Continuous EEG Monitoring Research Articles

A Commentary on Electrographic Seizure Management and Clinical Outcomes in Critically Ill Children

Continuous EEG (cEEG) monitoring is the gold standard for detecting electrographic seizures in critically ill children and the current consensus-based guidelines recommend urgent cEEG to detect electrographic seizures that would otherwise be undetected. The detection of seizures usually leads to the use of antiseizure medications, even though current evidence that treatment leads to important improvements in outcomes is limited, raising the question of whether the current strategies need re-evaluation. There is emerging evidence indicating that the presence of electrographic seizures is not associated with unfavorable neurological outcome, and thus treatment is unlikely to alter the outcomes in these children. However, a high seizure burden and electrographic status epilepticus is associated with unfavorable outcome and the treatment of status epilepticus is currently warranted. Ultimately, outcomes are more likely a function of etiology than of a direct effect of the seizures themselves. We suggest re-examining our current consensus toward aggressive treatment to abolish all electrographic seizures and recommend a tailored approach where therapeutic interventions are indicated when seizure burden breaches above a critical threshold that may be associated with adverse outcomes. Future studies should explicitly evaluate whether there is a positive impact of treating electrographic seizures or electrographic status epilepticus in order to justify continuing current approaches.

Implications of stimulus-induced, rhythmic, periodic, or ictal discharges (SIRPIDs) in hospitalized patients.

Stimulus-induced electroencephalographic (EEG) patterns are commonly seen in acutely ill patients undergoing continuous EEG monitoring. Despite ongoing investigations, the pathophysiology, therapeutic and prognostic significance of stimulus-induced rhythmic, periodic or ictal discharges (SIRPIDs) and how it applies to specific pathologies remain unclear. We aimed to investigate the clinical implications of SIRPIDs in hospitalized patients. This is a retrospective single-center study of hospitalized patients from May 2016 to August 2017. We included patients above the age of 18 years who underwent >16 h of EEG monitoring during a single admission. We excluded patients with cardiac arrest and anoxic brain injury. Demographic data were obtained as well as admission GCS, and discharge modified Rankin Score (mRS). EEGs were reviewed for background activity in addition to epileptiform, periodic, and rhythmic patterns. The presence or absence of SIRPIDs was recorded. Our outcome was discharge mRS defined as good outcome, mRS 0-4, and poor outcome mRS, 5-6. A total of 351 patients were included in the final analysis. The median age was 63 years and 175 (50%) were women. SIRPIDs were identified in 82 patients (23.4%). Patients with SIRPIDs had a median initial GCS of 12 (IQR, 6-15) and a length of stay of 12 days (IQR, 6-15). They were more likely to have absent posterior dominant rhythm, decreased reactivity, and more likely to have spontaneous periodic and rhythmic patterns and higher frequency of burst suppression. After adjusting for baseline clinical variables, underlying disease type and severity, and EEG background features, the presence of SIRPIDs was also associated with poor outcomes classified as MRS 5 or 6 (OR 4.75 [2.74-8.24] p ≤ 0.0001). In our cohort of hospitalized patients excluding anoxic brain injury, SIRPIDs were identified in 23.4% and were seen most commonly in patients with primary systemic illness. We found SIRPIDs were independently associated with poor neurologic outcomes. Several studies are indicated to validate these findings and determine the risks vs. benefits of anti-seizure treatment.

