Ictal Central Apnea Is Predictive of Mesial Temporal Seizure Onset: An Intracranial Investigation.

OBJECTIVE: Ictal central apnea (ICA) is a semiological sign of focal epilepsy, associated with temporal and frontal lobe seizures. In this study, using qualitative and quantitative approaches, we aimed to assess the localizational value of ICA. We also aimed to compare ICA clinical utility in relation to other seizure semiological features of focal epilepsy. METHODS: We analyzed seizures in patients with medically refractory focal epilepsy undergoing intracranial stereotactic electroencephalographic (SEEG) evaluations with simultaneous multimodal cardiorespiratory monitoring. A total of 179 seizures in 72 patients with reliable artifact-free respiratory signal were analyzed. RESULTS: ICA was seen in 55 of 179 (30.7%) seizures. Presence of ICA predicted a mesial temporal seizure onset compared to those without ICA (odds ratio = 3.8, 95% confidence interval = 1.3-11.6, p = 0.01). ICA specificity was 0.82. ICA onset was correlated with increased high-frequency broadband gamma (60-150Hz) activity in specific mesial or basal temporal regions, including amygdala, hippocampus, and fusiform and lingual gyri. Based on our results, ICA has an almost 4-fold greater association with mesial temporal seizure onset zones compared to those without ICA and is highly specific for mesial temporal seizure onset zones. As evidence of symptomatogenic areas, onset-synchronous increase in high gamma activity in mesial or basal temporal structures was seen in early onset ICA, likely representing anatomical substrates for ICA generation. INTERPRETATION: ICA recognition may help anatomoelectroclinical localization of clinical seizure onset to specific mesial and basal temporal brain regions, and the inclusion of these regions in SEEG evaluations may help accurately pinpoint seizure onset zones for resection. ANN NEUROL 2024;95:998-1008.

The value of additional electrodes when stereo-electroencephalography is inconclusive.

OBJECTIVE: Stereo-electroencephalography (SEEG) is the preferred method for intracranial localization of the seizure-onset zone (SOZ) in drug-resistant focal epilepsy. Occasionally SEEG evaluation fails to confirm the pre-implantation hypothesis. This leads to a decision tree regarding whether the addition of SEEG electrodes (two-step SEEG - 2sSEEG) or placement of subdural electrodes (SDEs) after SEEG (SEEG2SDE) would help. There is a dearth of literature encompassing this scenario, and here we aimed to characterize outcomes following unplanned two-step intracranial EEG (iEEG). METHODS: All 225 adult SEEG cases over 8 years at our institution were reviewed to extract patient data and outcomes following a two-step evaluation. Three raters independently quantified benefits of additional intracranial electrodes. The relationship between two-step iEEG benefit and clinical outcome was then analyzed. RESULTS: Fourteen patients underwent 2sSEEG and nine underwent SEEG2SDE. In the former cohort, the second SEEG procedure was performed for these reasons-precise localization of the SOZ (36%); defining margins of eloquent cortex (21%); and broadening coverage in the setting of non-localizable seizure onsets (43% of cases). Sixty-four percent of 2sSEEG cases were consistently deemed beneficial (Light's κ = 0.80). 2sSEEG performed for the first two indications was much more beneficial than when onsets were not localizable (100% vs 17%, p = .02). In the SEEG2SDE cohort, SDEs identified the SOZ and enabled delineation of margins relative to eloquent cortex in all cases. SIGNIFICANCE: The two-step iEEG is useful if the initial evaluation is broadly concordant with the original electroclinical hypothesis, where it can clarify onset zones or delineate safe surgical margins; however, it provides minimal benefit when the implantation hypothesis is erroneous, and we recommend that 2sSEEG not be generally utilized in such cases. SDE implantation after SEEG minimizes the need for SDEs and is helpful in delineating surgical boundaries relative to ictal-onset zones and eloquent cortex.

Reliable detection of generalized convulsive seizures using an off-the-shelf digital watch: A multisite phase 2 study.

