Quantitative EEG Analysis Research Articles

The association of quantitative EEG and MRI in Creutzfeldt-Jakob Disease.

Previous studies showed concordance between the typical Periodic Sharp Wave Complex (PSWC) activity in EEG of Creutzfeldt-Jakob Disease (CJD) patients and the MRI findings, while the concordance with slow activity in EEG is less established. The aim of this study was to better characterize the association between MRI findings and EEG changes using quantitative EEG (qEEG) analysis. The demographics, clinical features, and the MRI findings of 12 familial E200K patients with CJD were gathered. EEG test was done and reviewed for the typical PSWC and for the non-specific slow activity. A possible association between the MRI findings and the EEG activity was examined. Then, EEG was analyzed using qEEG tool, and the association between the qEEG finding and the MRI was examined. Twelve patients were included in the study (67% women). Cortical MRI lesions finding were seen in 6/12 (50%) of the patients, and deep gray mater lesions were seen in 8/12 patients (67%). EEG showed the classic PSWC in 6/12 (50%) of the patients where slow activity was seen in 10/12 (83%). Slow activity and cortical MRI findings were associated in only 2/6 (33%) where deep gray matter findings and the slow activity had concordance of 4/8 (50%). qEEG analysis improved this concordance between slow activity and cortical findings to 3/6 (50%) and with the deep gray matter findings to 5/8 (63%). Quantitative EEG analysis modesty but not significantly, improves the association of EEG slow activity in familial E200K CJD patients with MRI findings.

Global and Intertuberal Epileptic Networks in Tuberous Sclerosis Based on Stereoelectroencephalographic (sEEG) Findings: A Quantitative EEG Analysis in Pediatric Subjects and Surgical Implications

Abstract INTRODUCTION Recent evidence favors a network concept in tuberous sclerosis complex (TSC) with seizure generation and propagation related to changes in global and regional connectivity between multiple, anatomically distant tubers. Direct exploration of network dynamics in TSC has been made possible through intracranial sampling with stereoelectroencephalography (sEEG). The objective of this study is to define epileptic networks in TSC using quantitative analysis of sEEG recordings. METHODS Intracranial sEEG recordings were obtained from pediatric patients who presented with medically refractory epilepsy secondary to TSC and subjected to quantitative signal analysis methods. Cortical connectivity was quantified by calculating pairwise coherence between all contacts and constructing an association matrix. The global coherence, defined as the ratio of the largest eigenvalue to the sum of all the eigenvalues, was calculated for each frequency band (delta, theta, alpha, beta, gamma). Spatial distribution of the connectivity was identified by plotting the leading principal component (product of the largest eigenvalue and its corresponding eigenvector). RESULTS Four pediatric subjects with TSC underwent invasive intracranial monitoring with sEEG, comprising 31 depth electrodes and 250 contacts, for localization of the epileptogenic focus and guidance of subsequent surgical intervention. Quantitative connectivity analysis revealed a sustained increase in global coherence during the ictal period in the beta/low gamma (14-30 Hz) and high gamma (31-80 Hz) bands. Our results corroborate findings from existing literature, which implicate higher frequencies as a driver of synchrony and desynchrony. Late delta activity was noted in the postictal period. CONCLUSION Epileptic networks, characterized by coordinated high frequency activity in the beta/low gamma and high gamma bands define the ictal period in TSC. This time-dependent increase in global coherence demonstrates evidence for intratuberal, intertuberal, and perituberal connectivity in TSC. These findings have surgical implications for targeted laser ablation or surgical resection of tubers.

