-eight patients were studied. Twenty (71.4%) patients had Engel I outcome. Thirteen patients had one seizure-onset pattern, 15 had two or more patterns. Five patterns of seizure-onset were identified and seizure-onset zones differed significantly across the 5 patterns. No difference was observed in surgical outcome between patients with single or multiple seizure-onset patterns. Periodic spike-onset pattern was associated with MTS (P = 0.003) while burst-onset was associated with non-MTS lesions (P = 0.003). Patients with seizure-onsets outside the resected temporal lobe (multiple onsets) had poorer prognosis (P = 0.0046). ConclusionWe identified 5 distinct onset patterns of MTE and correlated two of them with MRI findings. Multiple seizure-onset patterns in MTE may not necessarily suggest poor outcome. Patients with multi-focal seizure-onsets including seizures originating outside the resected temporal lobe have poorer outcome. SignificanceThis study identifies distinct onset patterns of MTE and their clinical implications.

The “Connectivity Epileptogenicity Index ” (cEI), a method for mapping the different seizure onset patterns in StereoElectroEncephalography recorded seizures

ObjectiveLocalization of epileptogenic brain regions is a crucial aim of pre-surgical evaluation of patients with drug-resistant epilepsy. Several methods have been proposed to identify the seizure onset zone, particularly based on the detection of fast activity. Most of these methods are inefficient to detect slower patterns of onset that account for 20–30% of commonly observed Stereo-Electro-Encephalography (SEEG) patterns. We seek to evaluate the performance of a new quantified measure called the Connectivity Epileptogenicity Index (cEI) in various types of seizure onset patterns. MethodsWe studied SEEG recorded seizures from 51 patients, suffering from focal drug-resistant epilepsy. The cEI combines a directed connectivity measure (“out-degrees”) and the original epileptogenicity index (EI). Quantified results (Out-degrees, cEI and EI) were compared to visually defined seizure onset zone (vSOZ). We computed recall (sensitivity) and precision (proportion of correct detections within all detections) with vSOZ as a reference. The quality of the detector was quantified by the area under the precision-recall curve. ResultsBest results (in terms of match with vSOZ) were obtained for cEI. For seizures with fast onset patterns, cEI and EI gave comparable results. For seizures with slow onset patterns, cEI gave a better estimation of the vSOZ than EI. ConclusionsWe observed that cEI discloses better performance than EI when seizures starts with slower patterns and equal to EI in seizures with fast onset patterns. SignificanceThe cEI is a promising new tool for epileptologists, that helps characterizing the seizure onset zone in sEEG, in a robust way despite variations in seizure onset patterns.

Perisylvian vulnerability to postencephalitic epilepsy

ObjectivePostencephalitic epilepsy is often resistant to antiseizure medications, leading to evaluation for epilepsy surgery. Characterizing its localization carries implications for optimal surgical approach. We aimed to determine whether a prior history of encephalitis is associated with specific epileptogenic networks among patients with drug resistant epilepsy undergoing stereotactic EEG (SEEG). MethodsWe conducted a retrospective cohort study of drug resistant epilepsy, with and without a prior history of encephalitis. We analyzed SEEG recordings to identify patterns of seizure onset and organization. Seventeen patients with a history of encephalitis (of infectious etiology in two subjects) were identified from a database of patients undergoing SEEG and were compared to seventeen drug-resistant epilepsy controls without a history of encephalitis matched for confounding variables including pre-implantation hypotheses, epilepsy duration, age, and sex. ResultsIndependent bilateral seizures were noted in 65% of the postencephalitic epilepsy cohort. We identified four SEEG-ictal patterns in patients with a prior history of encephalitis: (1) anteromesial temporal onset (24%), (2) anteromesial temporal onset with early spread to the perisylvian region (29%), (3) perisylvian (59%) and (4) synchronized anteromesial temporal and perisylvian (29%) onsets. Patterns 3 and 4, with perisylvian involvement at onset, were unique to the encephalitis group (p = 0.0003 and 0.04 respectively) and exhibited a “patchwork” organization. None of the encephalitis patients vs 5/7 matched controls had Engel I outcome (p = 0.0048). ConclusionsPostencephalitic epilepsies involve anteromesial temporal and perisylvian networks, often in a bilateral independent manner. Unique ictal patterns involving the perisylvian regions was identified in the encephalitis group, but not in the matched control group. Significance: These findings may reflect a selective vulnerability of the perisylvian regions to epilepsy resulting from encephalitis, significantly mitigating the chances of success with SEEG-guided temporal resections.

