Interrater reliability in visual identification of interictal high-frequency oscillations on electrocorticography and scalp EEG

Discovering Neurophysiological Characteristics of Pathological High-Frequency Oscillations in Epilepsy with an Explainable Deep Generative Model

Interictal high-frequency oscillation (HFO) is a promising biomarker of the epileptogenic zone (EZ). However, objective definitions to distinguish between pathological and physiological HFOs have remained elusive, impeding HFOs' clinical applications. We employed self-supervised deep generative variational autoencoders to learn such discriminative HFO features directly from their morphologies in a data-driven manner. We studied a large retrospective cohort of 185 patients who underwent intracranial monitoring and analyzed 686,410 candidate HFO events collected from 18,265 brain contacts across diverse brain regions. The model automatically clustered HFOs into distinct morphological groups in the latent space. One cluster consisted of putative morphologically defined pathological HFOs (mpHFOs): HFOs in that cluster were observed to be associated with spikes and exhibited high signal intensity both in the HFO band (>80 Hz) at detection and in the sub-HFO band (10-80 Hz) surrounding the detection and were primarily localized in the seizure onset zone (SOZ). Moreover, resection of brain regions based on a higher prevalence of interictal mpHFOs better predicted postoperative seizure outcomes than current clinical standards based on SOZ removal. Our self-supervised, explainable, deep generative model distills pathological HFOs and thus potentially helps delineate the EZ purely from interictal intracranial EEG data.

Pathological and Physiological High-frequency Oscillations on Electroencephalography in Patients with Epilepsy

High-frequency oscillations (HFOs) encompass ripples (80 Hz-200 Hz) and fast ripples (200 Hz-600 Hz), serving as a promising biomarker for localizing the epileptogenic zone in epilepsy. Spontaneous fast ripples are always pathological, while ripples may be physiological or pathological. Distinguishing physiological from pathological ripples is important not only for designating epileptogenic brain regions, but also for investigations that study ripples in the context of memory encoding, consolidation, and recall in patients with epilepsy. Many studies have sought to identify distinguishing features between pathological and physiological ripples over the past two decades. Physiological and pathological ripples differ with respect to their spatial location, cellular mechanisms, morphology, and coupling with background electroencephalographic activity. Retrospective studies have demonstrated that differentiating between pathological and physiological ripples can improve surgical outcome prediction. In this review, we summarize the characteristics, differences, and applications of pathological and physiological HFOs and discuss strategies for their clinical translation.

Interrater reliability of interictal EEG waveforms in Lennox-Gastaut Syndrome

OBJECTIVE

Identification of EEG waveforms is critical for diagnosing Lennox-Gastaut Syndrome (LGS) but is complicated by the progressive nature of the disease. Here, we assess the interrater reliability (IRR) among pediatric epileptologists for classifying EEG waveforms associated with LGS.

METHODS

A novel automated algorithm was used to objectively identify epochs of EEG with transient high power, which were termed events of interest (EOIs). The algorithm was applied to EEG from 20 LGS subjects and 20 healthy controls during NREM sleep, and 1350 EOIs were identified. Three raters independently reviewed the EOIs within isolated 15-second EEG segments in a randomized, blinded fashion. For each EOI, the raters assigned a waveform label (spike and slow wave, generalized paroxysmal fast activity, seizure, spindle, vertex, muscle, artifact, nothing, or other) and indicated the perceived subject type (LGS or control).

RESULTS

Labeling of subject type had 85% accuracy across all EOIs and an IRR of κ =0.790, suggesting that brief segments of EEG containing high-power waveforms can be reliably classified as pathological or normal. Waveform labels were less consistent, with κ =0.558, and the results were highly variable for different categories of waveforms. Label mismatches typically occurred when one reviewer selected "nothing," suggesting that reviewers had different thresholds for applying named labels.

SIGNIFICANCE

Classification of EEG waveforms associated with LGS has weak IRR, due in part to varying thresholds applied during visual review. Computational methods to objectively define EEG biomarkers of LGS may improve IRR and aid clinical decision-making.

