Clinical EEG Research Articles

High-Density (HD) Scalp EEG Findings in "Benign" Childhood Epilepsy with Centrotemporal Spikes (BCECTS).

Despite characteristic clinical and scalp EEG findings, BCECTS pathophysiology is unclear regarding involvement of large-scale neuronal networks. Higher number of scalp electrodes with HD-EEG may promote accurate localization of the cortical generators in BCECTS providing additional insight in those with neurocognitive problems. We aimed to determine the value of visual interpretation of topographical maps using 256 channels (when compared to standard 21 channel array) HD-EEG in BCECTS and attempted to source localize interictal discharges (IEDs) using Geosource 2 software. Patient records were reviewed for demographic, seizure, brain magnetic resonance imaging (MRI) details; scalp and HD-EEG findings. HD-EEG software was used to review raw EEG data (21 scalp EEG electrodes were compared to 256 HD-EEG electrodes); select, average, and source localize IEDs.Five BCECTS patients with HD-EEG were identified. Seizure onset age ranged from 5-11 years with 1-18 lifetime seizures; both focal (n = 3) and focal to bilateral tonic-clonic (n = 2). Neurocognitive co-morbidities noted in our cohort included attention deficit hyperactivity disorder, speech/developmental delay and a specific learning disorder. Scalp EEG showed typical findings with IEDs over the centrotemporal regions (bilateral n = 3; unilateral n = 2). Visualization and inspection of expanded coverage topographic maps with HD-EEG showed well-defined islands of maximum negativity and positivity of a dipole compared to conventional channels where boundary delineation was obscured. Further, HD-EEG localized IEDs/"spike-generator" to areas such as the pre-and post-central, middle-frontal and temporal gyrus, and the inferior parietal lobule. In BCECTS, HD-EEG may show affection for a broader neural network and may provide a better insight into the associated neurocognitive morbidities.

Effects of severe obstructive sleep apnea on functional prognosis in the acute phase of ischemic stroke and quantitative electroencephalographic markers

ObjectiveTo investigate the effect of severe obstructive sleep apnea (OSA) on functional prognosis in the acute phase and quantitative electroencephalography (EEG) markers during sleep in ischemic stroke patients. MethodsThis study included 125 mild-to-moderate acute ischemic stroke patients with OSA who underwent polysomnography (PSG) within one week of stroke onset between January 2015 and June 2020. Patients were grouped according to their apnea-hypopnea index (</≥ 30/h). Poor functional prognosis was defined as modified Rankin Scale score ≥3. The EEG recorded by PSG was extracted during different sleep stages for power spectrum analysis. The delta/alpha power ratio (DAR), (delta + theta)/(alpha + beta) ratio (the slowing ratio, TSR), and the relative power (RP) of each frequency band were calculated. Differences in clinical, PSG, and quantitative EEG characteristics were compared between the groups. Additionally, we explored predictors of poor functional prognosis. ResultsPatients with severe OSA had a higher proportion of hypertension, lower relative power of high-frequency bands, and higher delta RP, TSR, and DAR (p < 0.05). Severe OSA was associated with a 3.6-fold increase in risk of poor prognosis (p < 0.05). Increased delta RP and TSR, as well as decreased alpha, beta, and sigma RP, may be independent predictors of a poor functional prognosis. ConclusionsSevere OSA is an independent risk factor for a poor functional prognosis in patients with acute ischemic stroke, and quantitative EEG during sleep showed a significant slow wave enhancement, suggesting more severe brain dysfunction. The treatment of severe OSA may improve functional prognosis.

Suppressing interferences of EIT on synchronous recording EEG based on comb filter for seizure detection.

