Spike detection. I. Correlation and reliability of human experts

Pitfalls in scalp EEG: Current obstacles and future directions

Although electroencephalography (EEG) serves a critical role in the evaluation and management of seizure disorders, it is commonly misinterpreted, resulting in avoidable medical, social, and financial burdens to patients and health care systems. Overinterpretation of sharply contoured transient waveforms as being representative of interictal epileptiform abnormalities lies at the core of this problem. However, the magnitude of these errors is amplified by the high prevalence of paroxysmal events exhibited in clinical practice that compel investigation with EEG. Neurology training programs, which vary considerably both in the degree of exposure to EEG and the composition of EEG didactics, have not effectively addressed this widespread issue. Implementation of competency-based curricula in lieu of traditional educational approaches may enhance proficiency in EEG interpretation amongst general neurologists in the absence of formal subspecialty training. Efforts in this regard have led to the development of a systematic, high-fidelity approach to the interpretation of epileptiform discharges that is readily employable across medical centers. Additionally, machine learning techniques hold promise for accelerating accurate and reliable EEG interpretation, particularly in settings where subspecialty interpretive EEG services are not readily available. This review highlights common diagnostic errors in EEG interpretation, limitations in current educational paradigms, and initiatives aimed at resolving these challenges.

Feasibility of high-density electric source imaging in the presurgical workflow: Effect of number of spikes and automated spike detection

OBJECTIVE

Presurgical high-density electric source imaging (hdESI) of interictal epileptic discharges (IEDs) is only used by few epilepsy centers. One obstacle is the time-consuming workflow both for recording as well as for visual review. Therefore, we analyzed the effect of (a) an automated IED detection and (b) the number of IEDs on the accuracy of hdESI and time-effectiveness.

METHODS

In 22 patients with pharmacoresistant focal epilepsy receiving epilepsy surgery (Engel 1) we retrospectively detected IEDs both visually and semi-automatically using the EEG analysis software Persyst in 256-channel EEGs. The amount of IEDs, the Euclidean distance between hdESI maximum and resection zone, and the operator time were compared. Additionally, we evaluated the intra-individual effect of IED quantity on the distance between hdESI maximum of all IEDs and hdESI maximum when only a reduced amount of IEDs were included.

RESULTS

There was no significant difference in the number of IEDs between visually versus semi-automatically marked IEDs (74 ± 56 IEDs/patient vs 116 ± 115 IEDs/patient). The detection method of the IEDs had no significant effect on the mean distances between resection zone and hdESI maximum (visual: 26.07 ± 31.12 mm vs semi-automated: 33.6 ± 34.75 mm). However, the mean time needed to review the full datasets semi-automatically was shorter by 275 ± 46 min (305 ± 72 min vs 30 ± 26 min, P < 0.001). The distance between hdESI of the full versus reduced amount of IEDs of the same patient was smaller than 1 cm when at least a mean of 33 IEDs were analyzed. There was a significantly shorter intraindividual distance between resection zone and hdESI maximum when 30 IEDs were analyzed as compared to the analysis of only 10 IEDs (P < 0.001).

SIGNIFICANCE

Semi-automatized processing and limiting the amount of IEDs analyzed (~30-40 IEDs per cluster) appear to be time-saving clinical tools to increase the practicability of hdESI in the presurgical work-up.

Moving the field forward: detection of epileptiform abnormalities on scalp electroencephalography using deep learning—clinical application perspectives

The application of deep learning approaches for the detection of interictal epileptiform discharges is a nascent field, with most studies published in the past 5 years. Although many recent models have been published demonstrating promising results, deficiencies in descriptions of data sets, unstandardized methods, variation in performance evaluation and lack of demonstrable generalizability have made it difficult for these algorithms to be compared and progress to clinical validity. A few recent publications have provided a detailed breakdown of data sets and relevant performance metrics to exemplify the potential of deep learning in epileptiform discharge detection. This review provides an overview of the field and equips computer and data scientists with a synopsis of EEG data sets, background and epileptiform variation, model evaluation parameters and an awareness of the performance metrics of high impact and interest to the trained clinical and neuroscientist EEG end user. The gold standard and inter-rater disagreements in defining epileptiform abnormalities remain a challenge in the field, and a hierarchical proposal for epileptiform discharge labelling options is recommended. Standardized descriptions of data sets and reporting metrics are a priority. Source code-sharing and accessibility to public EEG data sets will increase the rigour, quality and progress in the field and allow validation and real-world clinical translation.

A Brief Introduction to Magnetoencephalography (MEG) and Its Clinical Applications

Magnetoencephalography (MEG) plays a pivotal role in the diagnosis of brain disorders. In this review, we have investigated potential MEG applications for analysing brain disorders. The signal-to-noise ratio (SNRMEG = 2.2 db, SNREEG < 1 db) and spatial resolution (SRMEG = 2−3 mm, SREEG = 7−10 mm) is higher for MEG than EEG, thus MEG potentially facilitates accurate monitoring of cortical activity. We found that the direct electrophysiological MEG signals reflected the physiological status of neurological disorders and play a vital role in disease diagnosis. Single-channel connectivity, as well as brain network analysis, using MEG data acquired during resting state and a given task has been used for the diagnosis of neurological disorders such as epilepsy, Alzheimer’s, Parkinsonism, autism, and schizophrenia. The workflow of MEG and its potential applications in the diagnosis of disease and therapeutic planning are also discussed. We forecast that computer-aided algorithms will play a prominent role in the diagnosis and prediction of neurological diseases in the future. The outcome of this narrative review will aid researchers to utilise MEG in diagnostics.

An interactive framework for the detection of ictal and interictal activities: Cross-species and stand-alone implementation

BACKGROUND AND OBJECTIVE

Despite advances on signal analysis and artificial intelligence, visual inspection is the gold standard in event detection on electroencephalographic recordings. This process requires much time of clinical experts on both annotating and training new experts for this same task. In scenarios where epilepsy is considered, the need for automatic tools is more prominent, as both seizures and interictal events can occur on hours- or days-long recordings. Although other solutions have already been proposed, most of them are not integrated on clinical and basic science environments due to their complexity and required specialization. Here we present a pipeline that arises from coordinated efforts between life-science researchers, clinicians and data scientists to develop an interactive and iterative workflow to train machine-learning tools for the automatic detection of electroencephalographic events in a variety of scenarios.

METHODS

The approach consists on a series of subsequent steps covering data loading and configuration, event annotation, model training/re-training and event detection. With slight modifications, the combination of these blocks can cope with a variety of scenarios. To illustrate the flexibility and robustness of the approach, three datasets from clinical (patients of Dravet Syndrome) and basic research environments (mice model of the same disease) were evaluated. From them, and in response to researchers' daily needs, four real world examples of interictal event detection and seizure classification tasks were selected and processed.

RESULTS

Results show that the current approach was of great aid for event annotation and model development. It was capable of creating custom machine-learning solutions for each scenario with slight adjustments on the analysis protocol, easily accessible to users without programming skills. Final annotator similarity metrics reached values above 80% on all cases of use, reaching 92.3% on interictal event detection on human recordings.