Scalp high-frequency oscillations differentiate neonates with seizures from healthy neonates

OBJECTIVE

We aimed to investigate (1) whether an automated detector can capture scalp high-frequency oscillations (HFO) in neonates and (2) whether scalp HFO rates can differentiate neonates with seizures from healthy neonates.

METHODS

We considered 20 neonates with EEG-confirmed seizures and four healthy neonates. We applied a previously validated automated HFO detector to determine scalp HFO rates in quiet sleep.

RESULTS

Etiology in neonates with seizures included hypoxic-ischemic encephalopathy in 11 cases, structural vascular lesions in 6, and genetic causes in 3. The HFO rates were significantly higher in neonates with seizures (0.098 ± 0.091 HFO/min) than in healthy neonates (0.038 ± 0.025 HFO/min; P = 0.02) with a Hedge's g value of 0.68 indicating a medium effect size. The HFO rate of 0.1 HFO/min/ch yielded the highest Youden index in discriminating neonates with seizures from healthy neonates. In neonates with seizures, etiology, status epilepticus, EEG background activity, and seizure patterns did not significantly impact HFO rates.

SIGNIFICANCE

Neonatal scalp HFO can be detected automatically and differentiate neonates with seizures from healthy neonates. Our observations have significant implications for neuromonitoring in neonates. This is the first step in establishing neonatal HFO as a biomarker for neonatal seizures.

Optimizing detection and deep learning-based classification of pathological high-frequency oscillations in epilepsy

OBJECTIVE

This study aimed to explore sensitive detection methods for pathological high-frequency oscillations (HFOs) to improve seizure outcomes in epilepsy surgery.

METHODS

We analyzed interictal HFOs (80-500 Hz) in 15 children with medication-resistant focal epilepsy who underwent chronic intracranial electroencephalogram via subdural grids. The HFOs were assessed using the short-term energy (STE) and Montreal Neurological Institute (MNI) detectors and examined for spike association and time-frequency plot characteristics. A deep learning (DL)-based classification was applied to purify pathological HFOs. Postoperative seizure outcomes were correlated with HFO-resection ratios to determine the optimal HFO detection method.

RESULTS

The MNI detector identified a higher percentage of pathological HFOs than the STE detector, but some pathological HFOs were detected only by the STE detector. HFOs detected by both detectors had the highest spike association rate. The Union detector, which detects HFOs identified by either the MNI or STE detector, outperformed other detectors in predicting postoperative seizure outcomes using HFO-resection ratios before and after DL-based purification.

CONCLUSIONS

HFOs detected by standard automated detectors displayed different signal and morphological characteristics. DL-based classification effectively purified pathological HFOs.

SIGNIFICANCE

Enhancing the detection and classification methods of HFOs will improve their utility in predicting postoperative seizure outcomes.

Quantitative approaches to guide epilepsy surgery from intracranial EEG

Over the past 10 years, the drive to improve outcomes from epilepsy surgery has stimulated widespread interest in methods to quantitatively guide epilepsy surgery from intracranial EEG (iEEG). Many patients fail to achieve seizure freedom, in part due to the challenges in subjective iEEG interpretation. To address this clinical need, quantitative iEEG analytics have been developed using a variety of approaches, spanning studies of seizures, interictal periods, and their transitions, and encompass a range of techniques including electrographic signal analysis, dynamical systems modeling, machine learning and graph theory. Unfortunately, many methods fail to generalize to new data and are sensitive to differences in pathology and electrode placement. Here, we critically review selected literature on computational methods of identifying the epileptogenic zone from iEEG. We highlight shared methodological challenges common to many studies in this field and propose ways that they can be addressed. One fundamental common pitfall is a lack of open-source, high-quality data, which we specifically address by sharing a centralized high-quality, well-annotated, multicentre dataset consisting of >100 patients to support larger and more rigorous studies. Ultimately, we provide a road map to help these tools reach clinical trials and hope to improve the lives of future patients.

Optimizing Detection and Deep Learning-based Classification of Pathological High-Frequency Oscillations in Epilepsy

Objective

This study aimed to explore sensitive detection methods and deep learning (DL)-based classification for pathological high-frequency oscillations (HFOs).