Special symposium: The upcoming 90th anniversary in 2019 of Berger’s discovery of EEG-4. New developments in quantitative EEG analysis

Special symposium: The upcoming 90th anniversary in 2019 of Berger’s discovery of EEG-4. New developments in quantitative EEG analysis

Microsleep episodes in the borderland between wakefulness and sleep

The wake-sleep transition zone represents a poorly defined borderland, containing, for example, microsleep episodes (MSEs), which are of potential relevance for diagnosis and may have consequences while driving. Yet, the scoring guidelines of the American Academy of Sleep Medicine (AASM) completely neglect it. We aimed to explore the borderland between wakefulness and sleep by developing the Bern continuous and high-resolution wake-sleep (BERN) criteria for visual scoring, focusing on MSEs visible in the electroencephalography (EEG), as opposed to purely behavior- or performance-defined MSEs. Maintenance of Wakefulness Test (MWT) trials of 76 randomly selected patients were retrospectively scored according to both the AASM and the newly developed BERN scoring criteria. The visual scoring was compared with spectral analysis of the EEG. The quantitative EEG analysis enabled a reliable objectification of the visually scored MSEs. For less distinct episodes within the borderland, either ambiguous or no quantitative patterns were found. As expected, the latency to the first MSE was significantly shorter in comparison to the sleep latency, defined according to the AASM criteria. In certain cases, a large difference between the two latencies was observed and a substantial number of MSEs occurred between the first MSE and sleep. Series of MSEs were more frequent in patients with shorter sleep latencies, while isolated MSEs were more frequent in patients who did not reach sleep. The BERN criteria extend the AASM criteria and represent a valuable tool for in-depth analysis of the wake-sleep transition zone, particularly important in the MWT.

A New Method of Diagnosing Attention-Deficit Hyperactivity Disorder in Male Patients by Quantitative EEG Analysis.

Attention-deficit hyperactivity disorder (ADHD) is one of the most common neuropsychiatric disorders of childhood. Diagnosis of ADHD is based on core symptoms and checklists. However, these are both subjective, which can lead to the problems of overdiagnosis and underdiagnosis. Elevated theta/beta ratio (TBR) of EEG band has been approved by the US Food and Drug Administration as a tool to assist in the diagnosis of ADHD. However, several recent studies have demonstrated that there are no significant differences in TBR between people with and without ADHD. In this study, we attempted to develop a new method for differentiating between male with and without ADHD by analyzing EEG features. Thirty boys with ADHD combined type (aged 8 years 5 months ± 1 year 11 months) and 30 age-matched controls (aged 8 years 5 months ± 1 year 8 months) were enrolled in this study. A classification analysis-based approach comprising training and classification phases was developed for classifying each subject's EEG features as ADHD or non-ADHD. Eight crucial feature descriptors were selected and ranked based on the t test. Compared with TBR in our study, the developed method had a higher area under the curve (87.78%), sensitivity (80.0%), and specificity (80.0%). Our method is more precise than using TBR in the diagnosis of ADHD. This newly developed method is a useful tool in identifying patients with ADHD and might reduce the possibility of overdiagnosis and underdiagnosis.

Changes in mu and beta amplitude of the EEG during upper limb movement correlate with motor impairment and structural damage in subacute stroke

ObjectiveMu and beta EEG oscillations show typical desynchronization patterns during movement. The aim of the current study was to assess whether in sub-acute stroke patients the magnitude of movement-related desynchronization reflects the extent of residual motor ability in the paretic upper limb. MethodsEEG and EMG data were recorded from 14 first-event stroke patients during repeated wrist extension movements of the paretic upper limb. Residual motor ability was assessed by the Fugl Meyer and Box and Blocks standardized clinical tests. Normalized lesion data was analyzed using the MEDx software. ResultsThe magnitude of event-related de-synchronization (ERD) of the high-mu and low-beta bands of the EEG, measured over the affected hemisphere, correlated significantly with (a) residual motor function in the paretic upper limb as measured by standard clinical tests; (b) the magnitude of EMG recorded from the paretic upper limb during wrist extension; and (c) the total hemispheric volume loss (negative correlation). ConclusionThe magnitude of high-mu and low-beta ERD recorded from the lesioned hemisphere of subacute stroke patients correlates with residual motor ability in the paretic upper limb. SignificanceMeasures derived from quantitative EEG analysis may play an important role in neurorehabilitation clinical practice.

