A skew-based method for identifying intracranial EEG channels with epileptic activity without detecting spikes, ripples, or fast ripples

Adaptive node feature extraction in graph-based neural networks for brain diseases diagnosis using self-supervised learning

Electroencephalography (EEG) has demonstrated significant value in diagnosing brain diseases. In particular, brain networks have gained prominence as they offer additional valuable insights by establishing connections between EEG signal channels. While brain connections are typically delineated by channel signal similarity, there lacks a consistent and reliable strategy for ascertaining node characteristics. Conventional node features such as temporal and frequency domain properties of EEG signals prove inadequate for capturing the extensive EEG information. In our investigation, we introduce a novel adaptive method for extracting node features from EEG signals utilizing a distinctive task-induced self-supervised learning technique. By amalgamating these extracted node features with fundamental edge features constructed using Pearson correlation coefficients, we showed that the proposed approach can function as a plug-in module that can be integrated to many common GNN networks (e.g., GCN, GraphSAGE, GAT) as a replacement of node feature selections module. Comprehensive experiments are then conducted to demonstrate the consistently superior performance and high generality of the proposed method over other feature selection methods in various of brain disorder prediction tasks, such as depression, schizophrenia, and Parkinson's disease. Furthermore, compared to other node features, our approach unveils profound spatial patterns through graph pooling and structural learning, shedding light on pivotal brain regions influencing various brain disorder prediction based on derived features.

Generalisability of epileptiform patterns across time and patients

The complexity of localising the epileptogenic zone (EZ) contributes to surgical resection failures in achieving seizure freedom. The distinct patterns of epileptiform activity during interictal and ictal phases, varying across patients, often lead to suboptimal localisation using electroencephalography (EEG) features. We posed two key questions: whether neural signals reflecting epileptogenicity generalise from interictal to ictal time windows within each patient, and whether epileptiform patterns generalise across patients. Utilising an intracranial EEG dataset from 55 patients, we extracted a large battery of simple to complex features from stereo-EEG (SEEG) and electrocorticographic (ECoG) neural signals during interictal and ictal windows. Our features (n = 34) quantified many aspects of the signals including statistical moments, complexities, frequency-domain and cross-channel network attributes. Decision tree classifiers were then trained and tested on distinct time windows and patients to evaluate the generalisability of epileptogenic patterns across time and patients, respectively. Evidence strongly supported generalisability from interictal to ictal time windows across patients, particularly in signal power and high-frequency network-based features. Consistent patterns of epileptogenicity were observed across time windows within most patients, and signal features of epileptogenic regions generalised across patients, with higher generalisability in the ictal window. Signal complexity features were particularly contributory in cross-patient generalisation across patients. These findings offer insights into generalisable features of epileptic neural activity across time and patients, with implications for future automated approaches to supplement other EZ localisation methods.

Neurophysiology, Neuropsychology, and Epilepsy, in 2022: Hills We Have Climbed and Hills Ahead. Neurophysiology in epilepsy

Since the discovery of the human electroencephalogram (EEG), neurophysiology techniques have become indispensable tools in our armamentarium to localize epileptic seizures. New signal analysis techniques and the prospects of artificial intelligence and big data will offer unprecedented opportunities to further advance the field in the near future, ultimately resulting in improved quality of life for many patients with drug-resistant epilepsy. This article summarizes selected presentations from Day 1 of the two-day symposium "Neurophysiology, Neuropsychology, Epilepsy, 2022: Hills We Have Climbed and the Hills Ahead". Day 1 was dedicated to highlighting and honoring the work of Dr. Jean Gotman, a pioneer in EEG, intracranial EEG, simultaneous EEG/ functional magnetic resonance imaging, and signal analysis of epilepsy. The program focused on two main research directions of Dr. Gotman, and was dedicated to "High-frequency oscillations, a new biomarker of epilepsy" and "Probing the epileptic focus from inside and outside". All talks were presented by colleagues and former trainees of Dr. Gotman. The extended summaries provide an overview of historical and current work in the neurophysiology of epilepsy with emphasis on novel EEG biomarkers of epilepsy and source imaging and concluded with an outlook on the future of epilepsy research, and what is needed to bring the field to the next level.