CONCLUSIONS

The presented framework is easily adaptable to multiple real world scenarios and the interactive and ease-to-use approach makes it manageable to clinical and basic researches without programming skills. Nevertheless, it is conceived so data scientists can optimize it for specific scenarios, improving the knowledge transfer between these fields.

Sensitivity of a 29-Channel MEG Source Montage

In this paper, we study the performance of a source montage corresponding to 29 brain regions reconstructed from whole-head magnetoencephalographic (MEG) recordings, with the aim of facilitating the review of MEG data containing epileptiform discharges. Test data were obtained by superposing simulated signals from 100-nAm dipolar sources to a resting state MEG recording from a healthy subject. Simulated sources were placed systematically to different cortical locations for defining the optimal regularization for the source montage reconstruction and for assessing the detectability of the source activity from the 29-channel MEG source montage. The signal-to-noise ratio (SNR), computed for each source from the sensor-level and source-montage signals, was used as the evaluation parameter. Without regularization, the SNR from the simulated sources was larger in the sensor-level signals than in the source montage reconstructions. Setting the regularization to 2% increased the source montage SNR to the same level as the sensor-level SNR, improving the detectability of the simulated events from the source montage reconstruction. Sources producing a SNR of at least 15 dB were visually detectable from the source-montage signals. Such sources are located closer than about 75 mm from the MEG sensors, in practice covering all areas in the grey matter. The 29-channel source montage creates more focal signals compared to the sensor space and can significantly shorten the detection time of epileptiform MEG discharges for focus localization.

An Algorithm (LaD) for Monitoring Childbirth in Settings Where Tracking All Parameters in the World Health Organization Partograph Is Not Feasible: Design and Expert Validation

Background

After determining the key childbirth monitoring items from experts, we designed an algorithm (LaD) to represent the experts’ suggestions and validated it. In this paper we describe an abridged algorithm for labor and delivery management and use theoretical case to compare its performance with human childbirth experts.

Objective

The objective of this study was to describe the LaD algorithm, its development, and its validation. In addition, in the validation phase we wanted to assess if the algorithm was inferior, equivalent, or superior to human experts in recommending the necessary clinical actions during childbirth decision making.

Methods

The LaD algorithm encompasses the tracking of 6 of the 12 childbirth parameters monitored using the World Health Organization (WHO) partograph. It has recommendations on how to manage a patient when parameters are outside the normal ranges. We validated the algorithm with purposively selected experts selecting actions for a stratified sample of patient case scenarios. The experts’ selections were compared to obtain pairwise sensitivity and false-positive rates (FPRs) between them and the algorithm.

Results

The mean weighted pairwise sensitivity among experts was 68.2% (SD 6.95; 95% CI 59.6-76.8), whereas that between experts and the LaD algorithm was 69.4% (SD 17.95; 95% CI 47.1-91.7). The pairwise FPR among the experts ranged from 12% to 33% with a mean of 23.9% (SD 9.14; 95% CI 12.6-35.2), whereas that between experts and the algorithm ranged from 18% to 43% (mean 26.3%; SD 10.4; 95% CI 13.3-39.3). The was a correlation (mean 0.67 [SD 0.06]) in the actions selected by the expert pairs for the different patient cases with a reliability coefficient (α) of .91.

Conclusions

The LaD algorithm was more sensitive, but had a higher FPR than the childbirth experts, although the differences were not statistically significant. An electronic tool for childbirth monitoring with fewer WHO-recommended parameters may not be inferior to human experts in labor and delivery clinical decision support.

Interrater Reliability of Experts in Identifying Interictal Epileptiform Discharges in Electroencephalograms

Importance

The validity of using electroencephalograms (EEGs) to diagnose epilepsy requires reliable detection of interictal epileptiform discharges (IEDs). Prior interrater reliability (IRR) studies are limited by small samples and selection bias.

Objective

To assess the reliability of experts in detecting IEDs in routine EEGs.

Design, Setting, and Participants

This prospective analysis conducted in 2 phases included as participants physicians with at least 1 year of subspecialty training in clinical neurophysiology. In phase 1, 9 experts independently identified candidate IEDs in 991 EEGs (1 expert per EEG) reported in the medical record to contain at least 1 IED, yielding 87 636 candidate IEDs. In phase 2, the candidate IEDs were clustered into groups with distinct morphological features, yielding 12 602 clusters, and a representative candidate IED was selected from each cluster. We added 660 waveforms (11 random samples each from 60 randomly selected EEGs reported as being free of IEDs) as negative controls. Eight experts independently scored all 13 262 candidates as IEDs or non-IEDs. The 1051 EEGs in the study were recorded at the Massachusetts General Hospital between 2012 and 2016.

Main Outcomes and Measures

Primary outcome measures were percentage of agreement (PA) and beyond-chance agreement (Gwet κ) for individual IEDs (IED-wise IRR) and for whether an EEG contained any IEDs (EEG-wise IRR). Secondary outcomes were the correlations between numbers of IEDs marked by experts across cases, calibration of expert scoring to group consensus, and receiver operating characteristic analysis of how well multivariate logistic regression models may account for differences in the IED scoring behavior between experts.

Results

Among the 1051 EEGs assessed in the study, 540 (51.4%) were those of females and 511 (48.6%) were those of males. In phase 1, 9 experts each marked potential IEDs in a median of 65 (interquartile range [IQR], 28-332) EEGs. The total number of IED candidates marked was 87 636. Expert IRR for the 13 262 individually annotated IED candidates was fair, with the mean PA being 72.4% (95% CI, 67.0%-77.8%) and mean κ being 48.7% (95% CI, 37.3%-60.1%). The EEG-wise IRR was substantial, with the mean PA being 80.9% (95% CI, 76.2%-85.7%) and mean κ being 69.4% (95% CI, 60.3%-78.5%). A statistical model based on waveform morphological features, when provided with individualized thresholds, explained the median binary scores of all experts with a high degree of accuracy of 80% (range, 73%-88%).

Conclusions and Relevance

This study's findings suggest that experts can identify whether EEGs contain IEDs with substantial reliability. Lower reliability regarding individual IEDs may be largely explained by various experts applying different thresholds to a common underlying statistical model.

Intravenous dexmedetomidine sedation for magnetoencephalography: A retrospective study

BACKGROUND

Magnetoencephalography (MEG) plays a preponderant role in the preoperative assessment of patients with drug-resistant epilepsy (DRE). However, the magnetoencephalography of patients with drug-resistant epilepsy can be difficult without sedation and/or general anesthesia. Our objective is to describe our experience with intravenous dexmedetomidine as sedation for magnetoencephalography and its effect, if any, on the ability to recognize epileptic spikes.

