Accurate identification of EEG recordings with interictal epileptiform discharges using a hybrid approach: Artificial intelligence supervised by human experts

A randomized controlled trial on evaluating clinician-supervised generative AI for decision support

BACKGROUND

The integration of generative artificial intelligence (AI) as clinical decision support systems (CDSS) into telemedicine presents a significant opportunity to enhance clinical outcomes, yet its application remains underexplored.

OBJECTIVE

This study investigates the efficacy of one of the most common generative AI tools, ChatGPT, for providing clinical guidance during cardiac arrest scenarios.

METHODS

We examined the performance, cognitive load, and trust associated with traditional methods (paper guide), autonomous ChatGPT, and clinician-supervised ChatGPT, where a clinician supervised the AI recommendations. Fifty-four subjects without medical backgrounds participated in randomized controlled trials, each assigned to one of three intervention groups: paper guide, ChatGPT, or supervised ChatGPT. Participants completed a standardized CPR scenario using an Augmented Reality (AR) headset, and performance, physiological, and self-reported metrics were recorded.

MAIN FINDINGS

Results indicate that the Supervised-ChatGPT group showed significantly higher decision accuracy compared to the paper guide and ChatGPT groups, although the scenario completion time was longer. Physiological data showed a reduced LF/HF ratio in the Supervised-ChatGPT group, suggesting potentially lower cognitive load. Trust in AI was also highest in the supervised condition. In one instance, ChatGPT suggested a risky option, highlighting the need for clinician supervision.

CONCLUSION

Our findings highlight the potential of supervised generative AI to enhance decision-making accuracy and user trust in emergency healthcare settings, despite trade-offs with response time. The study underscores the importance of clinician oversight and the need for further refinement of AI systems to improve safety. Future research should explore strategies to optimize AI supervision and assess the implementation of these systems in real-world clinical settings.

Detecting fast-ripples on both micro- and macro-electrodes in epilepsy: A wavelet-based CNN detector

BACKGROUND

Fast-ripples (FR) are short (∼10 ms) high-frequency oscillations (HFO) between 200 and 600 Hz that are helpful in epilepsy to identify the epileptogenic zone. Our aim is to propose a new method to detect FR that had to be efficient for intracerebral EEG (iEEG) recorded from both usual clinical macro-contacts (millimeter scale) and microwires (micrometer scale).

NEW METHOD

Step 1 of the detection method is based on a convolutional neural network (CNN) trained using a large database of > 11,000 FR recorded from the iEEG of 38 patients with epilepsy from both macro-contacts and microwires. The FR and non-FR events were fed to the CNN as normalized time-frequency maps. Step 2 is based on feature-based control techniques in order to reject false positives. In step 3, the human is reinstated in the decision-making process for final validation using a graphical user interface.

RESULTS

WALFRID achieved high performance on the realistically simulated data with sensitivity up to 99.95 % and precision up to 96.51 %. The detector was able to adapt to both macro and micro-EEG recordings. The real data was used without any pre-processing step such as artefact rejection. The precision of the automatic detection was of 57.5. Step 3 helped eliminating remaining false positives in a few minutes per subject.

COMPARISON WITH EXISTING METHODS

WALFRID performed as well or better than 6 other existing methods.

CONCLUSION

Since WALFRID was created to mimic the work-up of the neurologist, clinicians can easily use, understand, interpret and, if necessary, correct the output.

Of Pilots and Copilots: The Evolving Role of Artificial Intelligence in Clinical Neurophysiology

Artificial intelligence (AI) is revolutionizing clinical neurophysiology (CNP), particularly in its applications to electroencephalography (EEG), electromyography (EMG), and polysomnography (PSG). AI enhances diagnostic accuracy and efficiency while addressing interrater variability and the growing data volume. The evolution of AI tools, from early mimetic methods to advanced deep learning techniques, has significantly improved spike and seizure detection in EEG and facilitated whole EEG evaluations, reducing the workload on clinicians. In EMG, AI demonstrates promise in identifying motor unit abnormalities and analyzing audio signals, though challenges persist due to limited datasets and clinical context considerations. PSG scoring has seen substantial integration of AI, with systems achieving high accuracy through uncertainty estimation and selective manual review, but limitations remain in analyzing epileptic activity and classifying certain sleep stages. As a "co-pilot," AI augments human expertise by improving quality control, standardizing clinical trials, and enabling rapid data review, particularly for less experienced providers. Future AI advancements in CNP aim to shift from isolated data interpretation to providing clinical context, considering patient history, treatment options, and prognostic implications. While the potential of generative AI and "AI-omics" is transformative, the importance of thoughtful integration to augment rather than replace human expertise must be emphasized, ensuring that AI becomes a tool for collaboration and innovation in medicine.