Quantitative EEG analysis in Encephalopathy related to Status Epilepticus during slow Sleep.

Since its first description, quantifying the burden of epileptiform abnormalities in sleep EEG has played a fundamental role in the diagnosis of Encephalopathy related to Status Epilepticus during slow Sleep (ESES). In fact, in the 1971 seminal paper by Tassinari's group and in the following studies on this syndrome, the amount of epileptiform discharges (EDs) was calculated as the percentage of slow sleep occupied by spike-and-waves and referred to as "spike and wave index" (SWI). However, nowadays it is becoming increasingly clear that the SWI alone does not explain the whole clinical course of patients affected by ESES. In this paper, we aim to provide a state-of-the-art summary of the quantitative EEG methods currently used in the ESES/CSWS literature, highlighting the possible pitfalls and discrepancies explaining the unsatisfactory correlation between SWI and clinical course. Furthermore; we illustrate a number of methodological refinements - taking into account inter-individual, intra-individual, and temporal variability of EDs - alongside "new" quantitative variables -including ED-related and sleep-related features - potentially useful to reach a reliable electro-clinical correlation in patients with ESES.

P14-S Latency of Mismatch Negativity as a predictor of outcome in critically patients with subarachnoid haemorrhage

Background Multimodal neurophysiological markers support clinical assessments in individuals with disorders of consciousness (DoC). Few studies focused on subarachnoid haemorrhage (SAH) related coma. Furthermore, these patients often need sedation during the acute phase, making behavioral and neurophysiological monitoring challenging. We investigated which neurophysiological marker could better predict outcome in acute DoC related to SAH. Methods We studied 14 SAH patients admitted in the Sainte Anne hospital neurological ICU, between Oct 2017 and June 2018. All patients underwent within the first 72 h: clinical evaluation, EEG and multimodal evoked potentials (somatosensory potentials, brainstem auditory evoked potentials (BAEPs), late latency event-related potentials, N100 and mismatch negativity (MMN)). Quantitative EEG analysis using spectral and connectivity analysis was performed. According to the Glascow Outcome scale (GOS), we defined two clinical outcome groups: favorable (F) and unfavorable (UF). Results Clinical examination, EEG visual and quantitative analysis, presence of N20, BAEP, N100 and MMN were not significantly different in the two groups. However, mean latencies of N100 (F = 80 ms; UF = 106 ms;) and of MMN (F = 139 ms; UF = 197msec) were significantly different in the two groups (Mann Whitney test, respectively p = 0.017 and p = 0.035). Conclusions This study suggests that during the acute phase, in sedated patients with DoC related to SAH, latencies of N100 and latencies of MMN could be more discriminatory than the “classical” interpretation of these markers (present/absent) and the EEG analysis, in term of prognosis. These results need to be confirmed in a larger study.

Quantitative Preterm EEG Analysis: The Need for Caution in Using Modern Data Science Techniques.

Hemodynamic changes during neonatal transition increase the vulnerability of the preterm brain to injury. Real-time monitoring of brain function during this period would help identify the immediate impact of these changes on the brain. Neonatal EEG provides detailed real-time information about newborn brain function but can be difficult to interpret for non-experts; preterm neonatal EEG poses even greater challenges. An objective quantitative measure of preterm brain health would be invaluable during neonatal transition to help guide supportive care and ultimately protect the brain. Appropriate quantitative measures of preterm EEG must be calculated and care needs to be taken when applying the many techniques available for this task in the era of modern data science. This review provides valuable information about the factors that influence quantitative EEG analysis and describes the common pitfalls. Careful feature selection is required and attention must be paid to behavioral state given the variations encountered in newborn EEG during different states. Finally, the detrimental influence of artifacts on quantitative EEG analysis is illustrated.

