Automated Detector of High Frequency Oscillations in Epilepsy Based on Maximum Distributed Peak Points

Optimizing automated detection of high frequency oscillations using visual markings does not improve SOZ localization

OBJECTIVE

High frequency oscillations (HFOs) are a biomarker of the seizure onset zone (SOZ) and can be visually or automatically detected. In theory, one can optimize an automated algorithm's parameters to maximize SOZ localization accuracy; however, there is no consensus on whether or how this should be done. Therefore, we optimized an automated detector using visually identified HFOs and evaluated the impact on SOZ localization accuracy.

METHODS

We detected HFOs in intracranial EEG from 20 patients with refractory epilepsy from two centers using (1) unoptimized automated detection, (2) visual identification, and (3) automated detection optimized to match visually detected HFOs.

RESULTS

SOZ localization accuracy based on HFO rate was not significantly different between the three methods. Across patients, visually optimized detector settings varied, and no single set of settings produced universally accurate SOZ localization. Exploratory analysis suggests that, for many patients, detection settings exist that would improve SOZ localization.

CONCLUSIONS

SOZ localization accuracy was similar for all three methods, was not improved by visually optimizing detector settings, and may benefit from patient-specific parameter optimization.

SIGNIFICANCE

Visual HFO marking is laborious, and optimizing automated detection using visual markings does not improve localization accuracy. New patient-specific detector optimization methods are needed.

Automatic Detection of Scalp High-Frequency Oscillations Based on Deep Learning

Scalp high-frequency oscillations (sHFOs) are a promising non-invasive biomarker of epilepsy. However, the visual marking of sHFOs is a time-consuming and subjective process, existing automatic detectors based on single-dimensional analysis have difficulty with accurately eliminating artifacts and thus do not provide sufficient reliability to meet clinical needs. Therefore, we propose a high-performance sHFOs detector based on a deep learning algorithm. An initial detection module was designed to extract candidate high-frequency oscillations. Then, one-dimensional (1D) and two-dimensional (2D) deep learning models were designed, respectively. Finally, the weighted voting method is used to combine the outputs of the two model. In experiments, the precision, recall, specificity and F1-score were 83.44%, 83.60%, 96.61% and 83.42%, respectively, on average and the kappa coefficient was 80.02%. In addition, the proposed detector showed a stable performance on multi-centre datasets. Our sHFOs detector demonstrated high robustness and generalisation ability, which indicates its potential applicability as a clinical assistance tool. The proposed sHFOs detector achieves an accurate and robust method via deep learning algorithm.

Epileptic tissue localization using graph-based networks in the high frequency oscillation range of intracranial electroencephalography

PURPOSE

High frequency oscillations (HFOs) are an emerging biomarker of epilepsy. However, very few studies have investigated the functional connectivity of interictal iEEG signals in the frequency range of HFOs. Here, we study the corresponding functional networks using graph theory, and we assess their predictive value for automatic electrode classification in a cohort of 20 drug resistant patients.

METHODS

Coherence-based connectivity analysis was performed on the iEEG recordings, and six different local graph measures were computed in both sub-bands of the HFO frequency range (80-250 Hz and 250-500 Hz). Correlation analysis was implemented between the local graph measures and the ripple and fast ripple rates. Finally, the WEKA software was employed for training and testing different predictive models on the aforementioned local graph measures.

RESULTS

The ripple rate was significantly correlated with five out of six local graph measures in the functional network. For fast ripples, their rate was also significantly (but negatively) correlated with most of the local metrics. The results from WEKA showed that the Logistic Regression algorithm was able to classify highly HFO-contaminated electrodes with an accuracy of 82.5 % for ripples and 75.4 % for fast ripples.

CONCLUSION

Functional connectivity networks in the HFO band could represent an alternative to the direct use of distinct HFO events, while also providing important insights about hub epileptic areas that can represent possible surgical targets. Automatic electrode classification through FC-based classifiers can help bypass the burden of manual HFO annotation, providing at the same time similar amount of information about the epileptic tissue.

The association of magnetoencephalography high-frequency oscillations with epilepsy types and a ripple-based method with source-level connectivity for mapping epilepsy sources

OBJECTIVE

To explore the association between high-frequency oscillations (HFOs) and epilepsy types and to improve the accuracy of source localization.

METHODS

Magnetoencephalography (MEG) ripples of 63 drug-resistant epilepsy patients were detected. Ripple rates, distribution, spatial complexity, and the clustering coefficient of ripple channels were used for the preliminary classification of lateral temporal lobe epilepsy (LTLE), mesial temporal lobe epilepsy (MTLE), and nontemporal lobe epilepsy (NTLE), mainly frontal lobe epilepsy (FLE). Furthermore, the seizure site identification was improved using the Tucker LCMV method and source-level betweenness centrality.