Can ChatGPT 4.0 Diagnose Epilepsy? A Study on Artificial Intelligence's Diagnostic Capabilities

Objectives: This study investigates the potential of artificial intelligence (AI), specifically large language models (LLMs) like ChatGPT, to enhance decision support in diagnosing epilepsy. AI tools can improve diagnostic accuracy, efficiency, and decision-making speed. The aim of this study was to compare the level of agreement in epilepsy diagnosis between human experts (epileptologists) and AI (ChatGPT), using the 2014 International League Against Epilepsy (ILAE) criteria, and to identify potential predictors of diagnostic errors made by ChatGPT. Methods: A retrospective analysis was conducted on data from 597 patients who visited the emergency department for either a first epileptic seizure or a recurrence. Diagnoses made by experienced epileptologists were compared with those made by ChatGPT 4.0, which was trained on the 2014 ILAE epilepsy definition. The agreement between human and AI diagnoses was assessed using Cohen's kappa statistic. Sensitivity and specificity were compared using 2 × 2 contingency tables, and multivariate analyses were performed to identify variables associated with diagnostic errors. Results: Neurologists diagnosed epilepsy in 216 patients (36.2%), while ChatGPT diagnosed it in 109 patients (18.2%). The agreement between neurologists and ChatGPT was very low, with a Cohen's kappa value of -0.01 (95% confidence intervals, CI: -0.08 to 0.06). ChatGPT's sensitivity was 17.6% (95% CI: 14.5-20.6), specificity was 81.4% (95% CI: 78.2-84.5), positive predictive value was 34.8% (95% CI: 31.0-38.6), and negative predictive value was 63.5% (95% CI: 59.6-67.4). ChatGPT made diagnostic errors in 41.7% of the cases, with errors more frequent in older patients and those with specific medical conditions. The correct classification was associated with acute symptomatic seizures of unknown etiology. Conclusions: ChatGPT 4.0 does not reach human clinicians' performance in diagnosing epilepsy, showing poor performance in identifying epilepsy but better at recognizing non-epileptic cases. The overall concordance between human clinicians and AI is extremely low. Further research is needed to improve the diagnostic accuracy of ChatGPT and other LLMs.

Artificial Intelligence in Pediatric Epilepsy Detection: Balancing Effectiveness With Ethical Considerations for Welfare

Background and Aim

Epilepsy is a major neurological challenge, especially for pediatric populations. It profoundly impacts both developmental progress and quality of life in affected children. With the advent of artificial intelligence (AI), there's a growing interest in leveraging its capabilities to improve the diagnosis and management of pediatric epilepsy. This review aims to assess the effectiveness of AI in pediatric epilepsy detection while considering the ethical implications surrounding its implementation.

Methodology

A comprehensive systematic review was conducted across multiple databases including PubMed, EMBASE, Google Scholar, Scopus, and Medline. Search terms encompassed "pediatric epilepsy," "artificial intelligence," "machine learning," "ethical considerations," and "data security." Publications from the past decade were scrutinized for methodological rigor, with a focus on studies evaluating AI's efficacy in pediatric epilepsy detection and management.

Results

AI systems have demonstrated strong potential in diagnosing and monitoring pediatric epilepsy, often matching clinical accuracy. For example, AI-driven decision support achieved 93.4% accuracy in diagnosis, closely aligning with expert assessments. Specific methods, like EEG-based AI for detecting interictal discharges, showed high specificity (93.33%-96.67%) and sensitivity (76.67%-93.33%), while neuroimaging approaches using rs-fMRI and DTI reached up to 97.5% accuracy in identifying microstructural abnormalities. Deep learning models, such as CNN-LSTM, have also enhanced seizure detection from video by capturing subtle movement and expression cues. Non-EEG sensor-based methods effectively identified nocturnal seizures, offering promising support for pediatric care. However, ethical considerations around privacy, data security, and model bias remain crucial for responsible AI integration.