METHODS

In this retrospective study, we reviewed the records of 89 children who presented for Magnetoencephalography/electroencephalography (EEG) scans between August of 2008 and May of 2015. Data analyzed included demographics and the frequency of epileptic spikes. Sedated magnetoencephalography recordings were compared to nonsedated video-electroencephalography (vEEG) recordings in the same patients to determine the impact of dexmedetomidine.

RESULTS

Spike frequency between magnetoencephalography with sedation and video-electroencephalography without sedation was compared in 85 patients. Magnetoencephalography and video-electroencephalography were considered clinically concordant in 80 patients (94.1%) and discordant in 5 patients (5.9%), all with less spikes during Magnetoencephalography. The median (range) bolus dose of dexmedetomidine was 2 (1-2) mcg/kg. The median (range) infusion rate of dexmedetomidine was 2 (0.5-4) mcg/kg/h. All patients experienced reductions in heart rate after administration of dexmedetomidine; these reductions were statistically, but not clinically, significant.

CONCLUSIONS

Our results suggest that dexmedetomidine-based protocol provides reliable sedation in children undergoing MEG scanning because of the high success rate, limited interictal artifacts, and minimal impacts on spike frequency.

A fast machine learning approach to facilitate the detection of interictal epileptiform discharges in the scalp electroencephalogram

BACKGROUND

Finding interictal epileptiform discharges (IEDs) in the EEG is a part of diagnosing epilepsy. Automated software for annotating EEGs of patients with suspected epilepsy can therefore help with reaching a diagnosis. A large amount of data is required for training and evaluating an effective IED detection system. IEDs occur infrequently in the most patients' EEG, therefore, interictal EEG recordings contain mostly background waveforms.

NEW METHOD

As the first step to detect IEDs, we propose a machine learning technique eliminating most EEG background data using an ensemble of simple fast classifiers based on several EEG features. This could save computation time for an IED detection method, allowing the remaining waveforms to be classified by more computationally intensive methods. We consider several efficient features and reject background by applying thresholds on them in consecutive steps.

RESULTS

We applied the proposed algorithm on a dataset of 156 EEGs (93 and 63 with and without IEDs, respectively). We were able to eliminate 78% of background waveforms while retaining 97% of IEDs on our cross-validated dataset.

COMPARISON WITH EXISTING METHODS

We applied support vector machine, k-nearest neighbours, and random forest classifiers to detect IEDs with and without initial background rejection. Results show that rejecting background by our proposed method speeds up the overall classification by a factor ranging from 1.8 to 4.7 for the considered classifiers.

CONCLUSIONS

The proposed method successfully reduces computation time of an IED detection system. Therefore, it is beneficial in speeding up IED detection especially when utilizing large EEG datasets.

Interictal epileptiform discharge characteristics underlying expert interrater agreement

OBJECTIVE

The presence of interictal epileptiform discharges (IED) in the electroencephalogram (EEG) is a key finding in the medical workup of a patient with suspected epilepsy. However, inter-rater agreement (IRA) regarding the presence of IED is imperfect, leading to incorrect and delayed diagnoses. An improved understanding of which IED attributes mediate expert IRA might help in developing automatic methods for IED detection able to emulate the abilities of experts. Therefore, using a set of IED scored by a large number of experts, we set out to determine which attributes of IED predict expert agreement regarding the presence of IED.

METHODS

IED were annotated on a 5-point scale by 18 clinical neurophysiologists within 200 30-s EEG segments from recordings of 200 patients. 5538 signal analysis features were extracted from the waveforms, including wavelet coefficients, morphological features, signal energy, nonlinear energy operator response, electrode location, and spectrogram features. Feature selection was performed by applying elastic net regression and support vector regression (SVR) was applied to predict expert opinion, with and without the feature selection procedure and with and without several types of signal normalization.

RESULTS

Multiple types of features were useful for predicting expert annotations, but particular types of wavelet features performed best. Local EEG normalization also enhanced best model performance. As the size of the group of EEGers used to train the models was increased, the performance of the models leveled off at a group size of around 11.

CONCLUSIONS

The features that best predict inter-rater agreement among experts regarding the presence of IED are wavelet features, using locally standardized EEG. Our models for predicting expert opinion based on EEGer's scores perform best with a large group of EEGers (more than 10).

SIGNIFICANCE

By examining a large group of EEG signal analysis features we found that wavelet features with certain wavelet basis functions performed best to identify IEDs. Local normalization also improves predictability, suggesting the importance of IED morphology over amplitude-based features. Although most IED detection studies in the past have used opinion from three or fewer experts, our study suggests a "wisdom of the crowd" effect, such that pooling over a larger number of expert opinions produces a better correlation between expert opinion and objectively quantifiable features of the EEG.

Characteristics of EEG Interpreters Associated With Higher Interrater Agreement

PURPOSE

The goal of the project is to determine characteristics of academic neurophysiologist EEG interpreters (EEGers), which predict good interrater agreement (IRA) and to determine the number of EEGers needed to develop an ideal standardized testing and training data set for epileptiform transient (ET) detection algorithms.

METHODS

A three-phase scoring method was used. In phase 1, 19 EEGers marked the location of ETs in two hundred 30-second segments of EEG from 200 different patients. In phase 2, EEG events marked by at least 2 EEGers were annotated by 18 EEGers on a 5-point scale to indicate whether they were ETs. In phase 3, a third opinion was obtained from EEGers on any inconsistencies between phase 1 and phase 2 scoring.

RESULTS

The IRA for the 18 EEGers was only fair. A select group of the EEGers had good IRA and the other EEGers had low IRA. Board certification by the American Board of Clinical Neurophysiology was associated with better IRA performance but other board certifications, years of fellowship training, and years of practice were not. As the number of EEGers used for scoring is increased, the amount of change in the consensus opinion decreases steadily and is quite low as the group size approaches 10.

CONCLUSIONS

The IRA among EEGers varies considerably. The EEGers must be tested before use as scorers for ET annotation research projects. The American Board of Clinical Neurophysiology certification is associated with improved performance. The optimal size for a group of experts scoring ETs in EEG is probably in the 6 to 10 range.

Audit of EEG Reporting Temporal Abnormalities

Background:

To ensure the overall quality of our electroencephalogram (EEG) laboratory, we decided to perform an audit of EEGs interpreted at our institution, focusing initially on EEGs reporting temporal abnormalities.

Methods:

Reports of all EEGs performed between January 1st and June 30th, 2006 were reviewed in order to identify tracings mentioning abnormalities in the temporal regions. These records were then independently reviewed by two epileptologists on two distinct occasions, separated by an interval of at least six months. If the recording was considered normal after this process, the cause for misinterpretation was identified and the patient's chart was reviewed to determine if he was epileptic or not based on available evidence until June 2009.

Results:

In the first half of 2006,143 out of 773 EEGs mentioned temporal abnormalities (18.5%). In general, intra- and interrater agreement ratios between our two epileptologists were moderate to substantial for normality, presence of epileptic activity and presence of slowing. Forty-five recordings (31.5%) were reported as normal independently by them on two distinct sittings six months apart. The most common causes for misinterpretation were the presence of benign epileptiform variants, normal sharply contoured patterns of somnolence or hyperventilation. Chart review confirmed that most were non-epileptic patients (60% non-epileptic, 27% epileptic, 13% unknown).