Effects of Diurnal Intermittent Fasting on Daytime Sleepiness Reflected by EEG Absolute Power.

We hypothesized that diurnal intermittent fasting would not increase slow frequency EEG waves, which reflect increased sleepiness. Therefore, this study aimed at performing a quantitative EEG analysis of young healthy volunteers with and without diurnal intermittent fasting while controlling for potential confounders. Eight young healthy volunteers (mean age: 24.4 ± 3.7 years) reported to the laboratory during the following five occasions: (1) adaptation, (2) 4 weeks before Ramadan while performing diurnal intermittent fasting (fasting from dawn to sunset) for 1 week (fasting outside Ramadan), (3) 1 week before Ramadan (nonfasting baseline), (4) during the second week of Ramadan while fasting (Ramadan), and (5) 2 weeks after Ramadan (recovery). Frontal, central, and occipital EEG signals were monitored four times (09:15, 11:15, 13:15, and 15:15 for 15 minutes each time) while the participants were awake to assess EEG spectral power of the delta, theta, alpha, and beta frequency bands. During wakefulness, no difference in EEG slowing was observed between nonfasting baseline and fasting during and outside Ramadan. There were no differences in the mean absolute power of each band in any studied brain region in any of the four study periods. Moreover, there was no difference in EEG absolute powers in the delta, theta, alpha, and beta frequency bands for any individual brain region during the four study periods. Spectral analysis of EEG power indicates that diurnal intermittent fasting does not increase sleepiness. Future larger studies are needed to confirm the current findings.

Electrocortical networks in Parkinson's disease patients with Mild Cognitive Impairment. The PaCoS study

IntroductionParkinson's Disease (PD) is frequently associated with cognitive dysfunction ranging from Mild Cognitive Impairment (PD-MCI) to dementia. Few electrophysiological studies are available evaluating potential pathogenetic mechanisms linked to cognitive impairment in PD since its initial phases. The objective of the study is to analyze electrocortical networks related with cognitive decline in PD-MCI for identifying possible early electrophysiological markers of cognitive impairment in PD. MethodsFrom the PaCoS (Parkinson's disease Cognitive impairment Study) cohort, a sample of 102 subjects including 46 PD-MCI and 56 PD with normal cognition (PD-NC) was selected based on the presence of a neuropsychological assessment and at least one EEG recording. EEG signal epochs were analysed using Independent Component Analysis LORETA and spectral analysis by computing the Power Spectral Density (PSD) of site-specific signal epochs. ResultsLORETA analysis revealed significant differences in PD-MCI patients compared to PD-NC, with a decreased network involving alpha activity over the occipital lobe, an increased network involving beta activity over the frontal lobe associated with a reduction over the parietal lobe, an increased network involving theta and delta activity over the frontal lobe and a reduction of networks involving theta and delta activity in the parietal lobe. Quantitative EEG analysis showed a significant decrease of alpha PSD over the occipital regions and an increase of delta PSD over the left temporal region in PD-MCI as compared to PD-NC. ConclusionElectrocortical abnormalities detected in PD-MCI patients may represent the instrumental counterpart of early cognitive decline in PD.