Conclusion

While AI holds immense potential to enhance pediatric epilepsy management, ethical considerations surrounding transparency, fairness, and data security must be rigorously addressed. Collaborative efforts among stakeholders are imperative to navigate these ethical challenges effectively, ensuring responsible AI integration and optimizing patient outcomes in pediatric epilepsy care.

Expert level of detection of interictal discharges with a deep neural network

OBJECTIVE

Deep learning methods have shown potential in automating the detection of interictal epileptiform discharges (IEDs) in electroencephalography (EEG). We compared IED detection using our previously trained deep neural network with a group of experts to assess its potential applicability.

METHODS

First, we performed clinical validation on an internal data set. Seven experts reviewed all EEG studies. Performance agreement between experts and the network was compared at both the EEG and IED levels. All EEG recordings were also processed with Persyst. Subsequently, we performed external validation, with data from four centers, using a hybrid approach, where detections by the deep neural network were reviewed by an expert. In case of disagreement with the original report, the EEG recording was annotated independently by five experts.

RESULTS

For internal validation we included 22 EEG studies with IEDs and 28 EEG studies from controls. At the EEG level, our network showed performance similar to that of the experts. For individual IED detection, the sensitivities between experts ranged from 20.7%-86.4%, whereas the sensitivity of our network was 82.5% (confidence interval [CI]: 77.7%-87.4%) at 99% specificity and a false detection rate (FDR) of <.2/min, outperforming Persyst, with 64.6% sensitivity (CI: 61.4%-67.9%) at 98% specificity. External validation in 174 EEG studies demonstrated that all 85 EEG recordings classified as normal in the original report were classified correctly, with an FDR of .10/min. Of the 89 EEG studies with IEDs according to the report, 56 were correctly classified (Cohen's κ = .62). Visual analysis of the remaining 33 EEG recordings showed high interobserver variability among the five experts (Fleiss' κ = .13).

SIGNIFICANCE

Our deep neural network detects IEDs on par with clinical experts. The external validation in a hybrid approach showed substantial agreement with the original report. Disagreement was due mainly to high interobserver variability. Our deep neural network may support visual EEG analysis and assist in diagnostics, particularly when human resources are limited.

Advancing epilepsy diagnosis: A meta-analysis of artificial intelligence approaches for interictal epileptiform discharge detection

INTRODUCTION

Interictal epileptiform discharges (IEDs) in electroencephalograms (EEGs) are an important biomarker for epilepsy. Currently, the gold standard for IED detection is the visual analysis performed by experts. However, this process is expert-biased, and time-consuming. Developing fast, accurate, and robust detection methods for IEDs based on EEG may facilitate epilepsy diagnosis. We aim to assess the performance of deep learning (DL) and classic machine learning (ML) algorithms in classifying EEG segments into IED and non-IED categories, as well as distinguishing whether the entire EEG contains IED or not.

METHODS

We systematically searched PubMed, Embase, and Web of Science following PRISMA guidelines. We excluded studies that only performed the detection of IEDs instead of binary segment classification. Risk of Bias was evaluated with Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). Meta-analysis with the overall area under the Summary Receiver Operating Characteristic (SROC), sensitivity, and specificity as effect measures, was performed with R software.

RESULTS

A total of 23 studies, comprising 3,629 patients, were eligible for synthesis. Eighteen models performed discharge-level classification, and 6 whole-EEG classification. For the IED-level classification, 3 models were validated in an external dataset with more than 50 patients and achieved a sensitivity of 84.9 % (95 % CI: 82.3-87.2) and a specificity of 68.7 % (95 % CI: 7.9-98.2). Five studies reported model performance using both internal validation (cross-validation) and external datasets. The meta-analysis revealed higher performance for internal validation, with 90.4 % sensitivity and 99.6 % specificity, compared to external validation, which showed 78.1 % sensitivity and 80.1 % specificity.

CONCLUSION

Meta-analysis showed higher performance for models validated with resampling methods compared to those using external datasets. Only a minority of models use more robust validation techniques, which often leads to overfitting.