OBJECTIVE: The aim of this study was to develop a machine learning algorithm using an off-the-shelf digital watch, the Samsung watch (SM-R800), and evaluate its effectiveness for the detection of generalized convulsive seizures (GCS) in persons with epilepsy. METHODS: This multisite epilepsy monitoring unit (EMU) phase 2 study included 36 adult patients. Each patient wore a Samsung watch that contained accelerometer, gyroscope, and photoplethysmographic sensors. Sixty-eight time and frequency domain features were extracted from the sensor data and were used to train a random forest algorithm. A testing framework was developed that would better reflect the EMU setting, consisting of (1) leave-one-patient-out cross-validation (LOPO CV) on GCS patients, (2) false alarm rate (FAR) testing on nonseizure patients, and (3) "fixed-and-frozen" prospective testing on a prospective patient cohort. Balanced accuracy, precision, sensitivity, and FAR were used to quantify the performance of the algorithm. Seizure onsets and offsets were determined by using video-electroencephalographic (EEG) monitoring. Feature importance was calculated as the mean decrease in Gini impurity during the LOPO CV testing. RESULTS: LOPO CV results showed balanced accuracy of .93 (95% confidence interval [CI] = .8-.98), precision of .68 (95% CI = .46-.85), sensitivity of .87 (95% CI = .62-.96), and FAR of .21/24 h (interquartile range [IQR] = 0-.90). Testing the algorithm on patients without seizure resulted in an FAR of .28/24 h (IQR = 0-.61). During the "fixed-and-frozen" prospective testing, two patients had three GCS, which were detected by the algorithm, while generating an FAR of .25/24 h (IQR = 0-.89). Feature importance showed that heart rate-based features outperformed accelerometer/gyroscope-based features. SIGNIFICANCE: Commercially available wearable digital watches that reliably detect GCS, with minimum false alarm rates, may overcome usage adoption and other limitations of custom-built devices. Contingent on the outcomes of a prospective phase 3 study, such devices have the potential to provide non-EEG-based seizure surveillance and forecasting in the clinical setting.

Grading system for assessing the confidence in the epileptogenic zone reported in published studies: A Delphi consensus study.

OBJECTIVE: This study was undertaken to develop a standardized grading system based on expert consensus for evaluating the level of confidence in the localization of the epileptogenic zone (EZ) as reported in published studies, to harmonize and facilitate systematic reviews in the field of epilepsy surgery. METHODS: We conducted a Delphi study involving 22 experts from 18 countries, who were asked to rate their level of confidence in the localization of the EZ for various theoretical clinical scenarios, using different scales. Information provided in these scenarios included one or several of the following data: magnetic resonance imaging (MRI) findings, invasive electroencephalography summary, and postoperative seizure outcome. RESULTS: The first explorative phase showed an overall interrater agreement of .347, pointing to large heterogeneity among experts' assessments, with only 17% of the 42 proposed scenarios associated with a substantial level of agreement. A majority showed preferences for the simpler scale and single-item scenarios. The successive Delphi voting phases resulted in a majority consensus across experts, with more than two thirds of respondents agreeing on the rating of each of the tested single-item scenarios. High or very high levels of confidence were ascribed to patients with either an Engel class I or class IA postoperative seizure outcome, a well-delineated EZ according to all available invasive EEG (iEEG) data, or a well-delineated focal epileptogenic lesion on MRI. MRI signs of hippocampal sclerosis or atrophy were associated with a moderate level of confidence, whereas a low level was ascribed to other MRI findings, a poorly delineated EZ according to iEEG data, or an Engel class II-IV postoperative seizure outcome. SIGNIFICANCE: The proposed grading system, based on an expert consensus, provides a simple framework to rate the level of confidence in the EZ reported in published studies in a structured and harmonized way, offering an opportunity to facilitate and increase the quality of systematic reviews and guidelines in the field of epilepsy surgery.

Brain gray matter changes in children at risk for sudden unexpected death in epilepsy.