Acute effect of vaporized Cannabis on sleep and electrocortical activity

The use of Cannabis for medical purposes is rapidly expanding and is usually employed as a self-medication for the treatment of insomnia disorder. However, the effect on sleep seems to depend on multiple factors such as composition of the Cannabis, dosage and route of administration. Vaporization is the recommended route for the administration of Cannabis for medical purposes; however, there is no published research about the effects of vaporized Cannabis on sleep, neither in laboratory animals, nor in humans. Because previous reports suggested that low doses of THC have sedating effects, the aim of the present study was to characterize in rats, the acute effects on sleep induced by the administration of low doses of THC by means of vaporization of a specific type of Cannabis (THC 11.5% and negligible amounts of other cannabinoids).For this purpose, polysomnographic recordings in chronically prepared rats were performed during 6 h in the light and dark phases. Animals were treated with 0 (control), 40, 80 and 200 mg of Cannabis immediately before the beginning of recordings; the THC plasma concentrations with these doses were low (up to 6.7 ng/mL with 200 mg). A quantitative EEG analyses by means of the spectral power and coherence estimations was also performed for the highest Cannabis dose.Compared to control, 200 mg of Cannabis increased NREM sleep time during the light phase, but only during the first hour of recording. Interestingly, no changes on sleep were observed during the dark (active) phase or with lower doses of Cannabis.Cannabis 200 mg also produced EEG power reductions in different cortices, mainly for high frequency bands during W and REM sleep, but only during the light phase. On the contrary, a reduction in the sleep spindles intra-hemispheric coherence was observed during NREM sleep, but only during the dark phase.In conclusion, administration of low doses of THC by vaporization of a specific type of Cannabis produced a small increment of NREM sleep, but only during the light (resting) phase. This was accompanied by subtle modifications of high frequency bands power (during the light phase) and spindle coherence (during the dark phase), which are associated with cognitive processing. Our results reassure the importance of exploring the sleep-promoting properties of Cannabis.

Abstract TP542: EEG Detection of Secondary Worsening of Neuronal Function in Malignant MCA Infarction: Development of a Tool to Identify Patients for Decompressive Hemicraniectomy

Hemicraniectomy (HC) improves mortality in Malignant Middle cerebral artery Infarctions (MMI), but patient selection and optimal timing for HC remains unclear. A continuous method to monitor for the risk and development of herniation from infarction would allow a better selection of patients for HC We hypothesized that certain EEG features reflective of worsening mass effect will be easier to detect on quantitative EEG (qEEG) analysis trends We studied qEEG trends before, during clinical deterioration (drop in GCS by at least 3 points) and after administration of osmotherapy. qEEG trends included rhythmicity spectra 0-25 Hz (scale 0- 2uV/Hz), Color Density Power Spectra (CSA) 0-20Hz (scale 0-2uV/Hz), amplitude integrated EEG (aEEG): 0-100uV and VsRhythmicity spectrogram 1 - 25 Hz (-3-3 Zscore) Trends demonstrated bilateral loss of power on beta and alpha frequencies and increase in theta preceded clinical deterioration by approximately 1 hour (left > right). Patient2 subsequently had a robust theta power increase (+z score, red) in left > right hemisphere which coincide with clinical deterioration on vsRhythmmicity Spectrogram. Also, there was a decreased of fast and slow frequencies bands on rhythmicity spectrogram and CSA with a diffuse attenuation outlined by aEEG (Left > right). Following osmotherapy gradual improvement in all frequencies is seen but with an incomplete recovery in patient1 with lower alpha and theta power on right posterior and anterior quadrants suggested ongoing uncal and sub-falcine herniation, later confirmed on imaging Changes on qEEG reflecting impaired perfusion during acute herniation can manifest as decreased power in all frequency bands of varying severity prior to clinical deterioration. These changes can be completely or partially reversed with osmotherapy. Partial EEG response to osmotherapy may represent a potential sign of patients who will fail medical management and need DC. qEEG may be a helpful tool in HC patient selection