Nonictal electroencephalographic measures for the diagnosis of functional seizures

OBJECTIVE

Functional seizures (FS) look like epileptic seizures but are characterized by a lack of epileptic activity in the brain. Approximately one in five referrals to epilepsy clinics are diagnosed with this condition. FS are diagnosed by recording a seizure using video-electroencephalography (EEG), from which an expert inspects the semiology and the EEG. However, this method can be expensive and inaccessible and can present significant patient burden. No single biomarker has been found to diagnose FS. However, the current limitations in FS diagnosis could be improved with machine learning to classify signal features extracted from EEG, thus providing a potentially very useful aid to clinicians.

METHODS

The current study has investigated the use of seizure-free EEG signals with machine learning to identify subjects with FS from those with epilepsy. The dataset included interictal and preictal EEG recordings from 48 subjects with FS (mean age = 34.76 ± 10.55 years, 14 males) and 29 subjects with epilepsy (mean age = 38.95 ± 13.93 years, 18 males) from which various statistical, temporal, and spectral features from the five EEG frequency bands were extracted then analyzed with threshold accuracy, five machine learning classifiers, and two feature importance approaches.

RESULTS

The highest classification accuracy reported from thresholding was 60.67%. However, the temporal features were the best performing, with the highest balanced accuracy reported by the machine learning models: 95.71% with all frequency bands combined and a support vector machine classifier.

SIGNIFICANCE

Machine learning was much more effective than using individual features and could be a powerful aid in FS diagnosis. Furthermore, combining the frequency bands improved the accuracy of the classifiers in most cases, and the lowest performing EEG bands were consistently delta and gamma.

Real-Time Non-Invasive Imaging and Detection of Spreading Depolarizations through EEG: An Ultra-Light Explainable Deep Learning Approach

A core aim of neurocritical care is to prevent secondary brain injury. Spreading depolarizations (SDs) have been identified as an important independent cause of secondary brain injury. SDs are usually detected using invasive electrocorticography recorded at high sampling frequency. Recent pilot studies suggest a possible utility of scalp electrodes generated electroencephalogram (EEG) for non-invasive SD detection. However, noise and attenuation of EEG signals makes this detection task extremely challenging. Previous methods focus on detecting temporal power change of EEG over a fixed high-density map of scalp electrodes, which is not always clinically feasible. Having a specialized spectrogram as an input to the automatic SD detection model, this study is the first to transform SD identification problem from a detection task on a 1-D time-series wave to a task on a sequential 2-D rendered imaging. This study presented a novel ultra-light-weight multi-modal deep-learning network to fuse EEG spectrogram imaging and temporal power vectors to enhance SD identification accuracy over each single electrode, allowing flexible EEG map and paving the way for SD detection on ultra-low-density EEG with variable electrode positioning. Our proposed model has an ultra-fast processing speed (<0.3 sec). Compared to the conventional methods (2 hours), this is a huge advancement towards early SD detection and to facilitate instant brain injury prognosis. Seeing SDs with a new dimension - frequency on spectrograms, we demonstrated that such additional dimension could improve SD detection accuracy, providing preliminary evidence to support the hypothesis that SDs may show implicit features over the frequency profile.

vEpiNet: A multimodal interictal epileptiform discharge detection method based on video and electroencephalogram data

To enhance deep learning-based automated interictal epileptiform discharge (IED) detection, this study proposes a multimodal method, vEpiNet, that leverages video and electroencephalogram (EEG) data. Datasets comprise 24 931 IED (from 484 patients) and 166 094 non-IED 4-second video-EEG segments. The video data is processed by the proposed patient detection method, with frame difference and Simple Keypoints (SKPS) capturing patients' movements. EEG data is processed with EfficientNetV2. The video and EEG features are fused via a multilayer perceptron. We developed a comparative model, termed nEpiNet, to test the effectiveness of the video feature in vEpiNet. The 10-fold cross-validation was used for testing. The 10-fold cross-validation showed high areas under the receiver operating characteristic curve (AUROC) in both models, with a slightly superior AUROC (0.9902) in vEpiNet compared to nEpiNet (0.9878). Moreover, to test the model performance in real-world scenarios, we set a prospective test dataset, containing 215 h of raw video-EEG data from 50 patients. The result shows that the vEpiNet achieves an area under the precision-recall curve (AUPRC) of 0.8623, surpassing nEpiNet's 0.8316. Incorporating video data raises precision from 70% (95% CI, 69.8%-70.2%) to 76.6% (95% CI, 74.9%-78.2%) at 80% sensitivity and reduces false positives by nearly a third, with vEpiNet processing one-hour video-EEG data in 5.7 min on average. Our findings indicate that video data can significantly improve the performance and precision of IED detection, especially in prospective real clinic testing. It suggests that vEpiNet is a clinically viable and effective tool for IED analysis in real-world applications.