Conclusion:

Moderate to substantial intra- and interrater agreement as well as frequent misinterpretation of physiological variants indicate that some corrective measures need to be implemented to improve the consistency of EEG interpretation amongst our group of electroencephalographers.

The relationship between seizures, interictal spikes and antiepileptic drugs

OBJECTIVE

A considerable decrease in spike rate accompanies antiepileptic drug (AED) taper during intracranial EEG (icEEG) monitoring. Since spike rate during icEEG monitoring can be influenced by surgery to place intracranial electrodes, we studied spike rate during long-term scalp EEG monitoring to further test this observation.

METHODS

We analyzed spike rate, seizure occurrence and AED taper in 130 consecutive patients over an average of 8.9days (range 5-17days).

RESULTS

We observed a significant relationship between time to the first seizure, spike rate, AED taper and seizure occurrence (F (3,126)=19.77, p<0.0001). A high spike rate was related to a longer time to the first seizure. Further, in a subset of 79 patients who experienced seizures on or after day 4 of monitoring, spike rate decreased initially from an on- to off-AEDs epoch (from 505.0 to 382.3 spikes per hour, p<0.00001), and increased thereafter with the occurrence of seizures.

CONCLUSIONS

There is an interplay between seizures, spikes and AEDs such that spike rate decreases with AED taper and increases after seizure occurrence.

SIGNIFICANCE

The direct relationship between spike rate and AEDs and between spike rate and time to the first seizure suggests that spikes are a marker of inhibition rather than excitation.

Characterization of Subdural Stimulation-Induced Afterdischarge Activity Using the Continuous Wavelet Transform

OBJECTIVE

We address the problem of characterization of afterdischarges (ADs) that often arise in patients with intractable focal epilepsy who, as part of their evaluation, undergo cortical electrical stimulation: A standard diagnostic and evaluation procedure before respective surgery.

RESULTS

A total of 1333 channels of data recorded in 17 trials of seven patients whose EEG showed ADs (on a total of 156 channels) during cortical stimulation were examined in the time-scale domain using a complex Morlet scalogram. We found excellent characterization of the AD channels based on the distribution functions of the sum of the wavelet coefficients in the two lowest scales corresponding to the frequency range [20, 80] Hz, i.e., the β and γ ranges of EEG.

CONCLUSION

We suggest that the transient Morlet wavelet and the scale domain activity function of the EEG in the two lowest scales (as defined in this paper) could serve as a very useful decision aid in the identification of ADs during and after cortical electrical stimulation.

SIGNIFICANCE

In patients undergoing cortical electrical stimulation, AD waveforms can cause misleading test results by altering the ongoing electroencephalogram (EEG), and can become unwanted seizures. Any process to suppress the ADs rests on a reliable method to distinguish them from normal EEG channels, a task that is usually performed by visual inspection, and that is complicated by the fact that ADs have multiple distinct morphologies. The single feature of the EEG in our study resulted in average probability of detection of 0.99 with an average false alarm probability of 0.04. It is likely that the addition of one or two more features to our decision aid could improve sensitivity and selectivity to near perfection.

Studentized continuous wavelet transform (t-CWT) in the analysis of individual ERPs: real and simulated EEG data

This study aimed at evaluating the performance of the Studentized Continuous Wavelet Transform (t-CWT) as a method for the extraction and assessment of event-related brain potentials (ERP) in data from a single subject. Sensitivity, specificity, positive (PPV) and negative predictive values (NPV) of the t-CWT were assessed and compared to a variety of competing procedures using simulated EEG data at six low signal-to-noise ratios. Results show that the t-CWT combines high sensitivity and specificity with favorable PPV and NPV. Applying the t-CWT to authentic EEG data obtained from 14 healthy participants confirmed its high sensitivity. The t-CWT may thus be well suited for the assessment of weak ERPs in single-subject settings.

Detection of interictal epileptiform discharges using signal envelope distribution modelling: application to epileptic and non-epileptic intracranial recordings

Interictal epileptiform discharges (spikes, IEDs) are electrographic markers of epileptic tissue and their quantification is utilized in planning of surgical resection. Visual analysis of long-term multi-channel intracranial recordings is extremely laborious and prone to bias. Development of new and reliable techniques of automatic spike detection represents a crucial step towards increasing the information yield of intracranial recordings and to improve surgical outcome. In this study, we designed a novel and robust detection algorithm that adaptively models statistical distributions of signal envelopes and enables discrimination of signals containing IEDs from signals with background activity. This detector demonstrates performance superior both to human readers and to an established detector. It is even capable of identifying low-amplitude IEDs which are often missed by experts and which may represent an important source of clinical information. Application of the detector to non-epileptic intracranial data from patients with intractable facial pain revealed the existence of sharp transients with waveforms reminiscent of interictal discharges that can represent biological sources of false positive detections. Identification of these transients enabled us to develop and propose secondary processing steps, which may exclude these transients, improving the detector's specificity and having important implications for future development of spike detectors in general.

Standardized database development for EEG epileptiform transient detection: EEGnet scoring system and machine learning analysis

The routine scalp electroencephalogram (rsEEG) is the most common clinical neurophysiology procedure. The most important role of rsEEG is to detect evidence of epilepsy, in the form of epileptiform transients (ETs), also known as spike or sharp wave discharges. Due to the wide variety of morphologies of ETs and their similarity to artifacts and waves that are part of the normal background activity, the task of ET detection is difficult and mistakes are frequently made. The development of reliable computerized detection of ETs in the EEG could assist physicians in interpreting rsEEGs. We report progress in developing a standardized database for testing and training ET detection algorithms. We describe a new version of our EEGnet software system for collecting expert opinion on EEG datasets, a completely web-browser based system. We report results of EEG scoring from a group of 11 board-certified academic clinical neurophysiologists who annotated 30-s excepts from rsEEG recordings from 100 different patients. The scorers had moderate inter-scorer reliability and low to moderate intra-scorer reliability. In order to measure the optimal size of this standardized rsEEG database, we used machine learning models to classify paroxysmal EEG activity in our database into ET and non-ET classes. Based on our results, it appears that our database will need to be larger than its current size. Also, our non-parametric classifier, an artificial neural network, performed better than our parametric Bayesian classifier. Of our feature sets, the wavelet feature set proved most useful for classification.

Fast evaluation of interictal spikes in long-term EEG by hyper-clustering

PURPOSE

The burden of reviewing long-term scalp electroencephalography (EEG) is not much alleviated by automated spike detection if thousands of events need to be inspected and mentally classified by the reviewer. This study investigated a novel technique of clustering and 24-h hyper-clustering on top of automated detection to assess whether fast review of focal interictal spike types was feasible and comparable to the spikes types observed during routine EEG review in epilepsy monitoring.