Epileptic Tissue Localization through Skewness-Based Functional Connectivity in the High-Frequency Band of Intracranial EEG

Functional connectivity analysis of intracranial electroencephalography (iEEG) plays an important role in understanding the mechanism of epilepsy and seizure dynamics. However, existing connectivity analysis is only suitable for low-frequency bands below 80 Hz. High-frequency oscillations (HFOs) and high-frequency activity (HFA) in the high-frequency band (80-500 Hz) are thought to be specific biomarkers in epileptic tissue localization. However, the transience in duration and variability of occurrence time and amplitudes of these events pose a challenge for conducting effective connectivity analysis. To deal with this problem, we proposed skewness-based functional connectivity (SFC) in the high-frequency band and explored its utility in epileptic tissue localization and surgical outcome evaluation. SFC comprises three main steps. The first step is the quantitative measurement of amplitude distribution asymmetry between HFOs/HFA and baseline activity. The second step is functional network construction on the basis of rank correlation of asymmetry across time. The third step is connectivity strength extraction from the functional network. Experiments were conducted in two separate datasets which consist of iEEG recordings from 59 patients with drug-resistant epilepsy. Significant difference (p<0.001) in connectivity strength was found between epileptic and non-epileptic tissue. Results were quantified via the receiver operating characteristic curve and the area under the curve (AUC). Compared with low-frequency bands, SFC demonstrated superior performance. With respect to pooled and individual epileptic tissue localization for seizure-free patients, AUCs were 0.66 (95% confidence interval (CI): 0.63-0.69) and (0.63 95% CI 0.56-0.71), respectively. For surgical outcome classification, the AUC was 0.75 (95% CI 0.59-0.85). Therefore, SFC can act as a promising assessment tool in characterizing the epileptic network and potentially provide better treatment options for patients with drug-resistant epilepsy.

Automated Detection of Seizure Types from the Higher-Order Moments of Maximal Overlap Wavelet Distribution

In this work, an attempt has been made to develop an automated system for detecting electroclinical seizures such as tonic-clonic seizures, complex partial seizures, and electrographic seizures (EGSZ) using higher-order moments of scalp electroencephalography (EEG). The scalp EEGs of the publicly available Temple University database are utilized in this study. The higher-order moments, namely skewness and kurtosis, are extracted from the temporal, spectral, and maximal overlap wavelet distributions of EEG. The features are computed from overlapping and non-overlapping moving windowing functions. The results show that the wavelet and spectral skewness of EEG is higher in EGSZ than in other types. All the extracted features are found to have significant differences (p < 0.05), except for temporal kurtosis and skewness. A support vector machine with a radial basis kernel designed using maximal overlap wavelet skewness yields a maximum accuracy of 87%. In order to improve the performance, the Bayesian optimization technique is utilized to determine the suitable kernel parameters. The optimized model achieves the highest accuracy of 96% and an MCC of 91% in three-class classification. The study is found to be promising, and it could facilitate the rapid identification process of life-threatening seizures.

Interictal localization of the epileptogenic zone: Utilizing the observed resonance behavior in the spectral band of surrounding inhibition

Introduction

The gold standard for identification of the epileptogenic zone (EZ) continues to be the visual inspection of electrographic changes around seizures' onset by experienced electroencephalography (EEG) readers. Development of an epileptogenic focus localization tool that can delineate the EZ from analysis of interictal (seizure-free) periods is still an open question of great significance for improved diagnosis (e.g., presurgical evaluation) and treatment of epilepsy (e.g., surgical outcome).