Idiopathic generalized epilepsies in the epilepsy monitoring unit: Systematic quantification of focal EEG and semiological signs

OBJECTIVE

Focal seizure symptoms (FSS) and focal interictal epileptiform discharges (IEDs) are common in patients with idiopathic generalized epilepsies (IGEs), but dedicated studies systematically quantifying them both are lacking. We used automatic IED detection and localization algorithms and correlated these EEG findings with clinical FSS for the first time in IGE patients.

METHODS

32 patients with IGEs undergoing long-term video EEG monitoring were systematically analyzed regarding focal vs. generalized IEDs using automatic IED detection and localization algorithms. Quantitative EEG findings were correlated with FSS.

RESULTS

We observed FSS in 75% of patients, without significant differences between IGE subgroups. Mostly varying/shifting lateralizations of FSS across successive recorded seizures were seen. We detected a total of 81,949 IEDs, whereof 19,513 IEDs were focal (23.8%). Focal IEDs occurred in all patients (median 13% focal IEDs per patient, range 1.1 - 51.1%). Focal IED lateralization and localization predominance had no significant effect on FSS.

CONCLUSIONS

All included patients with IGE showed focal IEDs and three-quarter had focal seizure symptoms irrespective of the specific IGE subgroup. Focal IED localization had no significant effect on lateralization and localization of FSS.

SIGNIFICANCE

Our findings may facilitate diagnostic and treatment decisions in patients with suspected IGE and focal signs.

Artificial intelligence/machine learning for epilepsy and seizure diagnosis

Accurate seizure and epilepsy diagnosis remains a challenging task due to the complexity and variability of manifestations, which can lead to delayed or missed diagnosis. Machine learning (ML) and artificial intelligence (AI) is a rapidly developing field, with growing interest in integrating and applying these tools to aid clinicians facing diagnostic uncertainties. ML algorithms, particularly deep neural networks, are increasingly employed in interpreting electroencephalograms (EEG), neuroimaging, wearable data, and seizure videos. This review discusses the development and testing phases of AI/ML tools, emphasizing the importance of generalizability and interpretability in medical applications, and highlights recent publications that demonstrate the current and potential utility of AI to aid clinicians in diagnosing epilepsy. Current barriers of AI integration in patient care include dataset availability and heterogeneity, which limit studies' quality, interpretability, comparability, and generalizability. ML and AI offer substantial promise in improving the accuracy and efficiency of epilepsy diagnosis. The growing availability of diverse datasets, enhanced processing speed, and ongoing efforts to standardize reporting contribute to the evolving landscape of AI applications in clinical care.

Predictive models of epilepsy outcomes

PURPOSE OF REVIEW

Multiple complex medical decisions are necessary in the course of a chronic disease like epilepsy. Predictive tools to assist physicians and patients in navigating this complexity have emerged as a necessity and are summarized in this review.

RECENT FINDINGS

Nomograms and online risk calculators are user-friendly and offer individualized predictions for outcomes ranging from safety of antiseizure medication withdrawal (accuracy 65-73%) to seizure-freedom, naming, mood, and language outcomes of resective epilepsy surgery (accuracy 72-81%). Improving their predictive performance is limited by the nomograms' inability to ingest complex data inputs. Conversely, machine learning offers the potential of multimodal and expansive model inputs achieving human-expert level accuracy in automated scalp electroencephalogram (EEG) interpretation but lagging in predictive performance or requiring validation for other applications.

SUMMARY

Good to excellent predictive models are now available to guide medical and surgical epilepsy decision-making with nomograms offering individualized predictions and user-friendly tools, and machine learning approaches offering the potential of improved performance. Future research is necessary to bridge the two approaches for optimal translation to clinical care.