Reproducibility of Test-Retest Cortical Evoked Responses in Patients with Focal Epilepsy

This exploratory study examined the short-term reliability of cortical auditory evoked responses recorded from patients undergoing whole-head scalp elec-troencephalography (EEG) monitoring to assess their candidacy for surgical treatment of intractable focal seizures. Participants were 26 patients with either left-sided (N = 13) or right-sided focal epilepsy admitted to the hospital for continuous scalp EEG monitoring for possible epilepsy surgery planning. Cortical auditory evoked responses were recorded over multiple days from scalp EEG electrodes using tones presented binaurally in a passive oddball paradigm. Test-retest intervals were 1 - 6 days (mean 2 days). Test-retest reproducibility of the auditory N1 response was assessed by paired t-test (latency) and cross-correlation analysis (amplitude and latency). Within-patient comparisons of test-retest auditory N1 peak latencies revealed no significant differences. The cross-correlation coefficient indicated high test-retest reproducibility of the N1 waveform (rcc = 0.88). Seizure lateralization was not associated with asymmetries in N1 latencies or amplitudes. An N1 amplitude asymmetry (right > left) in patients with focal seizures originating from the left hemisphere was initially observed, but disappeared when patients with prior resections were excluded, suggesting that reduced left hemisphere tissue volume may account for the smaller N1 amplitudes. Test-retest reliability of cortical auditory evoked responses was unexpectedly high in patients with focal epilepsy regardless of seizure lateralization or localization. These findings challenge the view that neural responses are intrinsically unstable (unreliable) in patients with seizures.

Early EEG monitoring predicts clinical outcome in patients with moderate to severe traumatic brain injury

There is a need for reliable predictors in patients with moderate to severe traumatic brain injury to assist clinical decision making. We assess the ability of early continuous EEG monitoring at the intensive care unit (ICU) in patients with traumatic brain injury (TBI) to predict long term clinical outcome and evaluate its complementary value to current clinical standards. We performed continuous EEG measurements in patients with moderate to severe TBI during the first week of ICU admission. We assessed the Extended Glasgow Outcome Scale (GOSE) at 12 months, dichotomized into poor (GOSE 1–3) and good (GOSE 4–8) outcome. We extracted EEG spectral features, brain symmetry index, coherence, aperiodic exponent of the power spectrum, long range temporal correlations, and broken detailed balance. A random forest classifier using feature selection was trained to predict poor clinical outcome based on EEG features at 12, 24, 48, 72 and 96 h after trauma. We compared our predictor with the IMPACT score, the best available predictor, based on clinical, radiological and laboratory findings. In addition we created a combined model using EEG as well as the clinical, radiological and laboratory findings. We included hundred-seven patients. The best prediction model using EEG parameters was found at 72 h after trauma with an AUC of 0.82 (0.69–0.92), specificity of 0.83 (0.67–0.99) and sensitivity of 0.74 (0.63–0.93). The IMPACT score predicted poor outcome with an AUC of 0.81 (0.62–0.93), sensitivity of 0.86 (0.74–0.96) and specificity of 0.70 (0.43–0.83). A model using EEG and clinical, radiological and laboratory parameters resulted in a better prediction of poor outcome (p < 0.001) with an AUC of 0.89 (0.72–0.99), sensitivity of 0.83 (0.62–0.93) and specificity of 0.85 (0.75–1.00). EEG features have potential use for predicting clinical outcome and decision making in patients with moderate to severe TBI and provide complementary information to current clinical standards.