The purpose of this study was to eliminate the interferences of electrical impedance tomography (EIT) on synchronous recording electroencephalography (EEG) for seizure detection. The simulated EIT signal generated by COMSOL Multiphysics was superimposed on the clinical EEG signal obtained from the CHB-MIT Scalp EEG Database, and then the spectrum features of superimposed mixed signals were analyzed. According to the spectrum analysis, in addition to high-frequency interference at 51.2 kHz related to the drive current, there was also low-frequency interference caused by switching of electrode pairs, which were used to inject drive current. A low pass filter and a comb filter were used to suppress the high-frequency interference and low-frequency interference, respectively. Simulation results suggested the low-pass filter and comb filter working together effectively filtered out the interference of EIT on EEG in the process of synchronous monitoring. As a result, the normal EEG and epileptic EEG could be recognized effectively. Pearson correlation analysis further confirmed the interference of EIT on EEG was effectively suppressed. This study provides a simple and effective interference suppression method for the synchronous monitoring of EIT and EEG, which could be served as a reference for the synchronous monitoring of EEG and other medical electromagnetic devices.

A novel machine learning technique using resting‐state EEG to assess and monitor cognitive impairment

AbstractBackgroundMonitoring changes in cognitive impairment is an important aspect of care for aging patients. Traditionally, this is done with pen‐and‐paper assessments such as the Mini‐Mental State Exam (MMSE). EEG is a low‐cost and widely available neurophysiology tool that provides a data‐rich measure of brain health, but signal complexity has limited its impact. Advances in machine learning bring new possibilities to decode those complex signals and leverage EEG in patient care. Here, we engineer novel EEG features for assessing brain health with the [Banded Fractal Variability] (BFV) technique. We use [BFV] and other EEG metrics to assess patients’ cognitive function via a continuous score (a [Cognitive Impairment Index], or [CII]). We validate the diagnostic value of the EEG features by comparing our model against reduced models lacking the EEG features.MethodThe patient sample (N=97, mean (sd) age = 69.1 (11.0), 52.6% female) was collected at the Pacific Brain Health Center. Diagnostic categories included subjective cognitive impairment (n=44), mild cognitive impairment (n=26), and dementia (n=27). Mixed disease (AD, DLB, FTD, vascular, diabetes) was common in the impaired sample, and 39.6% were diagnosed with AD. We developed a continuous [CII] model trained on patients’ MMSE scores (range = 11 to 30, mean = 26.6) using a gradient boosting model with [Banded Fractal Variability] and other EEG features computed from clinical wake EEGs. Model development and testing took place within a rigorous nested cross‐validation (NCV) procedure to prevent overfitting. Performance was compared to reduced (no‐EEG features) models using the non‐parametric Wilcoxon sign‐rank test.ResultThe approach showed a Mean Absolute Error of 1.98 (sd = 0.28) across 5 test folds in the NCV outer loops, significantly outperforming both the baseline intercept‐only model (p = 0.031) and the demographics (age and years of education) model (p = 0.031).ConclusionThese preliminary results suggest that our [CII] derived from machine‐learning and quantitative EEG using [BFV] is a viable approach to assess and track patients’ cognitive impairment regardless of disease status across all impairment levels, raising the possibility that a brief EEG scan could shed light on patient health normally accessible only through extensive cognitive and neuropsychological testing.

Roadmap to EEGs: video-based e-learning modules addressing clinical EEG reading

Content available: Video

EPIAMNE: A New Scoring System for Differentiating Transient EPIleptic AMNEsia from Transient Global Amnesia.

Transient epileptic amnesia (TEA) is a rare cause of acute amnestic syndromes (AAS), often misdiagnosed as transient global amnesia (TGA). We proposed a scoring system-the EPIlepsy AMNEsia (EPIAMNE) score-using quantitative EEG (qEEG) analysis to obtain a tool for differentiating TEA from TGA. We retrospectively reviewed clinical information and standard EEGs (stEEG) of 19 patients with TEA and 21 with TGA. We computed and compared Power Spectral Density, demonstrating an increased relative theta power in TGA. We subsequently incorporated qEEG features in EPIAMNE score, together with clinical and stEEG features. ROC curve models and pairwise ROC curve comparison were used to evaluate and compare the diagnostic accuracy for TEA detection of EPIAMNE score, presence of symptoms atypical for TGA (pSymAT) and identification of anomalies (interictal epileptiform or temporal focal spiky transients) at stEEG (PosEEG). Area Under the Curve (AUC) of EPIAMNE score revealed to be higher than PosEEG and pSymAT (AUCEPIAMNE = 0.95, AUCpSymAT = 0.85, AUCPosEEG = 0.67) and this superiority proved to be statistically significant (p-valueEPIAMNE-PosEEG and p-valueEPIAMNE-pSymAT &lt; 0.05). In conclusion, EPIAMNE score classified TEA with higher accuracy than PosEEG and pSymAT. This approach could become a promising tool for the differential diagnosis of AAS, especially for early TEA detection.