Methods

We analyzed interictal HFOs (80-500 Hz) in 15 children with medication-resistant focal epilepsy who underwent resection after chronic intracranial electroencephalogram via subdural grids. The HFOs were assessed using the short-term energy (STE) and Montreal Neurological Institute (MNI) detectors and examined for pathological features based on spike association and time-frequency plot characteristics. A DL-based classification was applied to purify pathological HFOs. Postoperative seizure outcomes were correlated with HFO-resection ratios to determine the optimal HFO detection method.

Results

The MNI detector identified a higher percentage of pathological HFOs than the STE detector, but some pathological HFOs were detected only by the STE detector. HFOs detected by both detectors exhibited the most pathological features. The Union detector, which detects HFOs identified by either the MNI or STE detector, outperformed other detectors in predicting postoperative seizure outcomes using HFO-resection ratios before and after DL-based purification.

Conclusions

HFOs detected by standard automated detectors displayed different signal and morphological characteristics. DL-based classification effectively purified pathological HFOs.

Significance

Enhancing the detection and classification methods of HFOs will improve their utility in predicting postoperative seizure outcomes.

HIGHLIGHTS

HFOs detected by the MNI detector showed different traits and higher pathological bias than those detected by the STE detectorHFOs detected by both MNI and STE detectors (the Intersection HFOs) were deemed the most pathologicalA deep learning-based classification was able to distill pathological HFOs, regard-less of the initial HFO detection methods.

Characterizing physiological high-frequency oscillations using deep learning

Objective.Intracranially-recorded interictal high-frequency oscillations (HFOs) have been proposed as a promising spatial biomarker of the epileptogenic zone. However, HFOs can also be recorded in the healthy brain regions, which complicates the interpretation of HFOs. The present study aimed to characterize salient features of physiological HFOs using deep learning (DL).Approach.We studied children with neocortical epilepsy who underwent intracranial strip/grid evaluation. Time-series EEG data were transformed into DL training inputs. The eloquent cortex (EC) was defined by functional cortical mapping and used as a DL label. Morphological characteristics of HFOs obtained from EC (ecHFOs) were distilled and interpreted through a novel weakly supervised DL model.Main results.A total of 63 379 interictal intracranially-recorded HFOs from 18 children were analyzed. The ecHFOs had lower amplitude throughout the 80-500 Hz frequency band around the HFO onset and also had a lower signal amplitude in the low frequency band throughout a one-second time window than non-ecHFOs, resembling a bell-shaped template in the time-frequency map. A minority of ecHFOs were HFOs with spikes (22.9%). Such morphological characteristics were confirmed to influence DL model prediction via perturbation analyses. Using the resection ratio (removed HFOs/detected HFOs) of non-ecHFOs, the prediction of postoperative seizure outcomes improved compared to using uncorrected HFOs (area under the ROC curve of 0.82, increased from 0.76).Significance.We characterized salient features of physiological HFOs using a DL algorithm. Our results suggested that this DL-based HFO classification, once trained, might help separate physiological from pathological HFOs, and efficiently guide surgical resection using HFOs.

An approach for reliably identifying high-frequency oscillations and reducing false-positive detections

OBJECTIVE

Aiming to improve the feasibility and reliability of using high-frequency oscillations (HFOs) for translational studies of epilepsy, we present a pipeline with features specifically designed to reject false positives for HFOs to improve the automatic HFO detector.

METHODS

We presented an integrated, multi-layered procedure capable of automatically rejecting HFOs from a variety of common false positives, such as motion, background signals, and sharp transients. This method utilizes a time-frequency contour approach that embeds three different layers including peak constraints, power thresholds, and morphological identification to discard false positives. Four experts were involved in rating detected HFO events that were randomly selected from different posttraumatic epilepsy (PTE) animals for a comprehensive evaluation.