RESULTS

Ripple rates were significantly higher in MTLE than in LTLE and NTLE (p < 0.05). The LTLE and MTLE were mainly distributed in the temporal lobe, followed by the parietal lobe, occipital lobe, and frontal lobe, whereas MTLE ripples were mainly distributed in the frontal lobe, then parietal lobe and occipital lobe. Nevertheless, the NTLE ripples were primarily in the frontal lobe and partially in the occipital lobe (p < 0.05). Meanwhile, the spatial complexity of NTLE was significantly higher than that of LTLE and MTLE and was lowest in MTLE (p < 0.01). However, an opposite trend was observed for the standardized clustering coefficient compared with spatial complexity (p < 0.01). Finally, the tucker algorithm showed a higher percentage of ripples at the surgical site when the betweenness centrality was added (p < 0.01).

CONCLUSION

This study demonstrated that HFO rates, distribution, spatial complexity, and clustering coefficient of ripple channels varied considerably among the three epilepsy types. Additionally, tucker MEG estimation combined with ripple rates based on the source-level functional connectivity is a promising approach for presurgical epilepsy evaluation.

DSCNN-LSTMs: A Lightweight and Efficient Model for Epilepsy Recognition

Epilepsy is the second most common disease of the nervous system. Because of its high disability rate and the long course of the disease, it is a worldwide medical problem and social public health problem. Therefore, the timely detection and treatment of epilepsy are very important. Currently, medical professionals use their own diagnostic experience to identify seizures by visual inspection of the electroencephalogram (EEG). Not only does it require a lot of time and effort, but the process is also very cumbersome. Machine learning-based methods have recently been proposed for epilepsy detection, which can help clinicians make rapid and correct diagnoses. However, these methods often require extracting the features of EEG signals before using the data. In addition, the selection of features often requires domain knowledge, and feature types also have a significant impact on the performance of the classifier. In this paper, a one-dimensional depthwise separable convolutional neural network and long short-term memory networks (1D DSCNN-LSTMs) model is proposed to identify epileptic seizures by autonomously extracting the features of raw EEG. On the UCI dataset, the performance of the proposed 1D DSCNN-LSTMs model is verified by cross-validation and time complexity comparison. Compared with other previous models, the experimental results show that the highest recognition rates of binary and quintuple classification are 99.57% and 81.30%, respectively. It can be concluded that the 1D DSCNN-LSTMs model proposed in this paper is an effective method to identify seizures based on EEG signals.

Identification of Epileptogenic and Non-epileptogenic High-Frequency Oscillations Using a Multi-Feature Convolutional Neural Network Model

Accurately identifying epileptogenic zone (EZ) using high-frequency oscillations (HFOs) is a challenge that must be mastered to transfer HFOs into clinical use. We analyzed the ability of a convolutional neural network (CNN) model to distinguish EZ and non-EZ HFOs. Nineteen medically intractable epilepsy patients with good surgical outcomes 2 years after surgery were studied. Five-minute interictal intracranial electroencephalogram epochs of slow-wave sleep were selected randomly. Then 5 s segments of ripples (80–200 Hz) and fast ripples (FRs, 200–500 Hz) were detected automatically. The EZs and non-EZs were identified using the surgery resection range. We innovatively converted all epochs into four types of images using two scales: original waveforms, filtered waveforms, wavelet spectrum images, and smoothed pseudo Wigner–Ville distribution (SPWVD) spectrum images. Two scales were fixed and fitted scales. We then used a CNN model to classify the HFOs into EZ and non-EZ categories. As a result, 7,000 epochs of ripples and 2,000 epochs of FRs were randomly selected from the EZ and non-EZ data for analysis. Our CNN model can distinguish EZ and non-EZ HFOs successfully. Except for original ripple waveforms, the results from CNN models that are trained using fixed-scale images are significantly better than those from models trained using fitted-scale images (p < 0.05). Of the four fixed-scale transformations, the CNN based on the adjusted SPWVD (ASPWVD) produced the best accuracies (80.89 ± 1.43% and 77.85 ± 1.61% for ripples and FRs, respectively, p < 0.05). The CNN using ASPWVD transformation images is an effective deep learning method that can be used to classify EZ and non-EZ HFOs.