CUL4B-associated epilepsy: Report of a novel truncating variant promoting drug-resistant seizures and systematic review of the literature.

Background: Cabezas syndrome is a rare X-linked disease caused by mutations in CUL4B and characterized by developmental delay/intellectual disability, somatic dysmorphisms, behavioural disorder, ataxia/tremors. Although seizures have been formerly reported, their clinical semiology, EEG features and long-term outcome are largely unknown. PurposeThis study aims to expand knowledge on epilepsy associated with Cabezas syndrome and to understand whether different types of variants in the CUL4B gene or brain MRI abnormalities may influence seizure onset and epilepsy course. MethodsWith this in mind, we characterised the epileptic phenotype of a 17-year-old adolescent harbouring a CUL4B novel variant and performed a systematic literature review of CUL4B-associated seizures, analysing mutation types and neuroimaging features as epilepsy predictors. ResultsOur case observation indicates that CUL4B-associated epilepsy may also be drug-resistant and persist beyond infancy. Literature analysis shows that 43% of CUL4B patients develop seizures, with no statistically significant differences in epilepsy development according to mutation type and neuroimaging features. ConclusionOur study extends knowledge of CUL4B-associated epilepsy, offering new insights into disease progression.

Clinical EEG of Rett Syndrome: Group Analysis Supplemented with Longitudinal Case Report.

Rett syndrome (RTT), a severe neurodevelopmental disorder caused by MECP2 gene abnormalities, is characterized by atypical EEG activity, and its detailed examination is lacking. We combined the comparison of one-time eyes open EEG resting state activity from 32 girls with RTT and their 41 typically developing peers (age 2-16 years old) with longitudinal following of one girl with RTT to reveal EEG parameters which correspond to the RTT progression. Traditional measures, such as epileptiform abnormalities, generalized background activity, beta activity and the sensorimotor rhythm, were supplemented by a new frequency rate index measured as the ratio between high- and low-frequency power of sensorimotor rhythm. Almost all studied EEG parameters differentiated the groups; however, only the elevated generalized background slowing and decrease in our newly introduced frequency rate index which reflects attenuation in the proportion of the upper band of sensorimotor rhythm in RTT showed significant relation with RTT progression both in longitudinal case and group analysis. Moreover, only this novel index was linked to the breathing irregularities RTT symptom. The percentage of epileptiform activity was unrelated to RTT severity, confirming previous studies. Thus, resting EEG can provide information about the pathophysiological changes caused by MECP2 abnormalities and disease progression.

Selection of the optimal channel configuration for implementing wearable EEG devices for the diagnosis of mild cognitive impairment

BackgroundEarly diagnosis of mild cognitive impairment (MCI) is essential for timely treatment planning. With recent advances in the wearable technology, interest has increasingly shifted toward computer-aided self-diagnosis of MCI using wearable electroencephalography (EEG) devices in daily life. However, no study so far has investigated the optimal electrode configurations for the efficient diagnosis of MCI while considering the design factors of wearable EEG devices. In this study, we aimed to determine the optimal channel configurations of wearable EEG devices for the computer-aided diagnosis of MCI.MethodWe employed an EEG dataset collected from 21 patients with MCI and 21 healthy control subjects. After evaluating the classification accuracies for all possible electrode configurations for the two-, four-, six-, and eight-electrode conditions using a support vector machine, the optimal electrode configurations that provide the highest diagnostic accuracy were suggested for each electrode condition.ResultsThe highest classification accuracies of 74.04% ± 4.82, 82.43% ± 6.14, 86.28% ± 2.81, and 86.85% ± 4.97 were achieved for the optimal two-, four-, six-, and eight-electrode configurations, respectively, which demonstrated the possibility of precise machine-learning-based diagnosis of MCI with a limited number of EEG electrodes. Additionally, further simulations with the EEG dataset revealed that the optimal electrode configurations had significantly higher classification accuracies than commercial EEG devices with the same number of electrodes, which suggested the importance of electrode configuration optimization for wearable EEG devices based on clinical EEG datasets.ConclusionsThis study highlighted that the optimization of the electrode configuration, assuming the wearable EEG devices can potentially be utilized for daily life monitoring of MCI, is necessary to enhance the performance and portability.