RESULTS

The algorithm was run on 768-h recordings of intracranial electrodes in 48 PTE animals. A total of 453 917 HFOs were identified by initial HFO detection, of which 450 917 were implemented for HFO refinement and 203 531 events were retained. Random sampling was used to evaluate the performance of the detector. The HFO detection yielded an overall accuracy of 0.95 ± 0.03 , with precision, recall, and F1 scores of 0.92 ± 0.05 , 0.99 ± 0.01 , and 0.94 ± 0.03 , respectively. For the HFO classification, our algorithm obtained an accuracy of 0.97 ± 0.02 . For the inter-rater reliability of algorithm evaluation, the agreement among four experts was 0.94 ± 0.03 for HFO detection and 0.85 ± 0.04 for HFO classification.

SIGNIFICANCE

Our approach shows that a segregated pipeline design with a focus on false-positive rejection can improve the detection efficiency and provide reliable results. This pipeline does not require customization and uses fixed parameters, making it highly feasible and translatable for basic and clinical applications of epilepsy.

Refining epileptogenic high-frequency oscillations using deep learning: a reverse engineering approach

Intracranially recorded interictal high-frequency oscillations have been proposed as a promising spatial biomarker of the epileptogenic zone. However, its visual verification is time-consuming and exhibits poor inter-rater reliability. Furthermore, no method is currently available to distinguish high-frequency oscillations generated from the epileptogenic zone (epileptogenic high-frequency oscillations) from those generated from other areas (non-epileptogenic high-frequency oscillations). To address these issues, we constructed a deep learning-based algorithm using chronic intracranial EEG data via subdural grids from 19 children with medication-resistant neocortical epilepsy to: (i) replicate human expert annotation of artefacts and high-frequency oscillations with or without spikes, and (ii) discover epileptogenic high-frequency oscillations by designing a novel weakly supervised model. The ‘purification power’ of deep learning is then used to automatically relabel the high-frequency oscillations to distill epileptogenic high-frequency oscillations. Using 12 958 annotated high-frequency oscillation events from 19 patients, the model achieved 96.3% accuracy on artefact detection (F1 score = 96.8%) and 86.5% accuracy on classifying high-frequency oscillations with or without spikes (F1 score = 80.8%) using patient-wise cross-validation. Based on the algorithm trained from 84 602 high-frequency oscillation events from nine patients who achieved seizure-freedom after resection, the majority of such discovered epileptogenic high-frequency oscillations were found to be ones with spikes (78.6%, P < 0.001). While the resection ratio of detected high-frequency oscillations (number of resected events/number of detected events) did not correlate significantly with post-operative seizure freedom (the area under the curve = 0.76, P = 0.06), the resection ratio of epileptogenic high-frequency oscillations positively correlated with post-operative seizure freedom (the area under the curve = 0.87, P = 0.01). We discovered that epileptogenic high-frequency oscillations had a higher signal intensity associated with ripple (80–250 Hz) and fast ripple (250–500 Hz) bands at the high-frequency oscillation onset and with a lower frequency band throughout the event time window (the inverted T-shaped), compared to non-epileptogenic high-frequency oscillations. We then designed perturbations on the input of the trained model for non-epileptogenic high-frequency oscillations to determine the model’s decision-making logic. The model confidence significantly increased towards epileptogenic high-frequency oscillations by the artificial introduction of the inverted T-shaped signal template (mean probability increase: 0.285, P < 0.001), and by the artificial insertion of spike-like signals into the time domain (mean probability increase: 0.452, P < 0.001). With this deep learning-based framework, we reliably replicated high-frequency oscillation classification tasks by human experts. Using a reverse engineering technique, we distinguished epileptogenic high-frequency oscillations from others and identified its salient features that aligned with current knowledge.