Scalp-HFO indexes are biomarkers for the lateralization and localization of the epileptogenic zone in preoperative assessment

During the noninvasive evaluation phase for refractory epilepsy, the localization of the epileptogenic zone (EZ) is essential for the surgical protocols. Confirmation of laterality is required when the preoperative evaluation limits the EZ to bilateral anterior temporal lobes or bilateral frontal lobes. High-frequency oscillations (HFOs) are considered to be promising biological markers for the EZ. However, a large number of studies on HFOs stem from intracranial research. There were few quantitative measures for scalp HFOs, so we proposed a new method to quantify and analyze scalp HFOs. This method was called the "scalp-HFO index" (HI) and calculated in both the EZ and non-EZ. The calculation was based on the numbers and spectral power of scalp HFOs automatically detected. We labeled the brain lobes involved in the EZ as regions of interest (ROIs). The HIs based on the ripple numbers (n-HI) and spectral power (s-HI) were significantly higher in the ROI than in the contra-ROI (P = 0.012, P = 0.003), indicating that HIs contributed to the lateralization of EZ. The sensitivity and specificity of n-HI for the localization of the EZ were 90% and 79.58%, respectively, suggesting that n-HI was valuable in localizing the EZ. HI may contribute to the implantation strategy of invasive electrodes. However, few scalp HFOs were recorded when the EZ was located in the medial cortex region.NEW & NOTEWORTHY We proposed the scalp-high-frequency oscillation (HFO) index (HI) as a quantitative assessment method for scalp HFOs to locate the epileptogenic zone (EZ). Our results showed that the HI in regions of interest (ROIs) was significantly higher than in contra-ROIs. Sensitivity and specificity of HI based on ripple rates (n-HI) for EZ localization were 90% and 79.58%, respectively. If the n-HI of the brain region was >1.35, it was more likely to be an epileptogenic region. Clinical application of HIs as an indicator may facilitate localization of the EZ.

Automatic detection of interictal ripples on scalp EEG to evaluate the effect and prognosis of ACTH therapy in patients with infantile spasms

OBJECTIVE

We aimed to explore the feasibility of using scalp-recorded high-frequency oscillations (HFOs) to evaluate the efficacy and prognosis of adrenocorticotropic hormone (ACTH) treatment in patients with infantile spasms.

METHODS

Thirty-nine children with infantile spasms were enrolled and divided into seizure-free and non-seizure-free groups after ACTH treatment. Patients who were seizure-free were further divided into relapse and non-relapse subgroups based on the observations made during a 6-month follow-up period. Scalp ripples were detected and compared during the interictal periods before and after 2 weeks of treatment.

RESULTS

After ACTH treatment, the number and channels of ripples were significantly lower, whereas the percentage decrease in the number, spectral power, and channels of ripples was significantly higher in the seizure-free group than in the non-seizure-free group. In addition, the relapse subgroup showed higher number and spectral power and wider distribution of ripples than did the non-relapse subgroup. Changes in HFOs in terms of number, spectral power, and channel of ripples were closely related to the severity of epilepsy and can indicate disease susceptibility.

SIGNIFICANCE

Scalp HFOs can be used as an effective biomarker to monitor the effect and evaluate the prognosis of ACTH therapy in patients with infantile spasms.

Presurgical Evaluation of Epilepsy Using Resting-State MEG Functional Connectivity

Epilepsy is caused by abnormal electrical discharges (clinically identified by electrophysiological recording) in a specific part of the brain [originating in only one part of the brain, namely, the epileptogenic zone (EZ)]. Epilepsy is now defined as an archetypical hyperexcited neural network disorder. It can be investigated through the network analysis of interictal discharges, ictal discharges, and resting-state functional connectivity. Currently, there is an increasing interest in embedding resting-state connectivity analysis into the preoperative evaluation of epilepsy. Among the various neuroimaging technologies employed to achieve brain functional networks, magnetoencephalography (MEG) with the excellent temporal resolution is an ideal tool for estimating the resting-state connectivity between brain regions, which can reveal network abnormalities in epilepsy. What value does MEG resting-state functional connectivity offer for epileptic presurgical evaluation? Regarding this topic, this paper introduced the origin of MEG and the workflow of constructing source-space functional connectivity based on MEG signals. Resting-state functional connectivity abnormalities correlate with epileptogenic networks, which are defined by the brain regions involved in the production and propagation of epileptic activities. This paper reviewed the evidence of altered epileptic connectivity based on low- or high-frequency oscillations (HFOs) and the evidence of the advantage of using simultaneous MEG and intracranial electroencephalography (iEEG) recordings. More importantly, this review highlighted that MEG-based resting-state functional connectivity has the potential to predict postsurgical outcomes. In conclusion, resting-state MEG functional connectivity has made a substantial progress toward serving as a candidate biomarker included in epileptic presurgical evaluations.