BACKGROUND: Potential failing adult brain sites, stratified by risk, mediating Sudden Unexpected Death in Epilepsy (SUDEP) have been described, but are unknown in children. METHODS: We examined regional brain volumes using T1-weighted MRI images in 21 children with epilepsy at high SUDEP risk and 62 healthy children, together with SUDEP risk scores, calculated from focal seizure frequency. Gray matter tissue type was partitioned, maps normalized, smoothed, and compared between groups (SPM12; ANCOVA; covariates, age, sex, and BMI). Partial correlations between regional volumes and seizure frequency were examined (SPM12, covariates, age, sex, and BMI); 67% were at high risk for SUDEP. RESULTS: The cerebellar cortex, hippocampus, amygdala, putamen, cingulate, thalamus, and para-hippocampal gyrus showed increased gray matter volumes in epilepsy, and decreased volumes in the posterior thalamus, lingual gyrus, and temporal cortices. The cingulate, insula, and putamen showed significant positive relationships with focal seizure frequency indices using whole-brain voxel-by-voxel partial correlations. Tissue volume changes in selected sites differed in direction from adults; particularly, cerebellar sites, key for hypotensive recovery, increased rather than adult declines. CONCLUSION: The volume increases may represent expansion by inflammatory or other processes that, with sustained repetitive seizure discharge, lead to tissue volume declines described earlier in adults. IMPACT: Children with epilepsy, who are at risk for Sudden Unexplained Death, show changes in brain volume that often differ in direction of change from adults at risk for SUDEP. Sites of volume change play significant roles in mediating breathing and blood pressure, and include areas that serve recovery from prolonged apnea and marked loss of blood pressure. The extent of volume changes correlated with focal seizure frequency. Although the underlying processes contributing to regional volume changes remain speculative, regions of tissue swelling in pediatric brain areas may represent transitory conditions that later lead to tissue loss in the adult condition.

Interictal respiratory variability predicts severity of hypoxemia after generalized convulsive seizures.

OBJECTIVE: Severe respiratory dysfunction induced by generalized convulsive seizures (GCS) is now thought to be a common mechanism for sudden unexpected death in epilepsy (SUDEP). In a mouse model of seizure-induced death, increased interictal respiratory variability was reported in mice that later died of respiratory arrest after GCS. We studied respiratory variability in epilepsy patients as a predictive tool for severity of postictal hypoxemia, a potential biomarker for SUDEP risk. We then explored the relationship between respiratory variability and central CO2 drive, measured by the hypercapnic ventilatory response (HCVR). METHODS: We reviewed clinical, video-electroencephalography, and respiratory (belts, airflow, pulse oximeter, and HCVR) data of epilepsy patients. Mean, SD, and coefficient of variation (CV) of interbreath interval (IBI) were calculated. Primary outcomes were: (1) nadir of capillary oxygen saturation (SpO2 ) and (2) duration of oxygen desaturation. Poincaré plots of IBI were created. Covariates were evaluated in univariate models, then, based on Akaike information criteria (AIC), multivariate regression models were created. RESULTS: Of 66 GCS recorded in 131 subjects, 30 had interpretable respiratory data. In the multivariate model with the lowest AIC value, duration of epilepsy was a significant predictor of duration of oxygen desaturation. Duration of tonic phase and CV of IBI during the third postictal minute correlated with SpO2 nadir, whereas CV of IBI during non-rapid eye movement sleep had a negative correlation. Poincaré plots showed that long-term variability was significantly greater in subjects with ≥200 s of postictal oxygen desaturation after GCS compared to those with <200 s desaturation. Finally, HCVR slope showed a negative correlation with measures of respiratory variability. SIGNIFICANCE: These results indicate that interictal respiratory variability predicts severity of postictal oxygen desaturation, suggesting its utility as a potential biomarker. They also suggest that interictal respiratory control may be abnormal in some patients with epilepsy.

Altered amygdala volumes and microstructure in focal epilepsy patients with tonic-clonic seizures, ictal, and post-convulsive central apnea.