Intracranial EEG Biomarkers for Seizure Lateralization in Rapidly-Bisynchronous Epilepsy After Laser Corpus Callosotomy

Objective: It has been asserted that high-frequency analysis of intracranial EEG (iEEG) data may yield information useful in localizing epileptogenic foci. Methods: We tested whether proposed biomarkers could predict lateralization based on iEEG data collected prior to corpus callosotomy (CC) in three patients with bisynchronous epilepsy, whose seizures lateralized definitively post-CC. Lateralization data derived from algorithmically-computed ictal phase-locked high gamma (PLHG), high gamma amplitude (HGA), and low-frequency (filtered) line length (LFLL), as well as interictal high-frequency oscillation (HFO) and interictal epileptiform discharge (IED) rate metrics were compared against ground-truth lateralization from post-CC ictal iEEG. Results: Pre-CC unilateral IEDs were more frequent on the more-pathologic side in all subjects. HFO rate predicted lateralization in one subject, but was sensitive to detection threshold. On pre-CC data, no ictal metric showed better predictive power than any other. All post-corpus callosotomy seizures lateralized to the pathological hemisphere using PLHG, HGA, and LFLL metrics. Conclusions: While quantitative metrics of IED rate and ictal HGA, PHLG, and LFLL all accurately lateralize based on post-CC iEEG, only IED rate consistently did so based on pre-CC data. Significance: Quantitative analysis of IEDs may be useful in lateralizing seizure pathology. More work is needed to develop reliable techniques for high-frequency iEEG analysis.

High-Frequency Oscillations in Epilepsy: What Have We Learned and What Needs to be Addressed

For the past 2 decades, high-frequency oscillations (HFOs) have been enthusiastically studied by the epilepsy community. Emerging evidence shows that HFOs harbor great promise to delineate epileptogenic brain areas and possibly predict the likelihood of seizures. Investigations into HFOs in clinical epilepsy have advanced from small retrospective studies relying on visual identification and correlation analysis to larger prospective assessments using automatic detection and prediction strategies. Although most studies have yielded promising results, some have revealed significant obstacles to clinical application of HFOs, thus raising debate about the reliability and practicality of HFOs as clinical biomarkers. In this review, we give an overview of the current state of HFO research and pinpoint the conceptual and methodological issues that have hampered HFO translation. We highlight recent insights gained from long-term data, high-density recordings, and multicenter collaborations and discuss the open questions that need to be addressed in future research.

Scalp EEG interictal high frequency oscillations as an objective biomarker of infantile spasms

OBJECTIVE

To investigate the diagnostic utility of high frequency oscillations (HFOs) via scalp electroencephalogram (EEG) in infantile spasms.

METHODS

We retrospectively analyzed interictal slow-wave sleep EEGs sampled at 2,000 Hz recorded from 30 consecutive patients who were suspected of having infantile spasms. We measured the rate of HFOs (80-500 Hz) and the strength of the cross-frequency coupling between HFOs and slow-wave activity (SWA) at 3-4 Hz and 0.5-1 Hz as quantified with modulation indices (MIs).

RESULTS

Twenty-three patients (77%) exhibited active spasms during the overnight EEG recording. Although the HFOs were detected in all children, increased HFO rate and MIs correlated with the presence of active spasms (p < 0.001 by HFO rate; p < 0.01 by MIs at 3-4 Hz; p = 0.02 by MIs at 0.5-1 Hz). The presence of active spasms was predicted by the logistic regression models incorporating HFO-related metrics (AUC: 0.80-0.98) better than that incorporating hypsarrhythmia (AUC: 0.61). The predictive performance of the best model remained favorable (87.5% accuracy) after a cross-validation procedure.

CONCLUSIONS

Increased rate of HFOs and coupling between HFOs and SWA are associated with active epileptic spasms.

SIGNIFICANCE

Scalp-recorded HFOs may serve as an objective EEG biomarker for active epileptic spasms.

Pitfalls in Scalp High-Frequency Oscillation Detection From Long-Term EEG Monitoring