Fast Ripples as a Biomarker of Epileptogenic Tuber in Tuberous Sclerosis Complex Patients Using Stereo-Electroencephalograph

Objectives: To evaluate the value of fast ripples (FRs) (200–500 Hz) recorded with stereo-electroencephalograph (SEEG) in the localization of epileptogenic tubers in patients with tuberous sclerosis complex (TSC).

Methods: Seventeen TSC patients who underwent preoperative SEEG examination and resective epilepsy surgery were retrospectively enrolled. They were divided into two groups according to the seizure control at 1-year postoperative follow-up. The occurrence frequencies of FRs were automatically counted, and the FR rate was calculated. The high FR rate was defined as FR rate ≧0.5. According to different positions, the contacts’ locations were divided into three groups: inner of the tubers, the junction region of the tubers, and out of the tubers. The influence factors of postoperative seizure freedom were also analyzed.

Results: Twelve patients reached postoperative seizure freedom at 1-year follow-up. In total, FRs were found in 24.2% of the contacts and 67.1% of the tubers in all assessed patients. There were 47 high FR rate contacts localized in the junction region of the tubers, which was 62.7% of the 75 high FR rate contacts in total and was 8.4% of the total 561 contacts localized in the junction region of the tubers. Total removal of epileptogenic tubers and total resection of the high FR rate tubers/contacts were associated with postoperative seizure freedom (P < 0.05).

Conclusion: FRs could be extensively detected in TSC patients using SEEG, and high FR rate contacts were mostly localized in the junction region of the epileptogenic tuber, which could aid in the localization of epileptogenic tubers.

Localization of the Epileptogenic Zone by Multimodal Neuroimaging and High-Frequency Oscillation

Accurate localization of the epileptogenic zone (EZ) is a key factor to obtain good surgical outcome for refractory epilepsy patients. However, no technique, so far, can precisely locate the EZ, and there are barely any reports on the combined application of multiple technologies to improve the localization accuracy of the EZ. In this study, we aimed to explore the use of a multimodal method combining PET-MRI, fluid and white matter suppression (FLAWS)—a novel MRI sequence, and high-frequency oscillation (HFO) automated analysis to delineate EZ. We retrospectively collected 15 patients with refractory epilepsy who underwent surgery and used the above three methods to detect abnormal brain areas of all patients. We compared the PET-MRI, FLAWS, and HFO results with traditional methods to evaluate their diagnostic value. The sensitivities, specificities of locating the EZ, and marking extent removed versus not removed [RatioChann(ev)] of each method were compared with surgical outcome. We also tested the possibility of using different combinations to locate the EZ. The marked areas in every patient established using each method were also compared to determine the correlations among the three methods. The results showed that PET-MRI, FLAWS, and HFOs can provide more information about potential epileptic areas than traditional methods. When detecting the EZs, the sensitivities of PET-MRI, FLAWS, and HFOs were 68.75, 53.85, and 87.50%, and the specificities were 80.00, 33.33, and 100.00%. The RatioChann(ev) of HFO-marked contacts was significantly higher in patients with good outcome than those with poor outcome (p< 0.05). When intracranial electrodes covered all the abnormal areas indicated by neuroimaging with the overlapping EZs being completely removed referred to HFO analysis, patients could reach seizure-free (p < 0.01). The periphery of the lesion marked by neuroimaging may be epileptic, but not every lesion contributes to seizures. Therefore, approaches in multimodality can detect EZ more accurately, and HFO analysis may help in defining real epileptic areas that may be missed in the neuroimaging results. The implantation of intracranial electrodes guided by non-invasive PET-MRI and FLAWS findings as well as HFO analysis would be an optimized multimodal approach for locating EZ.

A Deep Fourier Neural Network for Seizure Prediction Using Convolutional Neural Network and Ratios of Spectral Power