OBJECTIVES: Sudden unexpected death in epilepsy (SUDEP) is a leading cause of death for patients with epilepsy; however, the pathophysiology remains unclear. Focal-to-bilateral tonic-clonic seizures (FBTCS) are a major risk factor, and centrally-mediated respiratory depression may increase the risk further. Here, we determined the volume and microstructure of the amygdala, a key structure that can trigger apnea in people with focal epilepsy, stratified by the presence or absence of FBTCS, ictal central apnea (ICA), and post-convulsive central apnea (PCCA). METHODS: Seventy-three patients with focal impaired awareness seizures without FBTC seizures (FBTCneg group) and 30 with FBTCS (FBTCpos group) recorded during video electroencephalography (VEEG) with respiratory monitoring were recruited prospectively during presurgical investigations. We acquired high-resolution T1-weighted anatomic and multi-shell diffusion images, and computed neurite orientation dispersion and density imaging (NODDI) metrics in all patients with epilepsy and 69 healthy controls. Amygdala volumetric and microstructure alterations were compared between three groups: healthy subjects, FBTCneg and FBTCpos groups. The FBTCpos group was further subdivided by the presence of ICA and PCCA, verified by VEEG. RESULTS: Bilateral amygdala volumes were significantly increased in the FBTCpos cohort compared to healthy controls and the FBTCneg group. Patients with recorded PCCA had the highest increase in bilateral amygdala volume of the FBTCpos cohort. Amygdala neurite density index (NDI) values were decreased significantly in both the FBTCneg and FBTCpos groups relative to healthy controls, with values in the FBTCpos group being the lowest of the two. The presence of PCCA was associated with significantly lower NDI values vs the non-apnea FBTCpos group (p = 0.004). SIGNIFICANCE: Individuals with FBTCpos and PCCA show significantly increased amygdala volumes and disrupted architecture bilaterally, with greater changes on the left side. The structural alterations reflected by NODDI and volume differences may be associated with inappropriate cardiorespiratory patterns mediated by the amygdala, particularly after FBTCS. Determination of amygdala volumetric and architectural changes may assist identification of individuals at risk.

Stimulation-induced respiratory enhancement in corticothalamic regions.

OBJECTIVE: We aimed to identify corticothalamic areas and electrical stimulation paradigms that optimally enhance breathing. METHODS: Twenty-nine patients with medically intractable epilepsy were prospectively recruited in an epilepsy monitoring unit while undergoing stereoelectroencephalographic evaluation. Direct electrical stimulation in cortical and thalamic regions was carried out using low (<1 Hz) and high (≥10 Hz) frequencies, and low (<5 mA) and high (≥5 mA) current intensities, with pulse width of .1 ms. Electrocardiography, arterial oxygen saturation (SpO2 ), end-tidal carbon dioxide (ETCO2 ), oronasal airflow, and abdominal and thoracic plethysmography were monitored continuously during stimulations. Airflow signal was used to estimate breathing rate, tidal volume, and minute ventilation (MV) changes during stimulation, compared to baseline. RESULTS: Electrical stimulation increased MV in the amygdala, anterior cingulate, anterior insula, temporal pole, and thalamus, with an average increase in MV of 20.8% ± 28.9% (range = 0.2%-165.6%) in 19 patients. MV changes were associated with SpO2 and ETCO2 changes (p < .001). Effects on respiration were parameter and site dependent. Within amygdala, low-frequency stimulation of the medial region produced 78.49% greater MV change (p < .001) compared to high-frequency stimulation. Longer stimulation produced greater MV changes (an increase of 4.47% in MV for every additional 10 s, p = .04). SIGNIFICANCE: Stimulation of amygdala, anterior cingulate gyrus, anterior insula, temporal pole, and thalamus, using certain stimulation paradigms, enhances respiration. Among tested paradigms, low-frequency, low-intensity, long-duration stimulation of the medial amygdala is the most effective breathing enhancement stimulation strategy. Such approaches may pave the way for the future development of neuromodulatory techniques that aid rescue from seizure-related apnea, potentially as a targeted sudden unexpected death in epilepsy prevention method.

Subject-Specific Activation of Central Respiratory Centers during Breath-Holding Functional Magnetic Resonance Imaging.

BACKGROUND: Voluntary breath-holding (BH) triggers responses from central neural control and respiratory centers in order to restore breathing. Such responses can be observed using functional MRI (fMRI). OBJECTIVES: We used this paradigm in healthy volunteers with the view to develop a biomarker that could be used to investigate disorders of the central control of breathing at the individual patient level. METHOD: In 21 healthy human subjects (mean age±SD, 32.8 ± 9.9 years old), fMRI was used to determine, at both the individual and group levels, the physiological neural response to expiratory and inspiratory voluntary apneas, within respiratory control centers in the brain and brainstem. RESULTS: Group analysis showed that expiratory BH, but not inspiratory BH, triggered activation of the pontine respiratory group and raphe nuclei at the group level, with a significant relationship between the levels of activation and drop in SpO2. Using predefined ROIs, expiratory BH, and to a lesser extent, inspiratory BH were associated with activation of most respiratory centers. The right ventrolateral nucleus of the thalamus, right pre-Bötzinger complex, right VRG, right nucleus ambiguus, and left Kölliker-Fuse-parabrachial complex were only activated during inspiratory BH. Individual analysis identified activations of cortical/subcortical and brainstem structures related to respiratory control in 19 out of 21 subjects. CONCLUSION: Our study shows that BH paradigm allows to reliably trigger fMRI response from brainstem and cortical areas involved in respiratory control at the individual level, suggesting that it might serve as a clinically relevant biomarker to investigate conditions associated with an altered central control of respiration.