The effect of cranial electrotherapy stimulation on sleep in healthy women

Objective: Cranial electrotherapy stimulation (CES) is considered to be a potential treatment for insomnia. Women are more likely to suffer from insomnia than men. Therefore we studied the effect of CES on sleep efficiency in young healthy women. Methods: A randomized, controlled clinical study was conducted on 40 women (age 18–35 years) without sleep disorders. Each subject underwent two nights of polysomnography in a sleep center. During the second night, we applied CES with a commercial device (Alpha-Stim 100) using either active or sham stimulation (double-blinded). Sleep was evaluated with respect to differences between the active and sham modes. Sleep electroencephalogram (EEG) analysis was applied to determine frequency changes. Results: In our study we found no evidence of any direct influence of the Alpha-Stim 100 on sleep. After application of CES, we determined no significant differences between the active group and the control group (sham). Using EEG spectral analysis there was evidence of a frequency-lowering influence on the low-α frequency band (8–10 Hz). Conclusions: At most we may assume a reproducible effect on the α frequency measured in the EEG for application of CES with current levels >100 µA and presumably also with frequencies >0.5 Hz, with application directly at the cranium. We found no influence on sleep parameters. The effect on the low-α band evidenced in quantitative EEG analysis would require further investigation in a study with sufficient effect size. Significance: This is the first study to investigate the effects of CES with polysomnography during and after therapy. Our study contributes to the few controlled trials that have been conducted to study CES and its effects on the EEG α band.Highlights• This is the first study to investigate the effects of the Alpha-Stim in polysomnography during and subsequent to therapy.• In a quantitative electroencephalogram analysis we studied the data obtained for systematic changes.• To minimize placebo effects in patients with sleep disorders, we conducted the tests on subjects without such disorders.

Viral encephalitis in quantitative EEG.

Our aim was to study the character of quantitative EEG of viral encephalitis. Collect the EEG data of hospitalized children with viral encephalitis diagnosed by pediatricians from January 2013 to September 2016, and EEG data of normal cases at the same age were control group. Using quantitative EEG analysis technologies to obtain power spectrum value, power spectrum value, 1 power spectrum value, 12 power spectrum value, 21 power spectrum value, and 12 power spectrum value. Relative power spectrum values of 2 and were obtained by calculation. All the cases were divided into 5 groups according to the EEG character: age 3 group, age 4 group, age 5-6 group, age 7-9-year group, and age 10-14-year group. Viral encephalitis group and normal cases group were statistically compared to obtain characters of quantitative EEG with viral encephalitis. Power spectrum values and power spectrum values of 3-14-year-old cases with encephalitis increased. 1 power spectrum values existed in age 3 group with viral encephalitis and declined at the post-head lead, while 11 and 12 power spectrum values existed in age 4-14 group with viral encephalitis declined at the post-head lead. The value of 2 Power spectrum values of age 3-9 group was limited in diagnosing viral encephalitis. 1 and 12 power spectrum values decrease in age 10-14 group with viral encephalitis. Relative power spectrum values of 2 and increased in age 3-14 group with viral encephalitis; Relative power spectrum values of decreased in age 3-14 group with viral encephalitis, most lead of relative power spectrum decreased in age 3-14 group with viral encephalitis. The character of quantitative EEG of cases with viral encephalitis is similar to EEG, but more detailed, more precise, more intuitive and can be used for clinical diagnose of viral encephalitis.

Quantitative EEG Tomography of Early Childhood Malnutrition.

The goal of this study is to identify the quantitative electroencephalographic (qEEG) signature of early childhood malnutrition [protein-energy malnutrition (PEM)]. To this end, archival digital EEG recordings of 108 participants in the Barbados Nutrition Study (BNS) were recovered and cleaned of artifacts (46 children who suffered an episode of PEM limited to the first year of life) and 62 healthy controls). The participants of the still ongoing BNS were initially enrolled in 1973, and EEGs for both groups were recorded in 1977–1978 (at 5–11 years). Scalp and source EEG Z-spectra (to correct for age effects) were obtained by comparison with the normative Cuban Human Brain Mapping database. Differences between both groups in the z spectra (for all electrode locations and frequency bins) were assessed by t-tests with thresholds corrected for multiple comparisons by permutation tests. Four clusters of differences were found: (a) increased theta activity (3.91–5.86 Hz) in electrodes T4, O2, Pz and in the sources of the supplementary motor area (SMA); b) decreased alpha1 (8.59–8.98 Hz) in Fronto-central electrodes and sources of widespread bilateral prefrontal are; (c) increased alpha2 (11.33–12.50 Hz) in Temporo-parietal electrodes as well as in sources in Central-parietal areas of the right hemisphere; and (d) increased beta (13.67–18.36 Hz), in T4, T5 and P4 electrodes and decreased in the sources of bilateral occipital-temporal areas. Multivariate Item Response Theory of EEGs scored visually by experts revealed a neurophysiological latent variable which indicated excessive paroxysmal and focal abnormality activity in the PEM group. A robust biomarker construction procedure based on elastic-net regressions and 1000-cross-validations was used to: (i) select stable variables and (ii) calculate the area under ROC curves (AUC). Thus, qEEG differentiate between the two nutrition groups (PEM vs Control) performing as well as visual inspection of the EEG scored by experts (AUC = 0.83). Since PEM is a global public health problem with lifelong neurodevelopmental consequences, our finding of consistent differences between PEM and controls, both in qualitative and quantitative EEG analysis, suggest that this technology may be a source of scalable and affordable biomarkers for assessing the long-term brain impact of early PEM.