Epileptic seizure prediction is one of the most used therapeutic adjuvant strategies for drug-resistant epilepsy. Conventional methods usually adopt handcrafted features and manual parameter setting. The over-reliance on the expertise of specialists may lead to weak exploitation of features and low popularization of clinical application. This paper proposes a novel parameterless patient-specific method based on Fourier Neural Network (FNN), where the Fourier transform and backpropagation learning are synthesized to make the predictor more efficient and practical. The employment of FNN is the first attempt in the field of seizure prediction due to its automatic extraction of immanent spectra in epileptic signals. Despite the self-adaptive superiority of FNN, we introduce Convolutional Neural Network (CNN) to further improve its search capability in high-dimensional feature spaces. The study also develops a multi-layer module to estimate spectral power ratios of raw recordings, which optimizes the prediction by enhancing feature diversity. Based on these modules, this paper proposes a two-channel deep neural network: Fourier Ratio Convolutional Neural Network (FRCNN). To demonstrate the reliability of the model, we explain the mathematical meaning of hidden-layer neurons in FRCNN theoretically. This approach is evaluated on both intracranial and scalp EEG datasets. It shows that the predictor achieved a sensitivity of 91.2% and a false prediction rate (FPR) of 0.06[Formula: see text]h[Formula: see text] across intracranial subjects and a sensitivity of 85.4% and an FPR of 0.14[Formula: see text]h[Formula: see text] over scalp subjects. The results indicate that FRCNN enables the convenience of epilepsy treatments while preserving a high degree of precision. In the end, a detailed comparison with the previous methods demonstrates that FRCNN has achieved higher performance and generalization ability.

An Automatic HFO Detection Method Combining Visual Inspection Features with Multi-Domain Features

As an important promising biomarker, high frequency oscillations (HFOs) can be used to track epileptic activity and localize epileptogenic zones. However, visual marking of HFOs from a large amount of intracranial electroencephalogram (iEEG) data requires a great deal of time and effort from researchers, and is also very dependent on visual features and easily influenced by subjective factors. Therefore, we proposed an automatic epileptic HFO detection method based on visual features and non-intuitive multi-domain features. To eliminate the interference of continuous oscillatory activity in detected sporadic short HFO events, the iEEG signals adjacent to the detected events were set as the neighboring environmental range while the number of oscillations and the peak-valley differences were calculated as the environmental reference features. The proposed method was developed as a MatLab-based HFO detector to automatically detect HFOs in multi-channel, long-distance iEEG signals. The performance of our detector was evaluated on iEEG recordings from epileptic mice and patients with intractable epilepsy. More than 90% of the HFO events detected by this method were confirmed by experts, while the average missed-detection rate was < 10%. Compared with recent related research, the proposed method achieved a synchronous improvement of sensitivity and specificity, and a balance between low false-alarm rate and high detection rate. Detection results demonstrated that the proposed method performs well in sensitivity, specificity, and precision. As an auxiliary tool, our detector can greatly improve the efficiency of clinical experts in inspecting HFO events during the diagnosis and treatment of epilepsy.

The Automatic Detection of Seizure Based on Tensor Distance And Bayesian Linear Discriminant Analysis

Electroencephalogram (EEG) plays an important role in recording brain activity to diagnose epilepsy. However, it is not only laborious, but also not very cost effective for medical experts to manually identify the features on EEG. Therefore, automatic seizure detection in accordance with the EEG recordings is significant for the diagnosis and treatment of epilepsy. Here, a new method for detecting seizures using tensor distance (TD) is proposed. First, the time-frequency characteristics of EEG signals are obtained by wavelet transformation, and the tensor representation of EEG signals is then obtained. Tucker decomposition is used to obtain the principal components of the EEG tensor. After, the distances between different categories of EEG tensors are calculated as the EEG features. Finally, the TD features are classified through the Bayesian Linear Discriminant Analysis (Bayesian LDA) classifier. The performance of this method is measured by the sensitivity, specificity, and recognition accuracy. Results indicate 95.12% sensitivity, 97.60% specificity, 97.60% recognition accuracy, and a false detection rate of 0.76 per hour in the invasive EEG dataset, which included 566.57[Formula: see text]h of EEG recording data from 21 patients. Taken together, the results show that TD has a good detection effect for seizure classification and that this method has high computational speed and great potential for real-time diagnosis.

Detection of Epileptic Seizure Using Pretrained Deep Convolutional Neural Network and Transfer Learning

INTRODUCTION

The diagnosis of epilepsy takes a certain process, depending entirely on the attending physician. However, the human factor may cause erroneous diagnosis in the analysis of the EEG signal. In the past 2 decades, many advanced signal processing and machine learning methods have been developed for the detection of epileptic seizures. However, many of these methods require large data sets and complex operations.

METHODS

In this study, an end-to-end machine learning model is presented for detection of epileptic seizure using the pretrained deep two-dimensional convolutional neural network (CNN) and the concept of transfer learning. The EEG signal is converted directly into visual data with a spectrogram and used directly as input data.

RESULTS

The authors analyzed the results of the training of the proposed pretrained AlexNet CNN model. Both binary and ternary classifications were performed without any extra procedure such as feature extraction. By performing data set creation from short-term spectrogram graphic images, the authors were able to achieve 100% accuracy for binary classification for epileptic seizure detection and 100% for ternary classification.