A Case of Pembrolizumab (Anti-PD-1) Induced Encephalitis

Objective N/A. Background PD-1 Immune checkpoint inhibitors (ICI) have been associated with neurologic immune-related adverse events including meningoencephalitis and limbic encephalitis that can manifest as paraneoplastic syndromes. We present a case of suspected pembrolizumab (anti PD-1) induced limbic encephalitis presenting as episodic aphasia. Design/Methods N/A. Results A 62-year-old man with poorly differentiated papillary thyroid carcinoma with extensive brain, lung, lymphatic metastases post thyroidectomy, radioiodine therapy, chemotherapy with dabrafenib and trametinib, and stereotactic radiation surgery presented with recurrence of brain metastases one year after the diagnosis of a 2.1 cm metastatic lesion in the left parieto-occipital region causing right homonymous hemianopsia. The metastases initially improved in size and number with radiation and serial brain MRIs were stable. Biopsy of the mets revealed a poorly differentiated carcinoma. He was subsequently treated with pembrolizumab for six months. Two months after treatment initiation, he reported episodic behavioral arrest, confusion, and expressive aphasia concerning for seizures. Continuous EEG monitoring revealed a left-sided focus without seizures, and the episodes persisted despite levetiracetam and clobazam. Four lumbar punctures revealed lymphocytic pleocytosis (10-14 cells), elevated protein (48–68 mg/dl), negative cytology, flow cytometry and viral studies including JC virus. Paraneoplastic panels in serum and cerebrospinal fluid were negative. Repeat MRIs findings were most consistent with radionecrosis and noted improvement of the metastatic lesions. Suspicion was raised for pembrolizumab-induced encephalitis and he received high-dose steroids with minimal response however, clinical improvement noted with reduced episode frequency after intravenous immunoglobulin induction therapy and rituximab maintenance therapy. Conclusions PD-1 ICI-related encephalitis is a diagnosis of exclusion that should be considered in patients with encephalopathy or other neurological deficits following 3 months of treatment initiation and response to immunosuppressive therapy. Higher incidences are reported in males. Early recognition is crucial to prevent long-term neurologic damage. Outcomes are depend on patient characteristics and clinical presentation.

Continuous EEG of memories of Near-death and mystical experiences: Preliminary research

We recently recorded, processed, and published brain activity by quantitative electroencephalography (QEEG) and quantitative electroencephalography (QEEGt), comparing separate subjects remembering their near-death experience (NDE) and mystical experience. Our studies refer to a mystical experience as a spiritual contemplative experience (SCE). We found neural and Greyson Scale (GS) correlations between them. This study is a real-time enhancement of that study using the emerging technology of continuous EEG monitoring (CEEG), which intensive care unit (ICU) professionals use to identify malignant EEG patterns quickly and provide care effectively. EEG monitoring encompasses a wide range of technical and clinical issues in successfully monitoring critically ill patients to detect significant changes in cerebral function and prevent serious neuronal injury over time. Because the brain undergoes continuous and dynamic changes, CEEG was considered an important method for real-time monitoring of functional brain changes while remembering near-death experiences (NDE) and spiritual contemplative experience (SCE). The CEEG envelope showed an incremental amplitude coinciding with NDE and SCE remembering the time, but an increase of amplitude was greater for SCE. CEEG also demonstrated a greater amplitude in frontal lobes for SCE. Statistical increments of absolute power for alpha and gamma bands were demonstrated during both NDE and SCE. We conclude that CEEG is a useful method for continuously assessing dynamic changes while remembering NDE and SCE.

Continuous autonomic nervous system assessment using heart rate variability methodology during near-death and mystical experiences

Introduction: We have recently recorded, processed, and published preliminary research (Part I) on brain activity by quantitative electroencephalography (EEG) tomography (QEEGt), comparing separate subjects remembering their near-death experiences (NDEs) and mystical experiences (SCE). Several reports have affirmed that NDE and SCE are related to important functional changes in the autonomic nervous system (ANS). The autonomic nervous system and the hypothalamus regulate pulse, blood pressure, breathing, and arousal in response to emotional cues. When activated, the sympathetic nervous system prepares the body for emergency actions by controlling the glands of the endocrine system. Heart rate variability (HRV) refers to how much an individual's heart rate (HR) varies, and it is a powerful method to assess the ANS. Several reports have affirmed that NDE and SCE are related to important functional changes in the ANS. We used the continuous EEG monitoring (CEEG) system to compare the memories of two groups with a NDE and a SCE. CEEG permits continuous electrocardiogram monitoring, allowing calculation of HR and HRV during SCE and NDE remembering. Hence, using HRV methodology, it is possible to assess the emotional effect of remembering NDE and SCE. Conclusion: We demonstrated the usefulness of using the CEEG methodology, which allows us to continuously assess the ANS through the HRV methodology, showing important significant functional changes in the autonomic nervous system (ANS), and comparing SCE and NDE.