Dynamic network properties of the interictal brain determine whether seizures appear focal or generalised

Current explanatory concepts suggest seizures emerge from ongoing dynamics of brain networks. It is unclear how brain network properties determine focal or generalised seizure onset, or how network properties can be described in a clinically-useful manner. Understanding network properties would cast light on seizure-generating mechanisms and allow to quantify to which extent a seizure is focal or generalised. Functional brain networks were estimated in segments of scalp-EEG without interictal discharges (68 people with epilepsy, 38 controls). Simplified brain dynamics were simulated using a computer model. We introduce: Critical Coupling (Cc), the ability of a network to generate seizures; Onset Index (OI), the tendency of a region to generate seizures; and Participation Index (PI), the tendency of a region to become involved in seizures. Cc was lower in both patient groups compared with controls. OI and PI were more variable in focal-onset than generalised-onset cases. In focal cases, the regions with highest OI and PI corresponded to the side of seizure onset. Properties of interictal functional networks from scalp EEG can be estimated using a computer model and used to predict seizure likelihood and onset patterns. This may offer potential to enhance diagnosis through quantification of seizure type using inter-ictal recordings.

Comparison of Responsive Neurostimulation System Efficacy Between Different Electrographic Seizure Onset Patterns (1255)

Comparison of Responsive Neurostimulation System Efficacy Between Different Electrographic Seizure Onset Patterns (1255)

A model for focal seizure onset, propagation, evolution, and progression

We developed a neural network model that can account for major elements common to human focal seizures. These include the tonic-clonic transition, slow advance of clinical semiology and corresponding seizure territory expansion, widespread EEG synchronization, and slowing of the ictal rhythm as the seizure approaches termination. These were reproduced by incorporating usage-dependent exhaustion of inhibition in an adaptive neural network that receives global feedback inhibition in addition to local recurrent projections. Our model proposes mechanisms that may underline common EEG seizure onset patterns and status epilepticus, and postulates a role for synaptic plasticity in the emergence of epileptic foci. Complex patterns of seizure activity and bi-stable seizure end-points arise when stochastic noise is included. With the rapid advancement of clinical and experimental tools, we believe that this model can provide a roadmap and potentially an in silico testbed for future explorations of seizure mechanisms and clinical therapies.

Automated classification of five seizure onset patterns from intracranial electroencephalogram signals

ObjectiveThe electroencephalographic (EEG) signals contain information about seizures and their onset location. There are several seizure onset patterns reported in the literature, and these patterns have clinical significance. In this work, we propose a system to automatically classify five seizure onset patterns from intracerebral EEG signals. MethodsThe EEG was segmented by clinicians indicating the start and end time of each seizure onset pattern, the channels involved at onset and the seizure onset pattern. Twelve features that represent the time domain characteristics and signal complexity were extracted from 663 seizures channels of 24 patients. The features were used for classification of the patterns with support vector machine - Error-Correcting Output Codes (SVM-ECOC). Three patient groups with a similar number of seizure segments were created, and one group was used for testing and the rest for training. This test was repeated by rotating the testing and training data. ResultsThe feature space formed by both time domain and multiscale sample entropy features perform well in classification of the data. An overall accuracy of 80.7% was obtained with these features and a linear kernel of SVM-ECOC. ConclusionsThe seizure onset patterns consist of varied time and complexity characteristics. It is possible to automatically classify various seizure onset patterns very similarly to visual classification. SignificanceThe proposed system could aid the medical team in assessing intracerebral EEG by providing an objective classification of seizure onset patterns.