DISCUSSION/CONCLUSION

The proposed automatic identification and classification model can help in the early diagnosis of epilepsy, thus providing the opportunity for effective early treatment.

Roles of Very Fast Ripple (500-1000[Formula: see text]Hz) in the Hippocampal Network During Status Epilepticus

Very fast ripples (VFRs, 500-1000[Formula: see text]Hz) are considered more specific than high-frequency oscillations (80-500[Formula: see text]Hz) as biomarkers of epileptogenic zones. Although VFRs are frequent abnormal phenomena in epileptic seizures, their functional roles remain unclear. Here, we detected the VFRs in the hippocampal network and tracked their roles during status epilepticus (SE) in rats with pilocarpine-induced temporal lobe epilepsy (TLE). All regions in the hippocampal network exhibited VFRs in the baseline, preictal, ictal and postictal states, with the ictal state containing the most VFRs. Moreover, strong phase-locking couplings existed between VFRs and slow oscillations (1-12[Formula: see text]Hz) in the ictal and postictal states for all regions. Further investigation indicated that during VFRs, the build-up of slow oscillations in the ictal state began from the temporal lobe and then spread through the whole hippocampal network via two different pathways, which might be associated with the underlying propagation of epileptiform discharges in the hippocampal network. Overall, we provide a functional description of the emergence of VFRs in the hippocampal network during SE, and we also establish that VFRs may be the physiological representation of the pathological alterations in hippocampal network activity during SE in TLE.

Automatic vs. Manual Detection of High Frequency Oscillations in Intracranial Recordings From the Human Temporal Lobe

Background: High frequency oscillations (HFOs) have attracted great interest among neuroscientists and epileptologists in recent years. Not only has their occurrence been linked to epileptogenesis, but also to physiologic processes, such as memory consolidation. There are at least two big challenges for HFO research. First, detection, when performed manually, is time consuming and prone to rater biases, but when performed automatically, it is biased by artifacts mimicking HFOs. Second, distinguishing physiologic from pathologic HFOs in patients with epilepsy is problematic. Here we automatically and manually detected HFOs in intracranial EEGs (iEEG) of patients with epilepsy, recorded during a visual memory task in order to assess the feasibility of the different detection approaches to identify task-related ripples, supporting the physiologic nature of HFOs in the temporal lobe. Methods: Ten patients with unclear seizure origin and bilaterally implanted macroelectrodes took part in a visual memory consolidation task. In addition to iEEG, scalp EEG, electrooculography (EOG), and facial electromyography (EMG) were recorded. iEEG channels contralateral to the suspected epileptogenic zone were inspected visually for HFOs. Furthermore, HFOs were marked automatically using an RMS detector and a Stockwell classifier. We compared the two detection approaches and assessed a possible link between task performance and HFO occurrence during encoding and retrieval trials. Results: HFO occurrence rates were significantly lower when events were marked manually. The automatic detection algorithm was greatly biased by filter-artifacts. Surprisingly, EOG artifacts as seen on scalp electrodes appeared to be linked to many HFOs in the iEEG. Occurrence rates could not be associated to memory performance, and we were not able to detect strictly defined "clear" ripples. Conclusion: Filtered graphoelements in the EEG are known to mimic HFOs and thus constitute a problem. So far, in invasive EEG recordings mostly technical artifacts and filtered epileptiform discharges have been considered as sources for these "false" HFOs. The data at hand suggests that even ocular artifacts might bias automatic detection in invasive recordings. Strict guidelines and standards for HFO detection are necessary in order to identify artifact-derived HFOs, especially in conditions when cognitive tasks might produce a high amount of artifacts.

Bi-Dimensional Approach Based on Transfer Learning for Alcoholism Pre-disposition Classification via EEG Signals

Recent statistics have shown that the main difficulty in detecting alcoholism is the unreliability of the information presented by patients with alcoholism; this factor confusing the early diagnosis and it can reduce the effectiveness of treatment. However, electroencephalogram (EEG) exams can provide more reliable data for analysis of this behavior. This paper proposes a new approach for the automatic diagnosis of patients with alcoholism and introduces an analysis of the EEG signals from a two-dimensional perspective according to changes in the neural activity, highlighting the influence of high and low-frequency signals. This approach uses a two-dimensional feature extraction method, as well as the application of recent Computer Vision (CV) techniques, such as Transfer Learning with Convolutional Neural Networks (CNN). The methodology to evaluate our proposal used 21 combinations of the traditional classification methods and 84 combinations of recent CNN architectures used as feature extractors combined with the following classical classifiers: Gaussian Naive Bayes, K-Nearest Neighbor (k-NN), Multilayer Perceptron (MLP), Random Forest (RF) and Support Vector Machine (SVM). CNN MobileNet combined with SVM achieved the best results in Accuracy (95.33%), Precision (95.68%), F1-Score (95.24%), and Recall (95.00%). This combination outperformed the traditional methods by up to 8%. Thus, this approach is applicable as a classification stage for computer-aided diagnoses, useful for the triage of patients, and clinical support for the early diagnosis of this disease.