AI and EEG: Should EEGers RIP (Rest in Peace)?

[Box: see text]

EEG based automated seizure detection - A survey of medical professionals

Diagnosing and managing seizures presents substantial challenges for clinicians caring for patients with epilepsy. Although machine learning (ML) has been proposed for automated seizure detection using EEG data, there is little evidence of these technologies being broadly adopted in clinical practice. Moreover, there is a noticeable lack of surveys investigating this topic from the perspective of medical practitioners, which limits the understanding of the obstacles for the development of effective automated seizure detection. Besides the issue of generalisability and replicability seen in a small amount of studies, obstacles to the adoption of automated seizure detection remain largely unknown. To understand the obstacles preventing the application of seizure detection tools in clinical practice, we conducted a survey targeting medical professionals involved in the management of epilepsy. Our study aimed to gather insights on various factors such as the clinical utility, professional sentiment, benchmark requirements, and perceived barriers associated with the use of automated seizure detection tools. Our key findings are: I) The minimum acceptable sensitivity reported by most of our respondents (80%) seems achievable based on studies reported from most currently available ML-based EEG seizure detection algorithms, but replication studies often fail to meet this minimum. II) Respondents are receptive to the adoption of ML seizure detection tools and willing to spend time in training. III) The top three barriers for usage of such tools in clinical practice are related to availability, lack of training, and the blackbox nature of ML algorithms. Based on our findings, we developed a guide that can serve as a basis for developing ML-based seizure detection tools that meet the requirements of medical professionals, and foster the integration of these tools into clinical practice.

Machine Learning and Artificial Intelligence Applications to Epilepsy: a Review for the Practicing Epileptologist

PURPOSE OF REVIEW

Machine Learning (ML) and Artificial Intelligence (AI) are data-driven techniques to translate raw data into applicable and interpretable insights that can assist in clinical decision making. Some of these tools have extremely promising initial results, earning both great excitement and creating hype. This non-technical article reviews recent developments in ML/AI in epilepsy to assist the current practicing epileptologist in understanding both the benefits and limitations of integrating ML/AI tools into their clinical practice.

RECENT FINDINGS

ML/AI tools have been developed to assist clinicians in almost every clinical decision including (1) predicting future epilepsy in people at risk, (2) detecting and monitoring for seizures, (3) differentiating epilepsy from mimics, (4) using data to improve neuroanatomic localization and lateralization, and (5) tracking and predicting response to medical and surgical treatments. We also discuss practical, ethical, and equity considerations in the development and application of ML/AI tools including chatbots based on Large Language Models (e.g., ChatGPT). ML/AI tools will change how clinical medicine is practiced, but, with rare exceptions, the transferability to other centers, effectiveness, and safety of these approaches have not yet been established rigorously. In the future, ML/AI will not replace epileptologists, but epileptologists with ML/AI will replace epileptologists without ML/AI.

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.

Passive and active markers of cortical excitability in epilepsy

Electroencephalography (EEG) has been the primary diagnostic tool in clinical epilepsy for nearly a century. Its review is performed using qualitative clinical methods that have changed little over time. However, the intersection of higher resolution digital EEG and analytical tools developed in the past decade invites a re-exploration of relevant methodology. In addition to the established spatial and temporal markers of spikes and high-frequency oscillations, novel markers involving advanced postprocessing and active probing of the interictal EEG are gaining ground. This review provides an overview of the EEG-based passive and active markers of cortical excitability in epilepsy and of the techniques developed to facilitate their identification. Several different emerging tools are discussed in the context of specific EEG applications and the barriers we must overcome to translate these tools into clinical practice.

Ultrafast review of ambulatory EEGs with deep learning

OBJECTIVE

Interictal epileptiform discharges (IED) are hallmark biomarkers of epilepsy which are typically detected through visual analysis. Deep learning has shown potential in automating IED detection, which could reduce the burden of visual analysis in clinical practice. This is particularly relevant for ambulatory electroencephalograms (EEGs), as these entail longer review times.

METHODS

We applied a previously trained neural network to an independent dataset of 100 ambulatory EEGs (average duration 20.6 h). From these, 42 EEGs contained IEDs, 25 were abnormal without IEDs and 33 were normal. The algorithm flagged 2 second epochs that it considered IEDs. The EEGs were provided to an expert, who used NeuroCenter EEG to review the recordings. The expert concluded if each recording contained IEDs, and was timed during the process.