Forecasting seizure clusters from chronic ambulatory electrocorticography.

Seizure clusters are seizures that occur in rapid succession during periods of heightened seizure risk and are associated with substantial morbidity and sudden unexpected death in epilepsy. The objective of this feasibility study was to evaluate the performance of a novel seizure cluster forecasting algorithm. Chronic ambulatory electrocorticography recorded over an average of 38 months in 10 subjects with drug-resistant epilepsies was analyzed pseudoprospectively by dividing data into training (first 85%) and validation periods. For each subject, the probability of seizure clustering, derived from the Kolmogorov-Smirnov statistic using a novel algorithm, was forecasted in the validation period using individualized autoregressive models that were optimized from training data. The primary outcome of this study was the mean absolute scaled error (MASE) of 1-day horizon forecasts. From 10 subjects, 394 ± 142 (mean ± SD) electrocorticography-based seizure events were extracted for analysis, representing a span of 38 ± 27 months of recording. MASE across all subjects was .74 ± .09, .78 ± .09, and .83 ± .07 at .5-, 1-, and 2-day horizons. The feasibility study demonstrates that seizure clusters are quasiperiodic and can be forecasted to clinically meaningful horizons. Pending validation in larger cohorts, the forecasting approach described herein may herald chronotherapy during imminent heightened seizure vulnerability.

Limbic and paralimbic respiratory modulation: From inhibition to enhancement.

OBJECTIVE: Increased understanding of the role of cortical structures in respiratory control may help the understanding of seizure-induced respiratory dysfunction that leads to sudden unexpected death in epilepsy (SUDEP). The aim of this study was to characterize respiratory responses to electrical stimulation (ES), including inhibition and enhancement of respiration. METHODS: We prospectively recruited 19 consecutive patients with intractable epilepsy undergoing stereotactic electroencephalography (EEG) evaluation from June 2015 to June 2018. Inclusion criteria were patients ≥18 years in whom ES was indicated for clinical mapping of ictal onset or eloquent cortex as part of the presurgical evaluation. ES was carried out at 50 Hz, 0.2 msec, and 1-10 mA current intensity. Common brain regions sampled across all patients were amygdala (AMY), hippocampus (HG), anterior cingulate gyrus (CING), orbitofrontal cortex (OrbF), temporal neocortex (TNC), temporal pole (TP), and entorhinal cortex (ERC). Seven hundred fifty-five stimulations were conducted. Quantitative analysis of breathing signal, that is, changes in breathing rate (BR), depth (TV), and minute ventilation (MV), was carried out during ES using the BreathMetrics breathing waveform analysis toolbox. Electrocardiography, arterial oxygen saturation, end-tidal and transcutaneous carbon dioxide, nasal airflow, and abdominal and thoracic plethysmography were monitored continuously during stimulations. RESULTS: Electrical stimulation of TP and CING (at lower current strengths <3 mA) increased TV and MV. At >7-10 mA, CING decreased TV and MV. On the other hand, decreased TV and MV occurred with stimulation of mesial temporal structures such as AMY and HG. Breathing changes were dependent on stimulation intensity. Lateral temporal, entorhinal, and orbitofrontal cortices did not affect breathing either way. SIGNIFICANCE: These findings suggest that breathing responses other than apnea can be induced by ES. Identification of two regions-the temporal pole and anterior cingulate gyrus-for enhancement of breathing may be important in paving the way to future development of strategies for prevention of SUDEP.

Can Big Data guide prognosis and clinical decisions in epilepsy?