Quantitative artifact reduction and pharmacologic paralysis improve detection of EEG epileptiform activity in critically ill patients

ObjectiveEpileptiform activity is common in critically ill patients, but movement-related artifacts—including electromyography (EMG) and myoclonus—can obscure EEG, limiting detection of epileptiform activity. We sought to determine the ability of pharmacologic paralysis and quantitative artifact reduction (AR) to improve epileptiform discharge detection. MethodsRetrospective analysis of patients who underwent continuous EEG monitoring with pharmacologic paralysis. Four reviewers read each patient’s EEG pre- and post- both paralysis and AR, and indicated the presence of epileptiform discharges. We compared the interrater reliability (IRR) of identifying discharges at baseline, post-AR, and post-paralysis, and compared the performance of AR and paralysis according to artifact type. ResultsIRR of identifying epileptiform discharges at baseline was slight (N = 30; κ = 0.10) with a trend toward increase post-AR (κ = 0.26, p = 0.053) and a significant increase post-paralysis (κ = 0.51, p = 0.001). AR was as effective as paralysis at improving IRR of identifying discharges in those with high EMG artifact (N = 15; post-AR κ = 0.63, p = 0.009; post-paralysis κ = 0.62, p = 0.006) but not with primarily myoclonus artifact (N = 15). ConclusionsParalysis improves detection of epileptiform activity in critically ill patients when movement-related artifact obscures EEG features. AR improves detection as much as paralysis when EMG artifact is high, but is ineffective when the primary source of artifact is myoclonus. SignificanceIn the appropriate setting, both AR and paralysis facilitate identification of epileptiform activity in critically ill patients.

NCOG-38. PEMBROLIZUMAB (ANTI-PD-1) INDUCED LIMBIC ENCEPHALITIS PRESENTING AS EPISODIC APHASIA

Abstract PD-1 immune checkpoint inhibitors (ICI) have been associated with neurological immune-related adverse events including meningioencephalitis and limbic encephalitis that can manifest as paraneoplastic syndromes. We present a case of pembrolizumab (Anti-PD-1) induced limbic encephalitis presenting as episodic aphasia. A 62-year-old man with poorly differentiated papillary thyroid carcinoma with extensive brain, lung, lymphatic metastases, post total-thyroidectomy, radioiodine therapy, chemotherapy with dabrafenib and trametinib, and stereotactic radiation presented with recurrence of brain metastasis 1 year after diagnosis of a 2.1 cm metastatic lesion in the left parieto-occipital region. The metastases had initially improved in size and number with radiation, followed on serial MRI Brain - that were stable. Biopsy of recurrence revealed a poorly differentiated cancer. He was subsequently treated with pembrolizumab for 6 months for refractory carcinoma. Two months after treatment initiation, he reported episodic behavioral arrest, confusion, and expressive aphasia, initially concerning for seizures. Continuous EEG monitoring revealed non-epileptogenic left-sided foci and no seizures were captured. Patient empirically started on levetiracetam and clobazam - however, episodes persisted. Lumbar puncture revealed lymphocytic pleocytosis (10-14 cells), elevated protein (48-68 mg/dL), negative cytology, flow cytometry, and viral studies including JC virus. Paraneoplastic panel in serum and cerebrospinal fluid were negative. Repeat MRI findings were most consistent with radionecrosis and on contrary improvement of the previous metastatic lesions. This rose suspicion for pembrolizumab-induced encephalitis and he received high dose steroids with minimal response but great clinical improvement with IVIG and induction/maintenance therapy with rituximab. Marked reduction in aphasic episodes were noticed thereafter. PD-1 ICI-related encephalitis is diagnosis of exclusion that should be considered in patients with encephalopathy or neurological deficit following 3 months of treatment initiation and response to immunosuppressive therapy. Higher incidences reported in males. Early recognition is crucial to prevent long-term neurological damage. Outcomes depend on patient characteristics and timeline of presentation.