Methods

We developed an EZ interictal localization algorithm (EZILA) based on novel analysis of intracranial EEG (iEEG) using a univariate periodogram-type power measure, a straight-forward ranking approach, a robust dimensional reduction method and a clustering technique. Ten patients with temporal and extra temporal lobe epilepsies, and matching the inclusion criteria of having iEEG recordings at the epilepsy monitoring unit (EMU) and being Engel Class I ≥12 months post-surgery, were recruited in this study.

Results

In a nested k-fold cross validation statistical framework, EZILA assigned the highest score to iEEG channels within the EZ in all patients (10/10) during the first hour of the iEEG recordings and up to their first typical clinical seizure in the EMU (i.e., early interictal period). To further validate EZILA's performance, data from two new (Engel Class I) patients were analyzed in a double-blinded fashion; the EZILA successfully localized iEEG channels within the EZ from interictal iEEG in both patients.

Discussion

Out of the sampled brain regions, iEEG channels in the EZ were most frequently and maximally active in seizure-free (interictal) periods across patients in specific narrow gamma frequency band (∼60-80 Hz), which we have termed focal frequency band (FFB). These findings are consistent with the hypothesis that the EZ may interictally be regulated (controlled) by surrounding inhibitory neurons with resonance characteristics within this narrow gamma band.

Improving automated diagnosis of epilepsy from EEGs beyond IEDs

Objective.Clinical diagnosis of epilepsy relies partially on identifying interictal epileptiform discharges (IEDs) in scalp electroencephalograms (EEGs). This process is expert-biased, tedious, and can delay the diagnosis procedure. Beyond automatically detecting IEDs, there are far fewer studies on automated methods to differentiate epileptic EEGs (potentially without IEDs) from normal EEGs. In addition, the diagnosis of epilepsy based on a single EEG tends to be low. Consequently, there is a strong need for automated systems for EEG interpretation. Traditionally, epilepsy diagnosis relies heavily on IEDs. However, since not all epileptic EEGs exhibit IEDs, it is essential to explore IED-independent EEG measures for epilepsy diagnosis. The main objective is to develop an automated system for detecting epileptic EEGs, both with or without IEDs. In order to detect epileptic EEGs without IEDs, it is crucial to include EEG features in the algorithm that are not directly related to IEDs.Approach.In this study, we explore the background characteristics of interictal EEG for automated and more reliable diagnosis of epilepsy. Specifically, we investigate features based on univariate temporal measures (UTMs), spectral, wavelet, Stockwell, connectivity, and graph metrics of EEGs, besides patient-related information (age and vigilance state). The evaluation is performed on a sizeable cohort of routine scalp EEGs (685 epileptic EEGs and 1229 normal EEGs) from five centers across Singapore, USA, and India.Main results.In comparison with the current literature, we obtained an improved Leave-One-Subject-Out (LOSO) cross-validation (CV) area under the curve (AUC) of 0.871 (Balanced Accuracy (BAC) of 80.9%) with a combination of three features (IED rate, and Daubechies and Morlet wavelets) for the classification of EEGs with IEDs vs. normal EEGs. The IED-independent feature UTM achieved a LOSO CV AUC of 0.809 (BAC of 74.4%). The inclusion of IED-independent features also helps to improve the EEG-level classification of epileptic EEGs with and without IEDs vs. normal EEGs, achieving an AUC of 0.822 (BAC of 77.6%) compared to 0.688 (BAC of 59.6%) for classification only based on the IED rate. Specifically, the addition of IED-independent features improved the BAC by 21% in detecting epileptic EEGs that do not contain IEDs.Significance.These results pave the way towards automated detection of epilepsy. We are one of the first to analyze epileptic EEGs without IEDs, thereby opening up an underexplored option in epilepsy diagnosis.