Quantitative EEG analysis in early neonatal hypoxic ischaemic encephalopathy (HIE) to grade the severity and predict long-term prognosis

Summary Objectives The selection of quantitative EEG markers according to their ability to grade the severity of cerebral disability after perinatal hypoxia and to predict long term prognosis. These markers will be associated in an automated analysis tool that can be used at the patient's bedside to facilitate visual analysis. Methods Full term neonatal EEG were recorded before the 6th hour of life after perinatal hypoxia. They were visually analysed and graded from 0 to 3 (French classification) by two neurophysiologists blinded to clinical data. Several quantitative EEG markers were calculated on different derivations: the minimal and maximal Amplitude Index (min and max AI), the Burst Suppression Rate (BSR), the Spectral Absolute Power and Relative Delta Power (AP and RDP), the Spectral Edge Frequency (SEF). Results Ninety EEG were selected between 2013 and 2017 for the quantitative analysis. The mean BSR on C3C4 discriminated the grades 3 from grades 1 and 2 (Youden index: 0.88) and the median min AI on C3T3 discriminated the grades 1 from grades 2 (Youden index 0.6). The mean BSR on C3C4 discriminated good outcome from poor outcome at 1 year (Air Under the Curve: AUC 0.89) and 2 years (AUC 0.87). Conclusion After perinatal hypoxia, quantitative EEG markers based on the analysis of the amplitude and the continuity of the signal are associated with the visual EEG grades of severity and with the evolution at 1 and 2 years. They can be used to develop an automated classifier to facilitate the visual interpretation, especially to evaluate the indication of therapeutic hypothermia (grades 2 and 3) and long-term prognosis.

Automated detection of electroencephalography artifacts in human, rodent and canine subjects using machine learning

BackgroundElectroencephalography (EEG) invariably contains extra-cranial artifacts that are commonly dealt with based on qualitative and subjective criteria. Failure to account for EEG artifacts compromises data interpretation. New MethodWe have developed a quantitative and automated support vector machine (SVM)-based algorithm to accurately classify artifactual EEG epochs in awake rodent, canine and humans subjects. An embodiment of this method also enables the determination of ‘eyes open/closed’ states in human subjects. ResultsThe levels of SVM accuracy for artifact classification in humans, Sprague Dawley rats and beagle dogs were 94.17%, 83.68%, and 85.37%, respectively, whereas 'eyes open/closed' states in humans were labeled with 88.60% accuracy. Each of these results was significantly higher than chance. Comparison with Existing MethodsOther existing methods, like those dependent on Independent Component Analysis, have not been tested in non-human subjects, and require full EEG montages, instead of only single channels, as this method does. ConclusionsWe conclude that our EEG artifact detection algorithm provides a valid and practical solution to a common problem in the quantitative analysis and assessment of EEG in pre-clinical research settings across evolutionary spectra.