Clinical profile of a cohort of neonates with seizures: Association between semiology, etiology, and electroencephalographic findings.

Seizures are the most common sign of neurologic dysfunction, reflecting a wide variety of central nervous system disorders. A retrospective cross-sectional study of neonates with a clinical diagnosis of seizures was conducted in order to verify relationships between clinical aspects and EEG findings. Patients were divided into 3 groups according to the EEG recording available as: 1) with confirmatory ictal EEG; 2) with altered but non-ictal EEG; and 3) without any EEG recording. Variables related to pregnancy and birth history, neonatal complications, and seizure semiology (by video or clinical description) were compared to EEG findings. 97 neonates were included (39.1% preterm, 54.6% male), 71 with available EEG data (56.3% with ictal EEG). The group without EEG presented clinical characteristics significantly different from the others such as extreme prematurity, low birth weight, and higher neonatal mortality (P = 0.002, 0.001, and 0.003, respectively). The most common etiology was hypoxic-ischemic encephalopathy (HIE) (46.4%) followed by vascular disorders, which predominated in extremely preterm neonates (P = 0.006). Sequential seizure was the most common type (44.6%) and was more frequently identified in term neonates (46%). In 51.2% of the ictal recordings the main finding was electrographic seizure without clinical manifestation. Discharge using antiseizure medication was higher among those with ictal or altered non-ictal EEG (P < 0.001). HIE is still a frequently etiology for neonatal seizures. Even if the patients in the sample were not under continuous EEG, the substantial proportion of electrographic seizures without clinical manifestations detected suggests the importance of continuous EEG monitoring in neonates at increased risk of seizures.

Management and prognosis of pediatric status epilepticus

Abstract Background Pediatric status epilepticus is a neurological emergency with the potential for severe developmental and neurological consequences. Prompt diagnosis and management are necessary. Objectives To outline the existing best available evidence for managing pediatric and neonatal status epilepticus, in the light of emerging randomized controlled studies. We also focus on short and long-term prognoses. Materials and methods This is a systematic overview of the existing literature. Results Status epilepticus, its treatment, and prognosis are usually based on the continuation of seizure activity at 5 and 30 min. Refractory and super-refractory status epilepticus further complicates management and requires continuous EEG monitoring with regular reassessment and adjustment of therapy. Benzodiazepines have been accepted as the first line of treatment on the basis of reasonable evidence. Emerging randomized controlled trials demonstrate equal efficacy for parenterally administered phenytoin, levetiracetam, and valproic acid as second-line agents. Beyond this, the evidence for third-line options is sparse. However, encouraging evidence for midazolam and ketamine exists with further data required for immunological, dietary, and surgical interventions. Conclusion Our overview of the management of pediatric and neonatal status epilepticus based on available evidence emphasizes the need for evidence-based guidelines to manage status epilepticus that fails to respond to second-line treatment.

Continuous EEG monitoring detects nonconvulsive seizure and Ictal-Interictal Continuum abnormalities in moderate to severe ICANS following systemic CAR-T therapy.