RESULTS

The conclusion of the reviewer was the same as the EEG report in 97% of the recordings. Three EEGs contained IEDs that were not detected based on the flagged epochs. Review time for the 100 EEGs was approximately 4 h, with half of the recordings taking <2 minutes to review.

CONCLUSIONS

Our network can be used to reduce time spent on visual analysis in the clinic by 50-75 times with high reliability.

SIGNIFICANCE

Given the large time reduction potential and high success rate, this algorithm can be used in the clinic to aid in visual analysis.

Machine-learning for the prediction of one-year seizure recurrence based on routine electroencephalography

Predicting seizure recurrence risk is critical to the diagnosis and management of epilepsy. Routine electroencephalography (EEG) is a cornerstone of the estimation of seizure recurrence risk. However, EEG interpretation relies on the visual identification of interictal epileptiform discharges (IEDs) by neurologists, with limited sensitivity. Automated processing of EEG could increase its diagnostic yield and accessibility. The main objective was to develop a prediction model based on automated EEG processing to predict one-year seizure recurrence in patients undergoing routine EEG. We retrospectively selected a consecutive cohort of 517 patients undergoing routine EEG at our institution (training set) and a separate, temporally shifted cohort of 261 patients (testing set). We developed an automated processing pipeline to extract linear and non-linear features from the EEGs. We trained machine learning algorithms on multichannel EEG segments to predict one-year seizure recurrence. We evaluated the impact of IEDs and clinical confounders on performances and validated the performances on the testing set. The receiver operating characteristic area-under-the-curve for seizure recurrence after EEG in the testing set was 0.63 (95% CI 0.55-0.71). Predictions were still significantly above chance in EEGs with no IEDs. Our findings suggest that there are changes other than IEDs in the EEG signal embodying seizure propensity.

Deep net detection and onset prediction of electrographic seizure patterns in responsive neurostimulation

OBJECTIVE

Managing the progress of drug-resistant epilepsy patients implanted with the Responsive Neurostimulation (RNS) System requires the manual evaluation of hundreds of hours of intracranial recordings. The generation of these large amounts of data and the scarcity of experts' time for evaluation necessitate the development of automatic tools to detect intracranial electroencephalographic (iEEG) seizure patterns (iESPs) with expert-level accuracy. We developed an intelligent system for identifying the presence and onset time of iESPs in iEEG recordings from the RNS device.

METHODS

An iEEG dataset from 24 patients (36 293 recordings) recorded by the RNS System was used for training and evaluating a neural network model (iESPnet). The model was trained to identify the probability of seizure onset at each sample point of the iEEG. The reliability of the net was assessed and compared to baseline methods, including detections made by the device. iESPnet performance was measured using balanced accuracy and the F1 score for iESP detection. The prediction time was assessed via both the error and the mean absolute error. The model was evaluated following a hold-one-out strategy, and then validated in a separate cohort of 26 patients from a different medical center.

RESULTS

iESPnet detected the presence of an iESP with a mean accuracy value of 90% and an onset time prediction error of approximately 3.4 s. There was no relationship between electrode location and prediction outcome. Model outputs were well calibrated and unbiased by the RNS detections. Validation on a separate cohort further supported iESPnet applicability in real clinical scenarios. Importantly, RNS device detections were found to be less accurate and delayed in nonresponders; therefore, tools to improve the accuracy of seizure detection are critical for increasing therapeutic efficacy.

SIGNIFICANCE

iESPnet is a reliable and accurate tool with the potential to alleviate the time-consuming manual inspection of iESPs and facilitate the evaluation of therapeutic response in RNS-implanted patients.