Classification of Depth of Coma Using Complexity Measures and Nonlinear Features of Electroencephalogram Signals

In recent years, some electrophysiological analysis methods of consciousness have been proposed. Most of these studies are based on visual interpretation or statistical analysis, and there is hardly any work classifying the level of consciousness in a deep coma. In this study, we perform an analysis of electroencephalography complexity measures by quantifying features efficiency in differentiating patients in different consciousness levels. Several measures of complexity have been proposed to quantify the complexity of signals. Our aim is to lay the foundation of a system that will objectively define the level of consciousness by performing a complexity analysis of Electroencephalogram (EEG) signals. Therefore, a nonlinear analysis of EEG signals obtained with a recording scheme proposed by us from 39 patients with Glasgow Coma Scale (GCS) between 3 and 8 was performed. Various entropy values (approximate entropy, permutation entropy, etc.) obtained from different algorithms, Hjorth parameters, Lempel-Ziv complexity and Kolmogorov complexity values were extracted from the signals as features. The features were analyzed statistically and the success of features in classifying different levels of consciousness was measured by various classifiers. Consequently, levels of consciousness in deep coma (GCS between 3 and 8) were classified with an accuracy of 90.3%. To the authors' best knowledge, this is the first demonstration of the discriminative nonlinear features extracted from tactile and auditory stimuli EEG signals in distinguishing different GCSs of comatose patients.

Interictal high frequency background activity as a biomarker of epileptogenic tissue

High frequency oscillations (HFOs) are very brief events that are a well-established biomarker of the epileptogenic zone (EZ) but are rare and comprise only a tiny fraction of the total recorded EEG. We hypothesize that the interictal high frequency ‘background’ data, which has received little attention but represents the majority of the EEG record, also may contain additional, novel information for identifying the EZ. We analysed intracranial EEG (30–500 Hz frequency range) acquired from 24 patients who underwent resective surgery. We computed 38 quantitative features based on all usable, interictal data (63–307 h per subject), excluding all detected HFOs. We assessed association between each feature and the seizure onset zone (SOZ) and resected volume (RV) using logistic regression. A pathology score per channel was also created via principle component analysis and logistic regression, using hold-out-one-patient cross-validation to avoid in-sample training. Association of the pathology score with the SOZ and RV was quantified using an asymmetry measure. Many features were associated with the SOZ: 23/38 features had odds ratios >1.3 or <0.7 and 17/38 had odds ratios different than zero with high significance (P < 0.001/39, logistic regression with Bonferroni Correction). The pathology score, the rate of HFOs, and their channel-wise product were each strongly associated with the SOZ [median asymmetry ≥0.44, good surgery outcome patients; median asymmetry ≥0.40, patients with other outcomes; 95% confidence interval (CI) > 0.27 in both cases]. The pathology score and the channel-wise product also had higher asymmetry with respect to the SOZ than the HFO rate alone (median difference in asymmetry ≥0.18, 95% CI >0.05). These results support that the high frequency background data contains useful information for determining the EZ, distinct and complementary to information from detected HFOs. The concordance between the high frequency activity pathology score and the rate of HFOs appears to be a better biomarker of epileptic tissue than either measure alone.

Delineation of epileptogenic zones with high frequency magnetic source imaging based on kurtosis and skewness

BACKGROUND

Neuromagnetic high frequency brain signals (HFBS, > 80 Hz) are a new biomarker for localization of epileptogenic zones (EZs) for pediatric epilepsy.

METHODS

Twenty three children with drug-resistant epilepsy and age/sex matched healthy controls were studied with magnetoencephalography (MEG). Epileptic HFBS in 80-250 Hz and 250-600 Hz were quantitatively determined by comparing with normative controls in terms of kurtosis and skewness. Magnetic sources of epileptic HFBS were localized and then compared to clinical EZs determined by invasive recordings and surgical outcomes.

RESULTS

Kurtosis and skewness of HFBS were significantly elevated in epilepsy patients compared to healthy controls (p < 0,001 and p < 0.0001, respectively). Sources of elevated MEG signals in comparison to normative data were co-localized to EZs for 22 (22/23, 96 %) patients.