Immune Cell Effector Associated Neurotoxicity Syndrome (ICANS) is common amongst patients receiving CD19 targeted Chimeric Antigen Receptor T-cell (CAR-T) therapy. The purpose of this study is to characterize the incidence of seizures and ictal-interictal continuum (IIC) abnormalities in patients with ICANS. Retrospective review of consecutive patients treated with axicabtagene ciloleucel (axi-cel) for recurrent high-grade systemic lymphoma at Stanford Medical Center between 2/2016-6/2019. Electronic medical records (EMR) were reviewed for clinical features, treatment information, EEG data, CRS (cytokine release syndrome)/ICANS severity, and clinical outcomes. Fifty-six patients met inclusion criteria. 85.7% of patients developed CRS, and 58.9% developed ICANS. Twenty-eight patients had EEG monitoring, of whom 26 had ICANS. Median duration of EEG monitoring was 30 hours (range .5-126 hours). Four patients (7.1%) had seizures (1 patient had a clinical generalized seizure, 2 patients had clinical and nonconvulsive seizures, and 1 patient had an isolated non-convulsive seizure). Ictal-interictal continuum abnormalities were common, of which generalized periodic discharges (GPDs) with triphasic morphology and GPDs with epileptiform morphology were most frequently seen. Generalized periodic discharges with triphasic wave morphology were found across Grade 2-3 peak ICANS severity, however the majority (86%) of patients with epileptiform GPDs had Grade 3 peak ICANS severity. Among patients receiving axi-cel, seizure occurred in 7.1% of the total cohort, representing 12% of patients with ICANS. Ictal-interictal continuum abnormalities are also seen in patients with ICANS, most commonly GPDs. 75% of patients with seizures had nonconvulsive seizures supporting the use of continuous video EEG monitoring in this population.

Acute Symptomatic Seizure Associated With Chronic Antiseizure Medication Use After Stroke.

Patients with acute symptomatic seizures (ASyS) after stroke are discharged on antiseizure medications (ASMs) and stay on them for an extended period. We analyzed the current ASM management practice, 6 months, and at the last follow-up after stroke-related ASyS concerns to identify chronic and long-term ASM use predictors. A single-center, retrospective cohort study of adults who underwent continuous EEG monitoring for ASyS concerns after stroke (April 1, 2012 to March 31, 2018) with at least 6 months of follow-up was performed. ASM use beyond 6 months after the initial ASyS concern was defined as "chronic" among patients discharged on them. "Long-term" ASM use at the last follow-up in all patients with ASyS concerns was analyzed. Logistic regression and Cox regression multivariable modeling to analyze predictors of "chronic" and "long-term" ASM use, respectively, was performed. A total of 465 (mean age 61.7 ± 13.3 years and 52% female patients) patients (41.9% ischemic stroke, 36.1% intracerebral hemorrhage, and 21.9% subarachnoid hemorrhage) were included. Of the 179 (38.5%) patients discharged on ASMs, 132 (73.7%; 28.4% of study population) had chronic ASM use, despite 90% not experiencing any seizure (poststroke epilepsy [PSE]) during this time. The independent predictors of chronic ASM use were electrographic ASyS (odds ratio [OR] = 9.27, 95% CI = 2.53-60.4) and female sex (OR = 2.2, 95% CI = 1.02-4.83). After a median 61-month (5.1 years) follow-up, 101 (21.7%) patients in the study population were on long-term ASM use, including 67 (14.4%) who developed PSE. Long-term ASM use was associated with NIH Stroke Scale Score (OR = 1.5, 95% CI = 1.015-1.98), cortical involvement (OR = 1.28, 95% CI = 1.02-1.6), convulsive ASyS (OR = 1.46, 95% CI = 1.02-2.09), epileptiform findings on outpatient EEG (OR = 4.03, 95% CI = 1.28-12.76), and PSE development (OR = 7.06, 95% CI = 3.7-13.4). Chronic ASM use is highly associated with electrographic, rather than convulsive, ASyS. However, long-term ASM use is independently associated with PSE and its risk factors, including convulsive ASyS. With the ubiquity of stroke-related ASyS concerns in routine clinical practice, comparative effectiveness studies to guide ASM management are needed.

Etiology matters for neuroprognostication: A multimodal electrophysiological investigation in a case of Bickerstaff's brainstem encephalitis

We report the case of a 19-year-old patient with an acute-onset non-traumatic coma. Brain MRI scan was normal, CSF showed mild pleocytosis and moderately elevated protein, and continuous EEG-monitoring was compatible with spindle-coma. Cortical somatosensory evoked potentials (SSEPs) and middle-latency auditory evoked potentials (MLAEPs) were bilaterally absent, and brainstem auditory evoked potentials suggested a brainstem dysfunction. Serum anti-GQ1b and anti-GT1a IgG antibodies positivity suggested Bickerstaff's brainstem encephalitis (BBE). The clinical and functional outcomes were favorable and normal cortical SSEPs/MLAEPs reappeared in a few weeks. Based on this report, in cases of unexplained MRI-negative coma with neurophysiological evidence of brainstem dysfunction, BBE should be eliminated before considering withdrawal of life-sustaining therapy (WLST).