Sporadic Creutzfeldt–Jakob Disease in the young (50 and below): 10-year review of United Kingdom surveillance

IntroductionSporadic Creutzfeldt–Jakob Disease (sCJD) is the commonest human prion disease, with a median age of onset of 68 years. We characterise the clinical, investigation, and neuropathological features in young individuals with sCJD using data from UK national CJD surveillance.MethodsReferrals between 2011 and 2021 were examined, with definite (post-mortem confirmed) or probable sCJD cases included. Clinical features, MRI, EEG, CSF RT-QuIC, 14-3-3, PRNP sequencing and neuropathological findings were examined. We compared younger (≤ 50 years age of onset) with older individuals. Records of Non-sCJD referrals were also reviewed.Results46 (4%) young individuals were identified (age at onset 25–50) from 1178 cases. 15 (33%) were autopsy confirmed. Psychiatric disturbance (37% vs 22%, p = 0.02) and headache (11% vs 3%, p = 0.01) at presentation, and longer disease duration (by 1.45 months, 95% CI 0.43–2.79, logrank p = 0.007) were commoner. CSF RT-QuIC showed lower sensitivity (82% vs 93%, p = 0.02). There was no difference in sensitivity of MR brain or CSF 14-3-3. There were no significant co-pathologies in autopsy-confirmed cases. For non-sCJD referrals, 41 cases were of other CJD subtypes, and 7 non-prion diagnoses.ConclusionsYoung-onset sCJD is more likely to present with neuropsychiatric symptoms and headache, longer disease duration, and lower sensitivity of RT-QuIC. These findings may be driven by the underlying molecular subtypes. Our results guide the evaluation of younger individuals presenting with rapidly progressive cognitive, neuropsychiatric, and motor decline, and emphasise the need for additional vigilance for atypical features by clinicians and CJD surveillance programmes worldwide.

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.

Neurodevelopmental and Epilepsy Phenotypes in Individuals With Missense Variants in the Voltage-Sensing and Pore Domains of KCNH5.

KCNH5 encodes the voltage-gated potassium channel EAG2/Kv10.2. We aimed to delineate the neurodevelopmental and epilepsy phenotypic spectrum associated with de novo KCNH5 variants. We screened 893 individuals with developmental and epileptic encephalopathies for KCNH5 variants using targeted or exome sequencing. Additional individuals with KCNH5 variants were identified through an international collaboration. Clinical history, EEG, and imaging data were analyzed; seizure types and epilepsy syndromes were classified. We included 3 previously published individuals including additional phenotypic details. We report a cohort of 17 patients, including 9 with a recurrent de novo missense variant p.Arg327His, 4 with a recurrent missense variant p.Arg333His, and 4 additional novel missense variants. All variants were located in or near the functionally critical voltage-sensing or pore domains, absent in the general population, and classified as pathogenic or likely pathogenic using the American College of Medical Genetics and Genomics criteria. All individuals presented with epilepsy with a median seizure onset at 6 months. They had a wide range of seizure types, including focal and generalized seizures. Cognitive outcomes ranged from normal intellect to profound impairment. Individuals with the recurrent p.Arg333His variant had a self-limited drug-responsive focal or generalized epilepsy and normal intellect, whereas the recurrent p.Arg327His variant was associated with infantile-onset DEE. Two individuals with variants in the pore domain were more severely affected, with a neonatal-onset movement disorder, early-infantile DEE, profound disability, and childhood death. We describe a cohort of 17 individuals with pathogenic or likely pathogenic missense variants in the voltage-sensing and pore domains of Kv10.2, including 14 previously unreported individuals. We present evidence for a putative emerging genotype-phenotype correlation with a spectrum of epilepsy and cognitive outcomes. Overall, we expand the role of EAG proteins in human disease and establish KCNH5 as implicated in a spectrum of neurodevelopmental disorders and epilepsy.