METHODS

Spike detection used a transformation of scalp EEG into 29 regional source activities and adaptive thresholds to increase sensitivity. Our rule-based algorithm estimated 18 parameters around each detected peak and combined multichannel detections into one event. Similarity measures were derived from equivalent location, scalp topography, and source waveform of each event to form clusters over 2-h epochs using a density-based algorithm. Similar measures were applied to all 2-h clusters to form 24-h hyper-clusters. Independent raters evaluated electroencephalography data of 50 patients with epilepsy (25 children) using traditional visual spike review and optimized hyper-cluster inspection. Congruence between visual spike types and epileptiform hyper-clusters was assessed on a sublobar level using three-dimensional (3D) peak topographies.

KEY FINDINGS

Visual rating found 126 different epileptiform spike types (2.5 per patient). Independently, 129 hyper-clusters were classified as epileptiform and originating in separate sublobar regions (2.6 per patient). Ninety-one percent of visual spike types matched with hyper-clusters (temporal lobe spikes 94%, extratemporal 89%). Conversely, 11% of hyper-clusters rated epileptiform had no corresponding visual spike type. Numbers were comparable in adults and children. On average, 15 hyper-clusters had to be inspected and rated per patient with an evaluation time of around 5 min.

SIGNIFICANCE

Hyper-clustering over 24 h provides an independent tool for rapid daily evaluation of interictal spikes in long-term video-EEG monitoring. If used in addition to routine review of 2-5 min EEG per hour, sensitivity and reliability in noninvasive diagnosis of focal epilepsy increases.

Combined process automation for large-scale EEG analysis

Epileptogenesis is a dynamic process producing increased seizure susceptibility. Electroencephalography (EEG) data provides information critical in understanding the evolution of epileptiform changes throughout epileptic foci. We designed an algorithm to facilitate efficient large-scale EEG analysis via linked automation of multiple data processing steps. Using EEG recordings obtained from electrical stimulation studies, the following steps of EEG analysis were automated: (1) alignment and isolation of pre- and post-stimulation intervals, (2) generation of user-defined band frequency waveforms, (3) spike-sorting, (4) quantification of spike and burst data and (5) power spectral density analysis. This algorithm allows for quicker, more efficient EEG analysis.

High inter-reviewer variability of spike detection on intracranial EEG addressed by an automated multi-channel algorithm

OBJECTIVE

The goal of this study was to determine the consistency of human reviewer spike detection and then develop a computer algorithm to make the intracranial spike detection process more objective and reliable.

METHODS

Three human reviewers marked interictal spikes on samples of intracranial EEGs from 10 patients. The sensitivity, precision and agreement in channel ranking by activity were calculated between reviewers. A computer algorithm was developed to parallel the way human reviewers detect spikes by first identifying all potential spikes on each channel using frequency filtering and then block scaling all channels at the same time in order to exclude potential spikes that fall below an amplitude and slope threshold. Its performance was compared to the human reviewers on the same set of patients.

RESULTS

Human reviewers showed surprisingly poor inter-reviewer agreement, but did broadly agree on the ranking of channels for spike activity. The computer algorithm performed as well as the human reviewers and did especially well at ranking channels from highest to lowest spike frequency.

CONCLUSIONS

Our algorithm showed good agreement with the different human reviewers, even though they demonstrated different criteria for what constitutes a 'spike' and performed especially well at the clinically important task of ranking channels by spike activity.

SIGNIFICANCE

An automated, objective method to detect interictal spikes on intracranial recordings will improve both research and the surgical management of epilepsy patients.

Web-based collection of expert opinion on routine scalp EEG: software development and interrater reliability

Computerized detection of epileptiform transients (ETs), characterized by interictal spikes and sharp waves in the EEG, has been a research goal for the last 40 years. A reliable method for detecting ETs would assist physicians in interpretation and improve efficiency in reviewing long-term EEG recordings. Computer algorithms developed thus far for detecting ETs are not as reliable as human experts, primarily due to the large number of false-positive detections. Comparing the performance of different algorithms is difficult because each study uses individual EEG test datasets. In this article, we present EEGnet, a distributed web-based platform for the acquisition and analysis of large-scale training datasets for comparison of different EEG ET detection algorithms. This software allows EEG scorers to log in through the web, mark EEG segments of interest, and categorize segments of interest using a conventional clinical EEG user interface. This software platform was used by seven board-certified academic epileptologists to score 40 short 30-second EEG segments from 40 patients, half containing ETs and half containing artifacts and normal variants. The software performance was adequate. Interrater reliability for marking the location of paroxysmal activity was low. Interrater reliability of marking artifacts and ETs was high and moderate, respectively.

EEG spike activity precedes epilepsy after kainate-induced status epilepticus

PURPOSE

Chronic epilepsy frequently develops after brain injury, but prediction of which individual patient will develop spontaneous recurrent seizures (i.e., epilepsy) is not currently possible. Here, we use continuous radiotelemetric electroencephalography (EEG) and video monitoring along with automated computer detection of EEG spikes and seizures to test the hypothesis that EEG spikes precede and are correlated with subsequent spontaneous recurrent seizures.

METHODS

The presence and pattern of EEG spikes was studied during long recording epochs between the end of status epilepticus (SE) induced by three different doses of kainate and the onset of chronic epilepsy.

RESULTS

The presence of spikes, and later spike clusters, over several days after SE before the first spontaneous seizure, was consistently associated with the development of chronic epilepsy. The rate of development of epilepsy (i.e., increase in seizure frequency) was strongly correlated with the frequency of EEG spikes and the cumulative number of EEG spikes after SE.

CONCLUSIONS

The temporal features of EEG spikes (i.e., their presence, frequency, and pattern [clusters]) when analyzed over prolonged periods, may be a predictive biomarker for the development of chronic epilepsy after brain injury. Future clinical trials using prolonged EEG recordings may reveal the diagnostic utility of EEG spikes as predictors of subsequent epilepsy in brain-injured humans.

Automatic EEG spike detection

Since the 1970s advances in science and technology during each succeeding decade have renewed the expectation of efficient, reliable automatic epileptiform spike detection (AESD). But even when reinforced with better, faster tools, clinically reliable unsupervised spike detection remains beyond our reach. Expert-selected spike parameters were the first and still most widely used for AESD. Thresholds for amplitude, duration, sharpness, rise-time, fall-time, after-coming slow waves, background frequency, and more have been used. It is still unclear which of these wave parameters are essential, beyond peak-peak amplitude and duration. Wavelet parameters are very appropriate to AESD but need to be combined with other parameters to achieve desired levels of spike detection efficiency. Artificial Neural Network (ANN) and expert-system methods may have reached peak efficiency. Support Vector Machine (SVM) technology focuses on outliers rather than centroids of spike and nonspike data clusters and should improve AESD efficiency. An exemplary spike/nonspike database is suggested as a tool for assessing parameters and methods for AESD and is available in CSV or Matlab formats from the author at brainvue@gmail.com. Exploratory Data Analysis (EDA) is presented as a graphic method for finding better spike parameters and for the step-wise evaluation of the spike detection process.