Power spectrum analysis and outcomes of non-convulsive status epilepticus: a single-center study.

Diagnosis of non-convulsive status epilepticus (NCSE) is challenging and outcomes during follow-up are not clear. This study aimed to conduct power spectrum analysis in NCSE and measure outcomes of patients. We searched continuous EEG monitoring (cEEG) recordings to identify patients of NCSE. An artifact-free cEEG epoch of continuous 60s was chosen for spectral power analysis. We also collected electronic medical records of the patients for extracting clinical information. Patients recruited were followed up at least every half a year. There were 48 patients with 64 independent NCSE episodes during different course of disease recruited in the study, with a mean age of 40.3 ± 19.1years (range, 12-72years), including 24 males (50%) and 24 females (50%). When the spectral power of 60s equaled to 11.30 μV2 for predicting impairment of consciousness, (sensitivity, specificity) = (0.979, 0.625). When the spectral power of 60s equaled to 52.70 μV2 for predicting myoclonic jerks, (sensitivity, specificity) = (0.783, 0.756). There were 27 patients (56.3%) followed up with a duration over 12months. Nineteen patients (70.4%) continued to have seizures. Eleven (40.7%) resisted to at least two kinds of appropriate anti-seizure medication at maximum tolerated levels. Five patients with prolonged NCSE suffered from loss of brain parenchymal volume on follow-up MRI scans. Spectral power analysis can be used to detect mental status and limb jerks. Early diagnosis and treatment of NCSE are important, which can influence outcomes of the patients during follow-up.

Seizure Burden, EEG, and Outcome in Neonates With Acute Intracranial Infections: A Prospective Multicenter Cohort Study

BackgroundLimited data exist regarding seizure burden, electroencephalogram (EEG) background, and associated outcomes in neonates with acute intracranial infections. MethodsThis secondary analysis was from a prospective, multicenter study of neonates enrolled in the Neonatal Seizure Registry with seizures due to intracranial infection. Sites used continuous EEG monitoring per American Clinical Neurophysiology Society guidelines. High seizure burden was defined a priori as seven or more EEG-confirmed seizures. EEG background was categorized using standardized terminology. Primary outcome was neurodevelopment at 24-months corrected age using Warner Initial Developmental Evaluation of Adaptive and Functional Skills (WIDEA-FS). Secondary outcomes were postneonatal epilepsy and motor disability. ResultsTwenty-seven of 303 neonates (8.9%) had seizures due to intracranial infection, including 16 (59.3%) bacterial, 5 (18.5%) viral, and 6 (22.2%) unknown. Twenty-three neonates (85%) had at least one subclinical seizure. Among 23 children with 24-month follow-up, the WIDEA-FS score was, on average, 23 points lower in children with high compared with low seizure burden (95% confidence interval, [−48.4, 2.1]; P = 0.07). After adjusting for gestational age, infection etiology, and presence of an additional potential acute seizure etiology, the effect size remained unchanged (β = −23.8, P = 0.09). EEG background was not significantly associated with WIDEA-FS score. All children with postneonatal epilepsy (n = 4) and motor disability (n = 5) had high seizure burden, although associations were not significant. ConclusionHigh seizure burden may be associated with worse neurodevelopment in neonates with intracranial infection and seizures. EEG monitoring can provide useful management and prognostic information in this population.