Developing Disposable EEG Cap for Infant Recordings at the Neonatal Intensive Care Unit.

Long-term EEG monitoring in neonatal intensive care units (NICU) is challenged with finding solutions for setting up and maintaining a sufficient recording quality with limited technical experience. The current study evaluates different solutions for the skin–electrode interface and develops a disposable EEG cap for newborn infants. Several alternative materials for the skin–electrode interface were compared to the conventional gel and paste: conductive textiles (textured and woven), conductive Velcro, sponge, super absorbent hydrogel (SAH), and hydro fiber sheets (HF). The comparisons included the assessment of dehydration and recordings of signal quality (skin interphase impedance and powerline (50 Hz) noise) for selected materials. The test recordings were performed using snap electrodes integrated into a forearm sleeve or a forehead band along with skin–electrode interfaces to mimic an EEG cap with the aim of long-term biosignal recording on unprepared skin. In the hydration test, conductive textiles and Velcro performed poorly. While the SAH and HF remained sufficiently hydrated for over 24 h in an incubator-mimicking environment, the sponge material was dehydrated during the first 12 h. Additionally, the SAH was found to have a fragile structure and was electrically prone to artifacts after 12 h. In the electrical impedance and recording comparisons of muscle activity, the results for thick-layer HF were comparable to the conventional gel on unprepared skin. Moreover, the mechanical instability measured by 1–2 Hz and 1–20 Hz normalized relative power spectrum density was comparable with clinical EEG recordings using subdermal electrodes. The results together suggest that thick-layer HF at the skin–electrode interface is an effective candidate for a preparation-free, long-term recording, with many advantages, such as long-lasting recording quality, easy use, and compatibility with sensitive infant skin contact.

MRI based composite parameter of multiple tissue types for improved patient-level hemispheric and regional level lateralization in pediatric epilepsy

Although voxel-based morphometry (VBM) of gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) changes aid in epileptic seizure lateralization, type of T1 pulse sequence, preprocessing steps and tissue segmentation methods lead to variation in tissue classification. Here, we test the prediction accuracy of individual MRI based tissue types and a novel composite ratio parameter [(GM + WM)/CSF], sensitive to parenchymal changes and independent of tissue classification variations. Pediatric patients with partial seizures (both simple and complex), but normal and lesion-free MRI were considered (33 patients; unilateral EEG; 17 female / 16 male; age mean ± SD = 11.5 ± 5 years). MRI based seizure lateralization was performed for each patient and verified with EEG findings alone or in combination with seizure semiology. T1 weighted MRI from patients and normal control subjects was spatially transformed to the Talairach atlas and automatically segmented into GM, WM and CSF tissue types. 41 age matched normal controls (11 female / 30 male; age mean ± SD = 14.6 ± 3 years) served as the null distribution to test tissue type deviations across each epilepsy patient. When verified with the patient EEG prediction, WM, GM and CSF had a hemispheric match of 76%, 70% and 55% respectively, while the composite ratio [(GM + WM)/CSF)] showed the highest accuracy of 85%. When EEG findings and seizure semiology were combined, MRI predictions using the composite ratio improved further to 88%. To further localize the epileptic focus, regional level (frontal, temporal, parietal and occipital) MRI predictions were obtained. The composite ratio performed at 88–91% accuracy, revealing regional MRI changes, not predictable with EEG. The results show inconsistent changes in GM and WM in majority of the pediatric epilepsy patients and demonstrate the applicability of the composite ratio [(GM + WM)/CSF)] as a superior predictor, independent of tissue classification variations. Clinical EEG findings combined with seizure semiology, can overcome scalp EEG's limitations and lean towards the MRI lateralization in specific cases.