CONCLUSIONS

The results indicate, for the first time, that epileptic HFBS can be noninvasively quantified by measuring kurtosis and skewness in MEG data. Magnetic source imaging based on kurtosis and skewness can accurately localize EZs.

SIGNIFICANCE

Source imaging of kurtosis and skewness of MEG HFBS provides a novel way for preoperative localization of EZs for epilepsy surgery.

Distinguishing false and true positive detections of high frequency oscillations

Objective.

High frequency oscillations (HFOs) are a promising biomarker of tissue that instigates seizures. However, ambiguous data and random background fluctuations can cause any HFO detector (human or automated) to falsely label non-HFO data as an HFO (a false positive detection). The objective of this paper was to identify quantitative features of HFOs that distinguish between true and false positive detections.

Approach.

Feature selection was performed using background data in multi-day, interictal intracranial recordings from ten patients. We selected the feature most similar between randomly selected segments of background data and HFOs detected in surrogate background data (false positive detections by construction). We then compared these results with fuzzy clustering of detected HFOs in clinical data to verify the feature’s applicability. We validated the feature is sensitive to false versus true positive HFO detections by using an independent data set (six subjects) scored for HFOs by three human reviewers. Lastly, we compared the effect of redacting putative false positive HFO detections on the distribution of HFOs across channels and their association with seizure onset zone (SOZ) and resected volume (RV).

Main results.

Of the 15 analyzed features, the analysis selected only skewness of the curvature (skewCurve). The feature was validated in human scored data to be associated with distinguishing true and false positive HFO detections. Automated HFO detections with higher skewCurve were more focal based on entropy measures and had increased localization to both the SOZ and RV.

Significance.

We identified a quantitative feature of HFOs which helps distinguish between true and false positive detections. Redacting putative false positive HFO detections improves the specificity of HFOs as a biomarker of epileptic tissue.

Non-Gaussianity Detection of EEG Signals Based on a Multivariate Scale Mixture Model for Diagnosis of Epileptic Seizures

OBJECTIVE

The detection of epileptic seizures from scalp electroencephalogram (EEG) signals can facilitate early diagnosis and treatment. Previous studies suggested that the Gaussianity of EEG distributions changes depending on the presence or absence of seizures; however, no general EEG signal models can explain such changes in distributions within a unified scheme.

METHODS

This article describes the formulation of a stochastic EEG model based on a multivariate scale mixture distribution that can represent changes in non-Gaussianity caused by stochastic fluctuations in EEG. In addition, we propose an EEG analysis method by combining the model with a filter bank and introduce a feature representing the non-Gaussianity latent in each EEG frequency band.

RESULTS

We applied the proposed method to multichannel EEG data from twenty patients with focal epilepsy. The results showed a significant increase in the proposed feature during epileptic seizures, particularly in the high-frequency band. The feature calculated in the high-frequency band allowed highly accurate classification of seizure and non-seizure segments [area under the receiver operating characteristic curve (AUC) = 0.881] using only a simple threshold.

CONCLUSION

This article proposed a multivariate scale mixture distribution-based stochastic EEG model capable of representing non-Gaussianity associated with epileptic seizures. Experiments using simulated and real EEG data demonstrated the validity of the model and its applicability to epileptic seizure detection.

SIGNIFICANCE

The stochastic fluctuations of EEG quantified by the proposed model can help detect epileptic seizures with high accuracy.

Trends in the use of automated algorithms for the detection of high-frequency oscillations associated with human epilepsy