Association Between Interictal High-Frequency Oscillations and Slow Wave in Refractory Focal Epilepsy With Good Surgical Outcome

High-frequency oscillations (HFOs) have been proposed as a promising biomarker of the epileptogenic zone (EZ). But accurate delineation of EZ based on HFOs is still challenging. Our study compared HFOs from EZ and non-EZ on the basis of their associations with interictal slow waves, aiming at exploring a new way to localize EZ. Nineteen medically intractable epilepsy patients with good surgical outcome were included. Five minute interictal intracranial electroencephalography (EEG) epochs of slow-wave sleep were randomly selected; then ripples (80–200 Hz), fast ripples (FRs; 200–500 Hz), and slow waves (0.1–4 Hz) were automatically analyzed. The EZ and non-EZ were identified by resection range during the surgeries. We found that both ripples and FRs superimposed more frequently on slow waves in EZ than in non-EZ (P < 0.01). Although ripples preferred to occur on the down state of slow waves in both two groups, ripples in EZ tended to be closer to the down-state peak of slow wave than in non-EZ (-174 vs. -231 ms, P = 0.008). As for FR, no statistical difference was found between the two groups (P = 0.430). Additionally, slow wave-containing ripples in EZ had a steeper slope (1.7 vs. 1.5 μV/ms, P < 0.001) and wider distribution ratio (32.3 vs. 30.1%, P < 0.001) than those in the non-EZ. But for slow wave-containing FR, only a steeper slope (1.7 vs. 1.4 μV/ms, P < 0.001) was observed. Our study innovatively compared the different features of association between HFOs and slow wave in EZ and non-EZ from refractory focal epilepsy with good surgical outcome, proposing a new method to localize EZ and facilitating the surgical plan.

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.

Double-Step Machine Learning Based Procedure for HFOs Detection and Classification

The need for automatic detection and classification of high-frequency oscillations (HFOs) as biomarkers of the epileptogenic tissue is strongly felt in the clinical field. In this context, the employment of artificial intelligence methods could be the missing piece to achieve this goal. This work proposed a double-step procedure based on machine learning algorithms and tested it on an intracranial electroencephalogram (iEEG) dataset available online. The first step aimed to define the optimal length for signal segmentation, allowing for an optimal discrimination of segments with HFO relative to those without. In this case, binary classifiers have been tested on a set of energy features. The second step aimed to classify these segments into ripples, fast ripples and fast ripples occurring during ripples. Results suggest that LDA applied to 10 ms segmentation could provide the highest sensitivity (0.874) and 0.776 specificity for the discrimination of HFOs from no-HFO segments. Regarding the three-class classification, non-linear methods provided the highest values (around 90%) in terms of specificity and sensitivity, significantly different to the other three employed algorithms. Therefore, this machine-learning-based procedure could help clinicians to automatically reduce the quantity of irrelevant data.

Automatic Seizure Detection using Fully Convolutional Nested LSTM

The automatic seizure detection system can effectively help doctors to monitor and diagnose epilepsy thus reducing their workload. Many outstanding studies have given good results in the two-class seizure detection problems, but most of them are based on hand-wrought feature extraction. This study proposes an end-to-end automatic seizure detection system based on deep learning, which does not require heavy preprocessing on the EEG data or feature engineering. The fully convolutional network with three convolution blocks is first used to learn the expressive seizure characteristics from EEG data. Then these robust EEG features pertinent to seizures are presented as an input to the Nested Long Short-Term Memory (NLSTM) model to explore the inherent temporal dependencies in EEG signals. Lastly, the high-level features obtained from the NLSTM model are fed into the softmax layer to output predicted labels. The proposed method yields an accuracy range of 98.44–100% in 10 different experiments based on the Bonn University database. A larger EEG database is then used to evaluate the performance of the proposed method in real-life situations. The average sensitivity of 97.47%, specificity of 96.17%, and false detection rate of 0.487 per hour are yielded. For CHB–MIT Scalp EEG database, the proposed model also achieves a segment-level sensitivity of 94.07% with a false detection rate of 0.66 per hour. The excellent results obtained on three different EEG databases demonstrate that the proposed method has good robustness and generalization power under ideal and real-life conditions.