Satelight: self-attention-based model for epileptic spike detection from multi-electrode EEG

Objective. Because of the lack of highly skilled experts, automated technologies that support electroencephalogram (EEG)-based in epilepsy diagnosis are advancing. Deep convolutional neural network-based models have been used successfully for detecting epileptic spikes, one of the biomarkers, from EEG. However, a sizeable number of supervised EEG records are required for training. Approach. This study introduces the Satelight model, which uses the self-attention (SA) mechanism. The model was trained using a clinical EEG dataset labeled by five specialists, including 16 008 epileptic spikes and 15 478 artifacts from 50 children. The SA mechanism is expected to reduce the number of parameters and efficiently extract features from a small amount of EEG data. To validate the effectiveness, we compared various spike detection approaches with the clinical EEG data. Main results. The experimental results showed that the proposed method detected epileptic spikes more effectively than other models (accuracy = 0.876 and false positive rate = 0.133). Significance. The proposed model had only one-tenth the number of parameters as the other effective model, despite having such a high detection performance. Further exploration of the hidden parameters revealed that the model automatically attended to the EEG’s characteristic waveform locations of interest.

Psychomotor development and seizure features in idiopathic myoclonic epilepsy in infancy.

Myoclonic epilepsy in infancy (MEI) is a rare syndrome characterized by generalized myoclonic seizures (MS) that occur within the first 3 years of life. In the present study, the form of onset, and clinical and electroencephalogram (EEG) features were analyzed. A retrospective chart review was conducted for 16 MEI patients between March 2009 and July 2022 in Peking Union Medical College. The clinical and video EEG (VEEG) characteristics, treatment strategy, and follow-up information were analyzed. Four cases presented with afebrile generalized tonic-clonic seizures (GTCS) at the onset of MEI (GTCS at onset or atypical MEI), while 12 cases presented with MS at onset (MS at onset or typical MEI). The 24-hour VEEG revealed a generalized discharge of polyspike (or spike)-and-wave complexes that lasted for 1–3 seconds in the ictal phase. All patients were treated with valproic acid monotherapy, and none of the patients experienced seizure recurrence. Furthermore, all patients had normal psychomotor development at the end of the follow up period. Typical MEI (MS at onset) and atypical MEI (GTCS at onset) were described in the present study. These 2 groups differed in form of onset, but there were no significant differences in clinical or EEG features.

Automatic detection of abnormal EEG signals using multiscale features with ensemble learning.

Electroencephalogram (EEG) is an economical and convenient auxiliary test to aid in the diagnosis and analysis of brain-related neurological diseases. In recent years, machine learning has shown great potential in clinical EEG abnormality detection. However, existing methods usually fail to consider the issue of feature redundancy when extracting the relevant EEG features. In addition, the importance of utilizing the patient age information in EEG detection is ignored. In this paper, a new framework is proposed for distinguishing an unknown EEG recording as either normal or abnormal by identifying different types of EEG-derived significant features. In the proposed framework, different hierarchical salient features are extracted using a time-wise multi-scale aggregation strategy, based on a selected group of statistical characteristics calculated from the optimum discrete wavelet transform coefficients. We also fuse the age information with multi-scale features for further improving discrimination. The integrated features are classified using three ensemble learning classifiers, CatBoost, LightGBM, and random forest. Experimental results show that our method with CatBoost classifier can yield superior performance vis-a-vis competing techniques, which indicates the great promise of our methodology in EEG pathology detection.

Multimodal prediction of residual consciousness in the intensive care unit: the CONNECT-ME study.