Detection of focal epileptiform events in the EEG by spatio-temporal dipole clustering

OBJECTIVE

Methods for the detection of epileptiform events can be broadly divided into two main categories: temporal detection methods that exploit the EEG's temporal characteristics, and spatial detection methods that base detection on the results of an implicit or explicit source analysis. We describe how the framework of a spatial detection method was extended to improve its performance by including temporal information. This results in a method that provides (i) automated localization of an epileptogenic focus and (ii) detection of focal epileptiform events in an EEG recording. For the detection, only one threshold value needs to be set.

METHODS

The method comprises five consecutive steps: (1) dipole source analysis in a moving window, (2) automatic selection of focal brain activity, (3) dipole clustering to arrive at the identification of the epileptiform cluster, (4) derivation of a spatio-temporal template of the epileptiform activity, and (5) template matching. Routine EEG recordings from eight paediatric patients with focal epilepsy were labelled independently by two experts. The method was evaluated in terms of (i) ability to identify the epileptic focus, (ii) validity of the derived template, and (iii) detection performance. The clustering performance was evaluated using a leave-one-out cross validation. Detection performance was evaluated using Precision-Recall curves and compared to the performance of two temporal (mimetic and wavelet based) and one spatial (dipole analysis based) detection methods.

RESULTS

The method succeeded in identifying the epileptogenic focus in seven of the eight recordings. For these recordings, the mean distance between the epileptic focus estimated by the method and the region indicated by the labelling of the experts was 8mm. Except for two EEG recordings where the dipole clustering step failed, the derived template corresponded to the epileptiform activity marked by the experts. Over the eight EEGs, the method showed a mean sensitivity and selectivity of 92 and 77%, respectively.

CONCLUSIONS

The method allows automated localization of the epileptogenic focus and shows good agreement with the region indicated by the labelling of the experts. If the dipole clustering step is successful, the method allows a detection of the focal epileptiform events, and gave a detection performance comparable or better to that of the other methods.

SIGNIFICANCE

The identification and quantification of epileptiform events is of considerable importance in the diagnosis of epilepsy. Our method allows the automatic identification of the epileptic focus, which is of value in epilepsy surgery. The method can also be used as an offline exploration tool for focal EEG activity, displaying the dipole clusters and corresponding time series.

Magnetoencephalography Is More Successful for Screening and Localizing Frontal Lobe Epilepsy than Electroencephalography

PURPOSE

The diagnosis of frontal lobe epilepsy may be compounded by poor electroclinical localization, due to distributed or rapidly propagating epileptiform activity. This study aimed at developing optimal procedures for localizing interictal epileptiform discharges (IEDs) of patients with localization related epilepsy in the frontal lobe. To this end the localization results obtained for magnetoencephalography (MEG) and electroencephalography (EEG) were compared systematically using automated analysis procedures.

METHODS

Simultaneous recording of interictal EEG and MEG was successful for 18 out of the 24 patients studied. Visual inspection of these recordings revealed IEDs with varying morphology and topography. Cluster analysis was used to classify these discharges on the basis of their spatial distribution followed by equivalent dipole analysis of the cluster averages. The locations of the equivalent dipoles were compared with the location of the epileptogenic lesions of the patient or, if these were not visible at MRI with the location of the interictal onset zones identified by subdural electroencephalography.

RESULTS

Generally IEDs were more abundantly in MEG than in the EEG recordings. Furthermore, the duration of the MEG spikes, measured from the onset till the spike maximum, was in most patients shorter than the EEG spikes. In most patients, distinct spike subpopulations were found with clearly different topographical field maps. Cluster analysis of MEG spikes followed by dipole localization was successful (n = 14) for twice as many patients as for EEG source analysis (n = 7), indicating that the localizability of interictal MEG is much better than of interictal EEG.

CONCLUSIONS

The automated procedures developed in this study provide a fast screening method for identifying the distinct categories of spikes and the brain areas responsible for these spikes. The results show that MEG spike yield and localization is superior compared with EEG. This finding is of importance for the diagnosis and preoperative evaluation of patients with frontal lobe epilepsy.

Comparison of novel computer detectors and human performance for spike detection in intracranial EEG

OBJECTIVE

Interictal spikes in intracranial EEG (iEEG) may correlate with epileptogenic cortex, but review of interictal iEEG is labor intensive. Accurate automated spike detectors are necessary for understanding the role of spikes in epileptogenesis.

METHODS

The sensitivity, accuracy and reproducibility of three automated iEEG spike detectors were compared against two human EEG readers using iEEG segments from eight patients. A consensus set of detections was generated for detector calibration. Spike verification was calculated after both human EEG readers independently reviewed all detections.

RESULTS

Humans and two of the three automated detectors demonstrated comparable accuracy. In four patients, automated spike detection sensitivity was >70% and accuracy was >50%. In the remaining four patients, EEG background morphology resulted in poorer performance. Blinded human verification accuracy was 76.7+/-6.6% for computer-detected spikes, and 84.5+/-4.1% for human-detected spikes.

CONCLUSIONS

Automated iEEG spike detectors perform comparably to humans, but sensitivity and accuracy are patient dependent. Humans verified the majority of computer-detected spikes.

SIGNIFICANCE

In some patients automated detectors may be used for mapping spike occurrences in epileptic networks. This may reveal associations between spike distribution, seizure onset, and pathology.

Efficient unsupervised algorithms for the detection of seizures in continuous EEG recordings from rats after brain injury

Long-term EEG monitoring in chronically epileptic animals produces very large EEG data files which require efficient algorithms to differentiate interictal spikes and seizures from normal brain activity, noise, and, artifact. We compared four methods for seizure detection based on (1) EEG power as computed using amplitude squared (the power method), (2) the sum of the distances between consecutive data points (the coastline method), (3) automated spike frequency and duration detection (the spike frequency method), and (4) data range autocorrelation combined with spike frequency (the autocorrelation method). These methods were used to analyze a randomly selected test set of 13 days of continuous EEG data in which 75 seizures were imbedded. The EEG recordings were from eight different rats representing two different models of chronic epilepsy (five kainate-treated and three hypoxic-ischemic). The EEG power method had a positive predictive value (PPV, or true positives divided by the sum of true positives and false positives) of 18% and a sensitivity (true positives divided by the sum of true positives and false negatives) of 95%, the coastline method had a PPV of 78% and sensitivity of 99.59, the spike frequency method had a PPV of 78% and a sensitivity of 95%, and the autocorrelation method yielded a PPV of 96% and a sensitivity of 100%. It is possible to detect seizures automatically in a prolonged EEG recording using computationally efficient unsupervised algorithms. Both the quality of the EEG and the analysis method employed affect PPV and sensitivity.