Big Data is no longer a novel concept in health care. Its promise of positive impact is not only undiminished, but daily enhanced by seemingly endless possibilities. Epilepsy is a disorder with wide heterogeneity in both clinical and research domains, and thus lends itself to Big Data concepts and techniques. It is therefore inevitable that Big Data will enable multimodal research, integrating various aspects of "-omics" domains, such as phenome, genome, microbiome, metabolome, and proteome. This scope and granularity have the potential to change our understanding of prognosis and mortality in epilepsy. The scale of new discovery is unprecedented due to the possibilities promised by advances in machine learning, in particular deep learning. The subsequent possibilities of personalized patient care through clinical decision support systems that are evidence-based, adaptive, and iterative seem to be within reach. A major objective is not only to inform decision-making, but also to reduce uncertainty in outcomes. Although the adoption of electronic health record (EHR) systems is near universal in the United States, for example, advanced clinical decision support in or ancillary to EHRs remains sporadic. In this review, we discuss the role of Big Data in the development of clinical decision support systems for epilepsy care, prognostication, and discovery.

A Bespoke Electronic Health Record for Epilepsy Care (EpiToMe): Development and Qualitative Evaluation.

BACKGROUND: While electronic health records (EHR) bring various benefits to health care, EHR systems are often criticized as cumbersome to use, failing to fulfill the promise of improved health care delivery with little more than a means of meeting regulatory and billing requirements. EHR has also been recognized as one of the contributing factors for physician burnout. OBJECTIVE: Specialty-specific EHR systems have been suggested as an alternative approach that can potentially address challenges associated with general-purpose EHRs. We introduce the Epilepsy Tracking and optimized Management engine (EpiToMe), an exemplar bespoke EHR system for epilepsy care. EpiToMe uses an agile, physician-centered development strategy to optimize clinical workflow and patient care documentation. We present the design and implementation of EpiToMe and report the initial feedback on its utility for physician burnout. METHODS: Using collaborative, asynchronous data capturing interfaces anchored to a domain ontology, EpiToMe distributes reporting and documentation workload among technicians, clinical fellows, and attending physicians. Results of documentation are transmitted to the parent EHR to meet patient care requirements with a push of a button. An HL7 (version 2.3) messaging engine exchanges information between EpiToMe and the parent EHR to optimize clinical workflow tasks without redundant data entry. EpiToMe also provides live, interactive patient tracking interfaces to ease the burden of care management. RESULTS: Since February 2019, 15,417 electroencephalogram reports, 2635 Epilepsy Monitoring Unit daily reports, and 1369 Epilepsy Monitoring Unit phase reports have been completed in EpiToMe for 6593 unique patients. A 10-question survey was completed by 11 (among 16 invited) senior clinical attending physicians. Consensus was found that EpiToMe eased the burden of care documentation for patient management, a contributing factor to physician burnout. CONCLUSIONS: EpiToMe offers an exemplar bespoke EHR system developed using a physician-centered design and latest advancements in information technology. The bespoke approach has the potential to ease the burden of care management in epilepsy. This approach is applicable to other clinical specialties.

Ictal Lid Movements: Blinks and Lid Saccades.

Two types of lid movements, blinks and lid saccades, have discrete kinematic properties and physiology. These differences are reflected in distinct phenomenology of disorders affecting their neural substrate. Proof of this principle was seen in two patients, one with parietal lobe epilepsy and the other with temporal lobe epilepsy. The lid movements in the patient with parietal lobe epilepsy were rhythmic, yoked, and had a rapid upward component that instantaneously followed a slow downward drift. These cyclic movements strikingly resembled nystagmus, but unlike typical eye nystagmus, the rapid upward component was pathological and seemed to involve a saccadic mechanism. We suggest the terms "ictal lid saccades" or "ictal lid nystagmus" to describe such a phenomenon. In contrast, the patient with temporal lobe epilepsy had ipsilateral lid movements with rapid downward trajectories resembling reflex or spontaneous blinks. The term "ictal blink" is appropriate for this phenomenon.

Big data in epilepsy: Clinical and research considerations. Report from the Epilepsy Big Data Task Force of the International League Against Epilepsy.

Epilepsy is a heterogeneous condition with disparate etiologies and phenotypic and genotypic characteristics. Clinical and research aspects are accordingly varied, ranging from epidemiological to molecular, spanning clinical trials and outcomes, gene and drug discovery, imaging, electroencephalography, pathology, epilepsy surgery, digital technologies, and numerous others. Epilepsy data are collected in the terabytes and petabytes, pushing the limits of current capabilities. Modern computing firepower and advances in machine and deep learning, pioneered in other diseases, open up exciting possibilities for epilepsy too. However, without carefully designed approaches to acquiring, standardizing, curating, and making available such data, there is a risk of failure. Thus, careful construction of relevant ontologies, with intimate stakeholder inputs, provides the requisite scaffolding for more ambitious big data undertakings, such as an epilepsy data commons. In this review, we assess the clinical and research epilepsy landscapes in the big data arena, current challenges, and future directions, and make the case for a systematic approach to epilepsy big data.