Personalized model to predict seizures based on dynamic and static continuous EEG monitoring data

Background/ObjectiveEarly recognition of patients who may be at risk of developing acute symptomatic seizures would be useful. We aimed to determine whether continuous electroencephalography (cEEG) data using machine learning techniques such as neural networks and decision trees could predict seizure occurrence in hospitalized patients. MethodsThis was a single center retrospective cohort analysis of cEEG data in patients aged 18–90 years who were admitted and underwent cEEG monitoring between 2010 and 2019 limited to 72 h excluding those who were seizing at the onset of recording. A total of 41,491 patients were reviewed; of these, 3874 were used to develop the static model and 1687 to develop the dynamic model (half with seizure and half without seizure in each cohort). Of these, 80% were randomly selected as derivation cohorts for each model and 20% were randomly selected as validation cohorts. Dynamic and static machine learning models (long short term memory (LSTM) and Extreme Gradient Boosting algorithm (XGBoost)) based on day-to-day dynamic EEG changes and binary static EEG features over the prior 72 h or until seizure, which ever was earlier, were used. ResultsThe static model was able to predict seizure occurrence based on cEEG data with sensitivity and specificity of 0.81 and 0.59, respectively, with an AUC of 0.70. The dynamic model was able to predict seizure occurrence with sensitivity and specificity of 0.72 and 0.80, respectively, and AUC of 0.81. ConclusionsMachine learning models could be applied to cEEG data to predict seizure occurrence based on available cEEG data. Dynamic day-to-day EEG data are more useful in predicting seizures than binary static EEG data. These models could potentially be used to determine the need for ongoing cEEG monitoring and to prioritize resources.

Continuous EEG Monitoring in Critically Ill Children and Prognostic Factors for Short-term Outcome: An Observational Study

Continuous EEG Monitoring in Critically Ill Children and Prognostic Factors for Short-term Outcome: An Observational Study

Use of Continuous EEG Monitoring and Short-Term Outcomes in Critically Ill Children

AbstractThis study aimed to compare short-term outcomes at pediatric intensive care unit (PICU) discharge in critically ill children with and without continuous electroencephalography (cEEG) monitoring. We retrospectively compared 234 patients who underwent cEEG with 2294 patients without cEEG. Propensity score matching was used to compare patients with seizures and status epilepticus between cEEG and historical cohorts. The EEG cohort had higher in-hospital mortality, worse Pediatric Cerebral Performance Category (PCPC) scores, and greater PCPC decline at discharge. In patients with status epilepticus, the PCPC decline was higher in the cEEG cohort. PCPC decline at PICU discharge was associated with cEEG monitoring in patients with status epilepticus.

Extreme value theory inspires explainable machine learning approach for seizure detection

Epilepsy is one of the brightest manifestations of extreme behavior in living systems. Extreme epileptic events are seizures, that arise suddenly and unpredictably. Usually, treatment strategies start by analyzing brain activity during the seizures revealing their type and onset mechanisms. This approach requires collecting data for a representative number of events which is only possible during the continuous EEG monitoring over several days. A big part of the further analysis is searching for seizures on these recordings. An experienced medical specialist spends hours checking the data of a single patient and needs assistance from the automative systems for seizure detection. Machine learning methods typically address this issue in a supervised fashion and exhibit a lack of generalization. The extreme value theory allows addressing this issue with the unsupervised machine learning methods of outlier detection. Here, we make the first step toward using this approach for the seizure detection. Based on our recent work, we specified the EEG features showing extreme behavior during seizures and loaded them to the one-class SVM, a popular outlier detection algorithm. Testing the proposed approach on 83 patients, we reported 77% sensitivity and 12% precision. In 60 patients, sensitivity was 100%. In the rest 23 subjects, we observed deviations from the extreme behavior. The one-class SVM used a single subject’s data for training; therefore, it was stable against between-subject variability. Our results demonstrate an effective convergence between the extreme value theory, a physical concept, and the outlier detection algorithms, a machine learning concept, toward solving the meaningful task of medicine.