Identifying the primary epileptogenic hemisphere from electroencephalographic (EEG) and magnetoencephalographic dipole lateralizations in children with intractable epilepsy

We used electroencephalographic (EEG) and magnetoencephalographic dipole lateralizations to identify the primary epileptogenic hemisphere in 41 children with intractable localization-related epilepsy. We compared EEG and magnetoencephalographic dipole lateralizations, EEG ictal onsets, and magnetic resonance images (MRIs). Concordant lateralization of EEG and magnetoencephalographic dipoles (> 50% of each lateralizing to the same hemisphere) occurred in 34 patients, with EEG ictal onsets in the same hemisphere in 23 (68%) and concordant MRI lesions in 23 (68%). Focal resection in 16 of 20 patients resulted in a good surgical outcome. Of the seven children with nonconcordant magnetoencephalographic and EEG lateralizations, one (14%) had EEG ictal onset and one (14%) had MRI lesions that lateralized; none had surgery. The relationship between lateralized EEG and magnetoencephalographic dipoles forecasts surgical candidacy. Concordant lateralizations predict good seizure control after surgery by identifying the primary epileptogenic hemisphere. Discordant lateralizations signify an undetermined epileptogenic hemisphere and contraindicate surgery without further testing.

Detection of epileptiform activity by human interpreters: blinded comparison between electroencephalography and magnetoencephalography

PURPOSE

Objectively to evaluate whether independent spike detection by human interpreters is clinically valid in magnetoencephalography (MEG) and to characterize detection differences between MEG and scalp electroencephalography (EEG).

METHODS

We simultaneously recorded scalp EEG and MEG data from 43 patients with intractable focal epilepsy. Raw EEG and MEG waveforms were reviewed independently by two experienced epileptologists, one for EEG and one for MEG, blinded to the other modality and to the clinical information. The number and localization of spikes detected by EEG and/or MEG were compared in relation to clinical diagnosis based on postoperative seizure freedom.

RESULTS

Interictal spikes were captured in both EEG and MEG in 31, in MEG alone in eight, in EEG alone in one, and in neither modality in three patients. The number of detections ranged widely with no statistical difference between modalities. A median of 25.7% of total spikes was detectable by both modalities. Spike localization was similarly consistent with the epilepsy diagnosis in 85.2% (EEG) and 78.1% (MEG) of the patients. Inaccurate localization occurred only in those cases with very few spikes detected, especially when the detections were in one modality alone.

CONCLUSIONS

Interictal epileptiform discharges are easily perceived in MEG. Independent spike identification in MEG can provide clinical results comparable, but not superior, to EEG. Many spikes were seen in only one modality or the other; therefore the use of both EEG and MEG may provide additional information.

Seizure detection: evaluation of the Reveal algorithm

OBJECTIVE

The aim of this study is to evaluate an improved seizure detection algorithm and to compare with two other algorithms and human experts.

METHODS

672 seizures from 426 epilepsy patients were examined with the (new) Reveal algorithm which utilizes 3 methods, novel in their application to seizure detection: Matching Pursuit, small neural network-rules and a new connected-object hierarchical clustering algorithm.

RESULTS

Reveal had a sensitivity of 76% with a false positive rate of 0.11/h. Two other algorithms (Sensa and CNet) were tested and had sensitivities of 35.4 and 48.2% and false positive rates of 0.11/h and 0.75/h, respectively.

CONCLUSIONS

This study validates the Reveal algorithm, and shows it to compare favorably with other methods.

SIGNIFICANCE

Improved seizure detection can improve patient care in both the epilepsy monitoring unit and the intensive care unit.

Patterns of interictal spike propagation across the central sulcus in benign rolandic epilepsy

It has been reported that the rolandic area generating spikes is hyperexcitable, and that rolandic spikes propagate across the central area. However, the pattern of rolandic spike propagation and how the dipolar distribution of the spikes is related to the propagation pattern have not yet been studied. Thirty-nine EEGs from 27 patients with benign rolandic epilepsy (BRE) were examined. Sequential topographic mapping in 4-ms steps was used to analyze the pattern of spike propagation. The locations of maximum negative foci, the presence and distribution of the dipolar field, and the propagation pattern were examined. Dipoles were present in 23 (85.2%) out of 27 patients and in 43 (72.9%) out of 59 foci. Thirty-two foci (54.2%) in 20 patients demonstrated a propagation pattern. The typical pattern consisted of propagation from central to mid-temporal locations across the central sulcus. Most spike foci exhibiting a propagation pattern had a dipolar distribution (87.5%; p=0.008). These results suggest that rolandic spikes originate from sulcal or gyral cortices on either side of the central sulcus, and that spike propagation can ensue by intracortical spreading across the central sulcus.

EEG quality during simultaneous functional MRI of interictal epileptiform discharges

This article concerns the evaluation of the quality of interictal epileptiform EEG discharges recorded throughout simultaneous echo planar imaging (EPI). BOLD (blood oxygen level dependent) functional MRI (fMRI) images were acquired continuously on a patient with intractable epilepsy. EEG was sampled simultaneously, during and after imaging, with removal of pulse and imaging artifacts by subtraction of channel-specific running averages. Contiguous EEG epochs recorded with and without fMRI (fMRI+ve vs. fMRI-ve) were next randomized and presented to two blinded observers. Epileptiform discharges were identified retrospectively, and comparison was made in terms of the number of identified events, their amplitude, and spatiotemporal distribution. A spectral analysis was also performed on the EEG. In the randomized comparison of EEG segments, 80 (fMRI+ve) vs. 69 (fMRI-ve) discharges were noted with good interobserver agreement (69%). There were no significant differences in amplitude or spatio-temporal distribution. Comparison of the events detected and measured by two expert observers demonstrated that the Interictal Epileptiform Discharge (IED) characteristics were indistinguishable with and without scanning. We review briefly the existing literature on EEG recording quality for combined EEG/fMRI.

Seizure detection: correlation of human experts

OBJECTIVE

The description and application of a new, overlap-integral comparison method and the quantification of human vs. human accuracies that can be used as goals for algorithms.

METHODS

Four human experts marked ten 8 h electroencephalography (EEG) records from seizure patients. The seizures varied in origin and type, including complex partial, generalized absence, secondarily generalized and primary generalized tonic-clonic. The traditional any-overlap comparison method is used in addition to the overlap-integral method, which is sensitive to the correct placement of the seizure endpoints.

RESULTS

The number of events marked by each reader ranged from 57 to 77. The average any-overlap sensitivity and false positives per hour rate are 0.92 and 0.117. The average overlap-integral correlation, sensitivity and specificity are 0.80, 0.82 and 0.9926. As expected, the correspondence between readers is high, but confounding issues resulted in overlap-integral sensitivities less than 0.5 for 10% of the records. Seven percent of the any-overlap sensitivities are less than 0.5. A comparison of the methods by record shows that the overlap-integral specificity and the any-overlap false positive rate measure different features.