Peri-ictal hypoxia is related to extent of regional brain volume loss accompanying generalized tonic-clonic seizures.

OBJECTIVES: Hypoxia, or abnormally low blood-oxygen levels, often accompanies seizures and may elicit brain structural changes in people with epilepsy which contribute to central processes underlying sudden unexpected death in epilepsy (SUDEP). The extent to which hypoxia may be related to brain structural alterations in this patient group remains unexplored. METHODS: We analyzed high-resolution T1-weighted magnetic resonance imaging (MRI) to determine brain morphometric and volumetric alterations in people with generalized tonic-clonic seizures (GTCS) recorded during long-term video-electroencephalography (VEEG), recruited from two sites (n = 22), together with data from age- and sex-matched healthy controls (n = 43). Subjects were sub-divided into those with mild/moderate (GTCS-hypox-mild/moderate, n = 12) and severe (GTCS-hypox-severe, n = 10) hypoxia, measured by peripheral oxygen saturation (SpO2 ) during VEEG. Whole-brain voxel-based morphometry (VBM) and regional volumetry were used to assess group comparisons and correlations between brain structural measurements as well as the duration and extent of hypoxia during GTCS. RESULTS: Morphometric and volumetric alterations appeared in association with peri-GTCS hypoxia, including volume loss in the periaqueductal gray (PAG), thalamus, hypothalamus, vermis, cerebellum, parabrachial pons, and medulla. Thalamic and PAG volume was significantly reduced in GTCS patients with severe hypoxia compared with GTCS patients with mild/moderate hypoxia. Brainstem volume loss appeared in both hypoxia groups, although it was more extensive in those with severe hypoxia. Significant negative partial correlations emerged between thalamic and hippocampal volume and extent of hypoxia, whereas vermis and accumbens volumes declined with increasing hypoxia duration. SIGNIFICANCE: Brain structural alterations in patients with GTCS are related to the extent of hypoxia in brain sites that serve vital functions. Although the changes are associative only, they provide evidence of injury to regulatory brain sites related to respiratory manifestations of seizures.

Temporal phenotyping for transitional disease progress: An application to epilepsy and Alzheimer's disease.

Complicated multifactorial diseases deteriorate from one disease to other diseases. For example, existing studies consider Alzheimer's disease (AD) a comorbidity of epilepsy, but also recognize epilepsy to occur more frequently in patients with AD than those without. It is important to understand the progress of disease that deteriorates to severe diseases. To this end, we develop a transitional phenotyping method based on both longitudinal and cross-sectional relationships between diseases and/or medications. For a cross-sectional approach, we utilized a skip-gram model to represent co-occurred disease or medication. For a longitudinal approach, we represented each patient as a transition probability between medical events and used supervised tensor factorization to decompose into groups of medical events that develop together. Then we harmonized both information to derive high-risk transitional patterns. We applied our method to disease progress from epilepsy to AD. An epilepsy-AD cohort of 600,000 patients were extracted from Cerner Health Facts data. Our experimental results suggested a causal relationship between epilepsy and later onset of AD, and also identified five epilepsy subgroups with distinct phenotypic patterns leading to AD. While such findings are preliminary, the proposed method combining representation learning with tensor factorization seems to be an effective approach for risk factor analysis.

Categorisation of EEG suppression using enhanced feature extraction for SUDEP risk assessment.