Automated Detection of High-Frequency Oscillations in Epilepsy Based on a Convolutional Neural Network

Epilepsy is one of the most common chronic neurological diseases. High-frequency oscillations (HFOs) have emerged as promising biomarkers for the epileptogenic zone. However, visual marking of HFOs is a time-consuming and laborious process. Several automated techniques have been proposed to detect HFOs, yet these are still far from being suitable for application in a clinical setting. Here, ripples and fast ripples from intracranial electroencephalograms were detected in six patients with intractable epilepsy using a convolutional neural network (CNN) method. This approach proved more accurate than using four other HFO detectors integrated in RIPPLELAB, providing a higher sensitivity (77.04% for ripples and 83.23% for fast ripples) and specificity (72.27% for ripples and 79.36% for fast ripples) for HFO detection. Furthermore, for one patient, the Cohen's kappa coefficients comparing automated detection and visual analysis results were 0.541 for ripples and 0.777 for fast ripples. Hence, our automated detector was capable of reliable estimates of ripples and fast ripples with higher sensitivity and specificity than four other HFO detectors. Our detector may be used to assist clinicians in locating epileptogenic zone in the future.

Localization of the Epileptogenic Zone Using High Frequency Oscillations

For patients with drug-resistant focal epilepsy, surgery is the therapy of choice in order to achieve seizure freedom. Epilepsy surgery foremost requires the identification of the epileptogenic zone (EZ), defined as the brain area indispensable for seizure generation. The current gold standard for identification of the EZ is the seizure-onset zone (SOZ). The fact, however that surgical outcomes are unfavorable in 40-50% of well-selected patients, suggests that the SOZ is a suboptimal biomarker of the EZ, and that new biomarkers resulting in better postsurgical outcomes are needed. Research of recent years suggested that high-frequency oscillations (HFOs) are a promising biomarker of the EZ, with a potential to improve surgical success in patients with drug-resistant epilepsy without the need to record seizures. Nonetheless, in order to establish HFOs as a clinical biomarker, the following issues need to be addressed. First, evidence on HFOs as a clinically relevant biomarker stems predominantly from retrospective assessments with visual marking, leading to problems of reproducibility and reliability. Prospective assessments of the use of HFOs for surgery planning using automatic detection of HFOs are needed in order to determine their clinical value. Second, disentangling physiologic from pathologic HFOs is still an unsolved issue. Considering the appearance and the topographic location of presumed physiologic HFOs could be immanent for the interpretation of HFO findings in a clinical context. Third, recording HFOs non-invasively via scalp electroencephalography (EEG) and magnetoencephalography (MEG) is highly desirable, as it would provide us with the possibility to translate the use of HFOs to the scalp in a large number of patients. This article reviews the literature regarding these three issues. The first part of the article focuses on the clinical value of invasively recorded HFOs in localizing the EZ, the detection of HFOs, as well as their separation from physiologic HFOs. The second part of the article focuses on the current state of the literature regarding non-invasively recorded HFOs with emphasis on findings and technical considerations regarding their localization.

Determining the Quantitative Threshold of High-Frequency Oscillation Distribution to Delineate the Epileptogenic Zone by Automated Detection

Objective: We proposed an improved automated high frequency oscillations (HFOs) detector that could not only be applied to various intracranial electrodes, but also automatically remove false HFOs caused by high-pass filtering. We proposed a continuous resection ratio of high order HFO channels and compared this ratio with each patient's post-surgical outcome, to determine the quantitative threshold of HFO distribution to delineate the epileptogenic zone (EZ). Methods: We enrolled a total of 43 patients diagnosed with refractory epilepsy. The patients were used to optimize the parameters for SEEG electrodes, to test the algorithm for identifying false HFOs, and to calculate the continuous resection ratio of high order HFO channels. The ratio can be used to determine a quantitative threshold to locate the epileptogenic zone. Results: Following optimization, the sensitivity, and specificity of our detector were 66.84 and 73.20% (ripples) and 69.76 and 66.13% (fast ripples, FRs), respectively. The sensitivity and specificity of our algorithm for removing false HFOs were 76.82 and 94.54% (ripples) and 72.55 and 94.87% (FRs), respectively. The median of the continuous resection ratio of high order HFO channels in patients with good surgical outcomes, was significantly higher than in patients with poor outcome, for both ripples and FRs (P < 0.05 ripples and P < 0.001 FRs). Conclusions: Our automated detector has the advantage of not only applying to various intracranial electrodes but also removing false HFOs. Based on the continuous resection ratio of high order HFO channels, we can set the quantitative threshold for locating epileptogenic zones.