CONCLUSIONS

There was little variation between readers and they were essentially interchangeable. High seizure rate (many per hour), short seizure durations (<10 s) and long seizure durations (approximately 10 min) with ambiguous offsets can complicate the analysis and result in poor correlation. There may be any number of unmarked events in rigorously marked records and it may be preferable to use records from non-epilepsy patients to compute the false positive rate. The any-overlap and overlap-integral comparison methods are complementary.

SIGNIFICANCE

Correlation between expert human readers can be low on some records, which will complicate testing of seizure detection algorithms.

Spike cluster analysis in neocortical localization related epilepsy yields clinically significant equivalent source localization results in magnetoencephalogram (MEG)

OBJECTIVE

In magnetoencephalogram (MEG) recordings of patients with epilepsy several types of sharp transients with different spatiotemporal distributions are commonly present. Our objective was to develop a computer based method to identify and classify groups of epileptiform spikes, as well as other transients, in order to improve the characterization of irritative areas in the brain of epileptic patients.

METHODS

MEG data centered on selected spikes were stored in signal matrices of C channels by T time samples. The matrices were normalized and euclidean distances between spike representations in vector space R(CxT) were input to a Ward's hierarchical clustering algorithm.

RESULTS

The method was applied to MEG data from 4 patients with localization-related epilepsy. For each patient, distinct spike subpopulations were found with clearly different topographical field maps. Inverse computations to selected spike subaverages yielded source solutions in agreement with seizure classification and location of structural lesions, if present, on magnetic resonance images.

CONCLUSIONS

With the proposed method a reliable categorization of epileptiform spikes is obtained, that can be applied in an automatic way. Computation of subaverages of similar spikes enhances the signal-to-noise ratio of spike field maps and allows for more accurate reconstruction of sources generating the epileptiform discharges.

An intervention to improve the interrater reliability of clinical EEG interpretations

Several studies have noted modest interrater reliability of clinical electroencephalogram (EEG) interpretations. Moreover, no study to date has investigated a means to improve the observed interrater agreement. The purpose of the present study was to examine (i). the interrater reliability of EEG interpretations among three raters (two psychiatrists and one pediatrician); and (ii). how to improve the reliability by establishing a consensus guideline for EEG interpretation. Three raters, two psychiatrists and a pediatrician, interpreted 100 consecutive EEG recorded at Tajimi General Hospital. After discussing the results of the first trial, the raters established a consensus guideline for EEG interpretation. They then interpreted 50 consecutive EEG recorded at Nagoya City University Hospital following this guideline. Kappa for global judgment of EEG abnormality in three grades (abnormal/borderline/normal) was 0.42 on the first and 0.63 on the second trial. Kappa significantly improved by using the guideline (P = 0.004). It is suggested that discussing and establishing the consensus guideline among the raters offers a feasible method to improve interrater reliability in clinical EEG interpretations.

Characteristics of dipoles in clustered individual spikes and averaged spikes

The aim of this study is to analyze the characteristics of dipoles in clustered individual spikes and averaged spikes, we compared electroencephalography (EEG) dipole localizations from patients with intractable extratemporal lobe epilepsy (IETLE) and from patients with benign epilepsy with centrotemporal spikes (BECTS). We studied 10 patients; five with IETLE who underwent epilepsy surgery after subdural EEG and five with BECTS. We recorded 19-channel digital scalp EEGs and used clustering analysis for individual spikes to characterize interictal spikes. We selected and averaged one representative spike group at the maximum negative peak electrode. We used a single dipole method with three-shell spherical head model. We compared dipole localizations of both averaged and individual spikes.IETLE data had more identifiable spike clusters and fewer spikes in each cluster than BECTS (P<0.05). Dipole sources with goodness-of-fit >or=95% in averaged spikes were less frequent in IETLE than in BECTS (P<0.05). For IETLE, averaged spikes showed no dipoles (two patients), while individual spikes gave dipole sources reliably in the epileptic region. For BECTS, individual and averaged spike sources were clustered. More than 80% of dipoles in averaged spikes were stable, in close proximity, for prolonged periods in BECTS. More spike groups after clustering and fewer acceptable dipoles from averaged spikes in IETLE reflect variable spike activity over extensive epileptic regions. Fewer spike groups producing more acceptable dipoles in BECTS correlate with stable spike sources within the isolated epileptic central region. Characteristics of clustered interictal spikes need careful examination before the use of dipole analysis of averaged spikes for epilepsy evaluation.

Spike detection: a review and comparison of algorithms

For algorithm developers, this review details recent approaches to the problem, compares the accuracy of various algorithms, identifies common testing issues and proposes some solutions. For the algorithm user, e.g. electroencephalograph (EEG) technician or neurologist, this review provides an estimate of algorithm accuracy and comparison to that of human experts. Manuscripts dated from 1975 are reviewed. Progress since Frost's 1985 review of the state of the art is discussed. Twenty-five manuscripts are reviewed. Many novel methods have been proposed including neural networks and high-resolution frequency methods. Algorithm accuracy is less than that of experts, but the accuracy of experts is probably less than what is commonly believed. Larger record sets will be required for expert-level detection algorithms.

Modality-specific spike identification in simultaneous magnetoencephalography/electroencephalography: a methodological approach

Epileptiform spikes may have a different morphology and signal-to-noise ratio in simultaneously recorded EEGs and magnetoencephalograms (MEGs) that may lead to differences in the identification of spikes if both the modalities are presented separately. Moreover, there are no criteria for MEG spikes. It is unknown to which extent the visual assessment of MEG data yields consistent and meaningful results. Nineteen patients were selected with mesial temporal lobe epilepsy who underwent whole-head simultaneous MEG/EEG. These data were split into MEG and EEG files and were assessed independently by three observers for the occurrence of spikes. Interobserver kappa values were calculated. A mean kappa value greater than 0.5 was taken as a criterion for the presence of unequivocal spikes. Index cases from the resulting four subgroups were studied further. One patient had unequivocal spikes in both modalities, one in EEG only, one in MEG only, and one did not show any unequivocal spike. Spikes on which at least two observers agreed were then subjected to a template match algorithm to test for equal morphology and distribution. Equal spikes were averaged and electrical and magnetic field maps were plotted. Unequivocal spikes were found in both MEG and EEG in one patient, in MEG only in two patients, in EEG only in two patients, and no spikes in either modality were seen in 14 patients. In the four index patients, MEG showed 50 to 80% more spikes than EEG. After averaging identical consensus spikes, MEG spikes revealed a concomitant spike in the EEG, but the reverse was not always true. Even in the patient with MEG and EEG spikes that met all selection criteria, simultaneous field maps showed unexpected inconsistencies. In most patients with mesial temporal lobe epilepsy, there are no unequivocal spikes during MEG/EEG. In some cases, however, experienced electroencephalographers can identify MEG spikes reliably. Because of a better signal-to-noise ratio, more spikes could be identified in MEG than in EEG. Simultaneous MEG/EEG recordings do not simply ensure the best of both, but one modality may improve the identification of spikes in the other. In addition, different aspects of a complex source can be revealed. Our three-step approach to combined data ensures a reproducible selection of spikes for source modeling.