High-frequency oscillations (HFOs) in intracranial electroencephalography (EEG) are a promising biomarker of the epileptogenic zone and tool for surgical planning. Many studies have shown that a high rate of HFOs (number per minute) is correlated with the seizure-onset zone, and complete removal of HFO-generating brain regions has been associated with seizure-free outcome after surgery. In order to use HFOs as a biomarker, these transient events must first be detected in electrophysiological data. Because visual detection of HFOs is time-consuming and subject to low interrater reliability, many automated algorithms have been developed, and they are being used increasingly for such studies. However, there is little guidance on how to select an algorithm, implement it in a clinical setting, and validate the performance. Therefore, we aim to review automated HFO detection algorithms, focusing on conceptual similarities and differences between them. We summarize the standard steps for data pre-processing, as well as post-processing strategies for rejection of false-positive detections. We also detail four methods for algorithm testing and validation, and we describe the specific goal achieved by each one. We briefly review direct comparisons of automated algorithms applied to the same data set, emphasizing the importance of optimizing detection parameters. Then, to assess trends in the use of automated algorithms and their potential for use in clinical studies, we review evidence for the relationship between automatically detected HFOs and surgical outcome. We conclude with practical recommendations and propose standards for the selection, implementation, and validation of automated HFO-detection algorithms.

Epileptogenicity Mapping: A Quantitative Approach to Identify the Seizure Onset

During the presurgical evaluation of patients with focal refractory epilepsies, the spatial mapping of the seizure onset zone (SOZ) and seizure propagation networks critically depends on the use of different features extracted from the intracranial electroencephalogram (IEEG). The identification of the SOZ is usually based on visual inspection by highly qualified neurophysiologists. However, quantitative IEEG analyses have recently been developed by exploiting signal and image characteristics in order to improve and expedite the SOZ detection. Here, the authors briefly review some of the latest methods proposed by different research groups and then present the recent implementation in Brainstorm software.

Kurtosis and skewness of high-frequency brain signals are altered in paediatric epilepsy

Intracranial studies provide solid evidence that high-frequency brain signals are a new biomarker for epilepsy. Unfortunately, epileptic (pathological) high-frequency signals can be intermingled with physiological high-frequency signals making these signals difficult to differentiate. Recent success in non-invasive detection of high-frequency brain signals opens a new avenue for distinguishing pathological from physiological high-frequency signals. The objective of the present study is to characterize pathological and physiological high-frequency signals at source levels by using kurtosis and skewness analyses. Twenty-three children with medically intractable epilepsy and age-/gender-matched healthy controls were studied using magnetoencephalography. Magnetoencephalographic data in three frequency bands, which included 2–80 Hz (the conventional low-frequency signals), 80–250 Hz (ripples) and 250–600 Hz (fast ripples), were analysed. The kurtosis and skewness of virtual electrode signals in eight brain regions, which included left/right frontal, temporal, parietal and occipital cortices, were calculated and analysed. Differences between epilepsy and controls were quantitatively compared for each cerebral lobe in each frequency band in terms of kurtosis and skewness measurements. Virtual electrode signals from clinical epileptogenic zones and brain areas outside of the epileptogenic zones were also compared with kurtosis and skewness analyses. Compared to controls, patients with epilepsy showed significant elevation in kurtosis and skewness of virtual electrode signals. The spatial and frequency patterns of the kurtosis and skewness of virtual electrode signals among the eight cerebral lobes in three frequency bands were also significantly different from that of the controls (2–80 Hz, P < 0.001; 80–250 Hz, P < 0.00001; 250–600 Hz, P < 0.0001). Compared to signals from non-epileptogenic zones, virtual electrode signals from epileptogenic zones showed significantly altered kurtosis and skewness (P < 0.001). Compared to normative data from the control group, aberrant virtual electrode signals were, for each patient, more pronounced in the epileptogenic lobes than in other lobes(kurtosis analysis of virtual electrode signals in 250–600 Hz; odds ratio = 27.9; P < 0.0001). The kurtosis values of virtual electrode signals in 80–250 and 250–600 Hz showed the highest sensitivity (88.23%) and specificity (89.09%) for revealing epileptogenic lobe, respectively. The combination of virtual electrode and kurtosis/skewness measurements provides a new quantitative approach to distinguishing pathological from physiological high-frequency signals for paediatric epilepsy. Non-invasive identification of pathological high-frequency signals may provide novel important information to guide clinical invasive recordings and direct surgical treatment of epilepsy.