Non-invasive measurements of ictal and interictal epileptiform activity using optically pumped magnetometers

Magneto- and electroencephalography (MEG/EEG) are important techniques for the diagnosis and pre-surgical evaluation of epilepsy. Yet, in current cryogen-based MEG systems the sensors are offset from the scalp, which limits the signal-to-noise ratio (SNR) and thereby the sensitivity to activity from deep structures such as the hippocampus. This effect is amplified in children, for whom adult-sized fixed-helmet systems are typically too big. Moreover, ictal recordings with fixed-helmet systems are problematic because of limited movement tolerance and/or logistical considerations. Optically Pumped Magnetometers (OPMs) can be placed directly on the scalp, thereby improving SNR and enabling recordings during seizures. We aimed to demonstrate the performance of OPMs in a clinical population. Seven patients with challenging cases of epilepsy underwent MEG recordings using a 12-channel OPM-system and a 306-channel cryogen-based whole-head system: three adults with known deep or weak (low SNR) sources of interictal epileptiform discharges (IEDs), along with three children with focal epilepsy and one adult with frequent seizures. The consistency of the recorded IEDs across the two systems was assessed. In one patient the OPMs detected IEDs that were not found with the SQUID-system, and in two patients no IEDs were found with either system. For the other patients the OPM data were remarkably consistent with the data from the cryogenic system, noting that these were recorded in different sessions, with comparable SNRs and IED-yields overall. Importantly, the wearability of OPMs enabled the recording of seizure activity in a patient with hyperkinetic movements during the seizure. The observed ictal onset and semiology were in agreement with previous video- and stereo-EEG recordings. The relatively affordable technology, in combination with reduced running and maintenance costs, means that OPM-based MEG could be used more widely than current MEG systems, and may become an affordable alternative to scalp EEG, with the potential benefits of increased spatial accuracy, reduced sensitivity to volume conduction/field spread, and increased sensitivity to deep sources. Wearable MEG thus provides an unprecedented opportunity for epilepsy, and given its patient-friendliness, we envisage that it will not only be used for presurgical evaluation of epilepsy patients, but also for diagnosis after a first seizure.

A quantitative approach to evaluating interictal epileptiform discharges based on interpretable quantitative criteria

OBJECTIVE

To provide quantitative measures of the six International Federation of Clinical Neurophysiology (IFCN) criteria for interictal epileptiform discharge (IED) identification and estimate the likelihood of a candidate IED being epileptiform.

METHODS

We designed an algorithm to identify five fiducial landmarks (onset, peak, trough, slow-wave peak, offset) of a candidate IED, and from these to quantify the six IFCN features of IEDs. Another model was trained with these features to quantify the probability that the waveform is epileptiform and incorporated into a user-friendly interface.

RESULTS

The model's performance is excellent (area under the receiver operating characteristic curve (AUROC) = 0.88; calibration error 0.03) but lower than human experts (receiver operating characteristic (ROC) curve is below experts' operating points) or a deep neural-network model (SpikeNet; AUCROC = 0.97; calibration error 0.04). The six features were all significant (p<0.001), but not equally important when determining potential epileptiform nature of candidate IEDs: waveform asymmetry was the most (coefficient 0.64) and duration the least discriminative (coefficient 0.09).

CONCLUSIONS

Our approach quantifies the six IFCN criteria for IED identification and combines them in an easily interpretable, accessible fashion that accurately captures the likelihood that a candidate waveform is epileptiform.

SIGNIFICANCE

This model may assist clinical electroencephalographers decide whether candidate waveforms are epileptiform and may assist trainees learn to identify IEDs.

From theory to practical fundamentals of electroencephalographic source imaging in localizing the epileptogenic zone

With continued advancement in computational technologies, the analysis of electroencephalography (EEG) has shifted from pure visual analysis to a noninvasive computational technique called EEG source imaging (ESI), which involves mathematical modeling of dipolar and distributed sources of a given scalp EEG pattern. ESI is a noninvasive phase I test for presurgical localization of the seizure onset zone in focal epilepsy. It is a relatively inexpensive modality, as it leverages scalp EEG and magnetic resonance imaging (MRI) data already collected typically during presurgical evaluation. With an adequate number of electrodes and combined with patient-specific MRI-based head models, ESI has proven to be a valuable and accurate clinical diagnostic tool for localizing the epileptogenic zone. Despite its advantages, however, ESI is routinely used at only a minority of epilepsy centers. This paper reviews the current evidence and practical fundamentals for using ESI of interictal and ictal epileptic activity during the presurgical evaluation of drug-resistant patients. We identify common errors in processing and interpreting ESI studies, describe the differences in approach needed for localizing interictal and ictal EEG discharges through practical examples, and describe best practices for optimizing the diagnostic information available from these studies.