Aims: Intracranially recorded high-frequency oscillations (>80 Hz) are considered a candidate epilepsy biomarker. Recent studies claimed their detectability on the scalp surface. We aimed to investigate the applicability of high-frequency oscillation analysis to routine surface EEG obtained at an epilepsy monitoring unit. Methods: We retrospectively analyzed surface EEGs of 18 patients with focal epilepsy and six controls, recorded during sleep under maximal medication withdrawal. As a proof of principle, the occurrence of motor task-related events during wakefulness was analyzed in a subsample of six patients with seizure- or syncope-related motor symptoms. Ripples (80-250 Hz) and fast ripples (>250 Hz) were identified by semi-automatic detection. Using semi-parametric statistics, differences in spontaneous and task-related occurrence rates were examined within subjects and between diagnostic groups considering the factors diagnosis, brain region, ripple type, and task condition. Results: We detected high-frequency oscillations in 17 out of 18 patients and in four out of six controls. Results did not show statistically significant differences in the mean rates of event occurrences, neither regarding the laterality of the epileptic focus, nor with respect to active and inactive task conditions, or the moving hand laterality. Significant differences in general spontaneous incidence [WTS(1) = 9.594; p = 0.005] that indicated higher rates of fast ripples compared to ripples, notably in patients with epilepsy compared to the control group, may be explained by variations in data quality. Conclusion: The current analysis methods are prone to biases. A common agreement on a standard operating procedure is needed to ensure reliable and economic detection of high-frequency oscillations.

Scalp EEG high frequency oscillations as a biomarker of treatment response in epileptic encephalopathy with continuous spike-and-wave during sleep (CSWS)

PURPOSE

We investigated whether the presence of interictal scalp EEG high frequency oscillations (HFOs) in children with epileptic encephalopathy with continuous spike-and-wave during sleep (CSWS) can predict seizure and cognitive outcome after steroid therapy.

METHODS

Twenty-two children with CSWS were prospectively enrolled and received methylprednisolone therapy. Interictal scalp HFOs, spike wave index (SWI) and intelligence quotient (IQ) were assessed before and after the treatment. The children were divided into two groups based on the early seizure reduction ratio at 2 weeks (≥50%, "response group"; otherwise "non-response group"). The "response group" was further divided into two subgroups ("relapse" and "non-relapse" subgroups) according to the late seizure outcome (after 3 months).

RESULTS

Interictal HFOs and electrical status epilepticus in sleep (ESES) (defined as SWI ≥ 85%) were detected in all children at the baseline. In the response with relapse group (n = 11), the detection ratio of HFOs was significantly higher than that of ESES at 2 weeks (81.2 vs. 27.3%), 3 months (90.9 vs. 36.4%), and 6 months (100 vs. 54.5%) post-therapy. In the non-response group (n = 4), both HFOs and ESES persisted in all children. The average IQ improved significantly only in the response with non-relapse group. The persistence of HFOs negatively correlated with both the average IQ, yet the persistence of ESES did not.

CONCLUSION

Interictal scalp HFOs may be a favorable non-invasive biomarker of predicting seizure and cognitive outcome in CSWS.

Prospective observational study: Fast ripple localization delineates the epileptogenic zone

OBJECTIVE

To investigate spatial correlation between interictal HFOs and neuroimaging abnormalities, and to determine if complete removal of prospectively identified interictal HFOs correlates with post-surgical seizure-freedom.

METHODS

Interictal fast ripples (FRs: 250-500 Hz) in 19 consecutive children with pharmacoresistant focal epilepsy who underwent extra-operative electrocorticography (ECoG) recording were prospectively analyzed. The interictal FRs were sampled at 2000 Hz and were visually identified during 10 min of slow wave sleep. Interictal FRs, MRI and FDG-PET were delineated on patient-specific reconstructed three-dimensional brain MRI.

RESULTS

Interictal FRs were observed in all patients except one. Thirteen out of 18 patients (72%) exhibited FRs beyond the extent of neuroimaging abnormalities. Fifteen of 19 children underwent resective surgery, and survival analysis with log-rank test demonstrated that complete resection of cortical sites showing interictal FRs correlated with longer post-operative seizure-freedom (p < 0.01). Complete resection of seizure onset zones (SOZ) also correlated with longer post-operative seizure-freedom (p = 0.01), yet complete resection of neuroimaging abnormalities did not (p = 0.43).

CONCLUSIONS

Prospective visual analysis of interictal FRs was feasible, and it seemed to accurately localize epileptogenic zones.

SIGNIFICANCE

Topological extent of epileptogenic region may exceed what is discernible by multimodal neuroimaging.