BACKGROUND: Sudden Unexpected Death in Epilepsy (SUDEP) has increased in awareness considerably over the last two decades and is acknowledged as a serious problem in epilepsy. However, the scientific community remains unclear on the reason or possible bio markers that can discern potentially fatal seizures from other non-fatal seizures. The duration of postictal generalized EEG suppression (PGES) is a promising candidate to aid in identifying SUDEP risk. The length of time a patient experiences PGES after a seizure may be used to infer the risk a patient may have of SUDEP later in life. However, the problem becomes identifying the duration, or marking the end, of PGES (Tomson et al. in Lancet Neurol 7(11):1021-1031, 2008; Nashef in Epilepsia 38:6-8, 1997). METHODS: This work addresses the problem of marking the end to PGES in EEG data, extracted from patients during a clinically supervised seizure. This work proposes a sensitivity analysis on EEG window size/delay, feature extraction and classifiers along with associated hyperparameters. The resulting sensitivity analysis includes the Gradient Boosted Decision Trees and Random Forest classifiers trained on 10 extracted features rooted in fundamental EEG behavior using an EEG specific feature extraction process (pyEEG) and 5 different window sizes or delays (Bao et al. in Comput Intell Neurosci 2011:1687-5265, 2011). RESULTS: The machine learning architecture described above scored a maximum AUC score of 76.02% with the Random Forest classifier trained on all extracted features. The highest performing features included SVD Entropy, Petrosan Fractal Dimension and Power Spectral Intensity. CONCLUSION: The methods described are effective in automatically marking the end to PGES. Future work should include integration of these methods into the clinical setting and using the results to be able to predict a patient's SUDEP risk.

Automated detection of activity onset after postictal generalized EEG suppression.

BACKGROUND: Sudden unexpected death in epilepsy (SUDEP) is a leading cause of premature death in patients with epilepsy. If timely assessment of SUDEP risk can be made, early interventions for optimized treatments might be provided. One of the biomarkers being investigated for SUDEP risk assessment is postictal generalized EEG suppression [postictal generalized EEG suppression (PGES)]. For example, prolonged PGES has been found to be associated with a higher risk for SUDEP. Accurate characterization of PGES requires correct identification of the end of PGES, which is often complicated due to signal noise and artifacts, and has been reported to be a difficult task even for trained clinical professionals. In this work we present a method for automatic detection of the end of PGES using multi-channel EEG recordings, thus enabling the downstream task of SUDEP risk assessment by PGES characterization. METHODS: We address the detection of the end of PGES as a classification problem. Given a short EEG snippet, a trained model classifies whether it consists of the end of PGES or not. Scalp EEG recordings from a total of 134 patients with epilepsy are used for training a random forest based classification model. Various time-series based features are used to characterize the EEG signal for the classification task. The features that we have used are computationally inexpensive, making it suitable for real-time implementations and low-power solutions. The reference labels for classification are based on annotations by trained clinicians identifying the end of PGES in an EEG recording. RESULTS: We evaluated our classification model on an independent test dataset from 34 epileptic patients and obtained an AUreceiver operating characteristic (ROC) (area under the curve) of 0.84. We found that inclusion of multiple EEG channels is important for better classification results, possibly owing to the generalized nature of PGES. Of among the channels included in our analysis, the central EEG channels were found to provide the best discriminative representation for the detection of the end of PGES. CONCLUSION: Accurate detection of the end of PGES is important for PGES characterization and SUDEP risk assessment. In this work, we showed that it is feasible to automatically detect the end of PGES-otherwise difficult to detect due to EEG noise and artifacts-using time-series features derived from multi-channel EEG recordings. In future work, we will explore deep learning based models for improved detection and investigate the downstream task of PGES characterization for SUDEP risk assessment.

Learning to detect the onset of slow activity after a generalized tonic-clonic seizure.

BACKGROUND: Sudden death in epilepsy (SUDEP) is a rare disease in US, however, they account for 8-17% of deaths in people with epilepsy. This disease involves complicated physiological patterns and it is still not clear what are the physio-/bio-makers that can be used as an indicator to predict SUDEP so that care providers can intervene and treat patients in a timely manner. For this sake, UTHealth School of Biomedical Informatics (SBMI) organized a machine learning Hackathon to call for advanced solutions https://sbmi.uth.edu/hackathon/archive/sept19.htm . METHODS: In recent years, deep learning has become state of the art for many domains with large amounts data. Although healthcare has accumulated a lot of data, they are often not abundant enough for subpopulation studies where deep learning could be beneficial. Taking these limitations into account, we present a framework to apply deep learning to the detection of the onset of slow activity after a generalized tonic-clonic seizure, as well as other EEG signal detection problems exhibiting data paucity. RESULTS: We conducted ten training runs for our full method and seven model variants, statistically demonstrating the impact of each technique used in our framework with a high degree of confidence. CONCLUSIONS: Our findings point toward deep learning being a viable method for detection of the onset of slow activity provided approperiate regularization is performed.