Functional MRI (fMRI) and EEG may reveal residual consciousness in patients with disorders of consciousness (DoC), as reflected by a rapidly expanding literature on chronic DoC. However, acute DoC is rarely investigated, although identifying residual consciousness is key to clinical decision-making in the intensive care unit (ICU). Therefore, the objective of the prospective, observational, tertiary center cohort, diagnostic phase IIb study 'Consciousness in neurocritical care cohort study using EEG and fMRI' (CONNECT-ME, NCT02644265) was to assess the accuracy of fMRI and EEG to identify residual consciousness in acute DoC in the ICU. Between April 2016 and November 2020, 87 acute DoC-patients with traumatic or non-traumatic brain injury were examined with repeated clinical assessments, fMRI and EEG. Resting-state EEG and EEG with external stimulations were evaluated by visual analysis, spectral band analysis and a Support Vector Machine (SVM) consciousness classifier. In addition, within- and between-network resting-state connectivity for canonical resting-state fMRI networks were assessed. Next, we used EEG and fMRI data at study enrollment in two different machine-learning algorithms (Random Forest and SVM with a linear kernel), to distinguish patients in a minimally conscious state or better (≥MCS) from those in coma or unresponsive wakefulness state (≤UWS), at time of study enrollment and at ICU-discharge (or before death). Prediction performances were assessed with area under the curve (AUC). Of 87 DoC-patients (mean age, 50.0 ± 18 years, 43% women), 51 (59%) were ≤ UWS and 36 (41%) were ≥ MCS at study enrollment. Thirty-one (36%) patients died in the ICU, including 28 who had life-sustaining therapy withdrawn. EEG and fMRI predicted consciousness levels at study enrollment and ICU-discharge, with maximum AUCs of 0.79 (95% CI 0.77-0.80) and 0.71 (95% CI 0.77-0.80), respectively. Models based on combined EEG and fMRI features predicted consciousness levels at study enrollment and ICU-discharge with maximum AUCs of 0.78 (95% CI 0.71-0.86) and 0.83 (95% CI 0.75-0.89), respectively, with improved positive predictive value and sensitivity. Overall, both machine-learning algorithms (SVM and Random Forest) performed equally well. In conclusion, we suggest that acute DoC prediction models in the ICU be based on a combination of fMRI and EEG features, regardless of the machine-learning algorithm used.

Differentiation of subclinical and clinical electrographic events in long-term electroencephalographic recordings.

With the advent of ultra-long-term recordings for monitoring of epilepsies, the interpretation of results of isolated electroencephalographic (EEG) recordings covering only selected brain regions attracts considerable interest. In this context, the question arises of whether detected ictal EEG patterns correspond to clinically manifest seizures or rather to purely electrographic events, that is, subclinical events. EEG patterns from 268 clinical seizures and 252 subclinical electrographic events from 50 patients undergoing video-EEG monitoring were analyzed. Features extracted included predominant frequency band, duration, association with rhythmic muscle artifacts, spatial extent, and propagation patterns. Classification using logistic regression was performed based on data from the whole dataset of 10-20 system EEG recordings and from a subset of two temporal electrode contacts. Correct separation of clinically manifest and purely electrographic events based on 10-20 system EEG recordings was possible in up to 83.8% of events, depending on the combination of features included. Correct classification based on two-channel recordings was only slightly inferior, achieving 78.6% accuracy; 74.4% and 74.8%, respectively, of events could be correctly classified when using duration alone with either electrode set, although classification accuracies were lower for some subgroups of seizures, particularly focal aware seizures and epileptic arousals. A correct classification of subclinical versus clinical EEG events was possible in 74%-83% of events based on full EEG recordings, and in 74%-78% when considering only a subset of two electrodes, matching the channel number available from new implantable diagnostic devices. This is a promising outcome, suggesting that ultra-long-term low-channel EEG recordings may provide sufficient information for objective seizure diaries. Intraindividual optimization using high numbers of ictal events may further improve separation, provided that supervised learning with external validation is feasible.

TU-145. Short and long-term clinical and resting-state EEG results of Transcranial Pulse Stimulation in Alzheimer’s Disease

TU-145. Short and long-term clinical and resting-state EEG results of Transcranial Pulse Stimulation in Alzheimer’s Disease