S82. Paroxysmal sympathetic hyperactivity evident as cyclical pattern on quantitative EEG analysis

Introduction Episodic fluctuations in sympathetic activity following traumatic brain injury (TBI) have been described for decades, with paroxysmal sympathetic hyperactivity (PSH) being the currently accepted term for this entity. While early reports postulated an epileptic mechanism, our understanding of the pathophysiology of PSH has evolved. Current hypotheses center around deregulation of central descending inhibitory inputs from the insula and cingulate cortex, resulting in increased sympathetic output from the hypothalamus, diencephalon and brainstem. We present a case of PSH with evidence of disrupted sympathetic arousal networks. Methods A 41-year-old woman suffered a TBI with multifocal supratentorial hemorrhages and intraparenchymal contusions. Eight days post-MVA she developed episodes of tachycardia, hypertension, diaphoresis, back-arching and extensor posturing of the upper extremities upon stimulus. These events were assumed to be epileptic and treated with lorazepam (up to 16 mg/24 h) and phenobarbital (up to 345 mg/24 h) to no avail. Episodes improved with dexmedetomidine and clonidine. Continuous video-EEG was ordered for event characterization. Results EEG background was characterized by reactive generalized theta activity without interictal discharges. Multiple stimulus-induced stereotypical events were captured on the first EEG, prior to initiation of anti-epileptic drugs and dexmedetomidine. These were time-locked with static generalized rhythmic delta activity, often sharply contoured (GRDA + S). Quantitative EEG cycled between GRDA+S and polymorphic theta; transitions were often associated with stimulation or arousal. Thus, the events were shown to be stimulus/arousal-induced PSH time-locked with GRDA+S. The second EEG on lorazepam (2 mg q6h) and dexmedetomidine (26 mcg/kg/h) demonstrated less GRDA and no PSH. The third EEG on phenobarbital (130 mg qhs) and less dexmedetomidine (1.5 mcg/kg/h) demonstrated polymorphic delta slowing during episodes of PSH. Therefore, PSH was seen prior to initiation of dexmedetomidine and while on a low dose, but not while on a higher dose. Phenobarbital did not improve PSH. Conclusion Episodes of PSH were non-epileptic and concordant with cyclical stimulus or arousal-induced GRDA. This pattern is thought to represent cortical-subcortical dysfunction rather than hyperexcitability, as shown in a recent retrospective analysis of critical care continuous EEG in which no association was found between GRDA and seizures. The cyclical pattern on scalp and quantitative EEG reflected increased generalized rhythmicity with arousal and episodes of PSH; akin to stimulus/arousal-induced rhythmic or periodic or ictal discharges (SIRPIDs) seen in critically ill patients. The pattern seen on the cEEG reflects network disruption during PSH rather than a hyperirritable cortex; we believe that treatment of PSH should target sympathetic arousal networks rather than cortical hyperexcitability.

S72. Case of acute stroke after high-risk carotid surgery: Visual and quantitative EEG analysis

EEG monitoring in the intensive care unit (ICU) has multiple potential uses outside of seizure detection by providing a continuous noninvasive assessment of neurologic function. Since the early 1990s, EEG has been widely employed intraoperatively during carotid endarterectomy to evaluate for cerebral ischemia. Its utility in the ICU, however, is contingent on access to frequent review by an expert neurophysiologist. Several studies found that quantitative EEG (QEEG), particularly a decrease in the alpha/delta ratio (ADR), can be sensitive in screening for delayed cerebral ischemia after subarachnoid hemorrhage. QEEG monitoring may be useful in other situations during which monitoring for ischemia is critical. The EEG for this case was reviewed by three neurophysiologists. QEEG analysis was performed retrospectively with Persyst InsightII software. A 56-year-old right-handed man presented with acute expressive aphasia. Imaging revealed a left frontal stroke and >70% left internal carotid artery stenosis. Six weeks later, he underwent left carotid endarterectomy, complicated by a diseased vessel wall necessitating complete ligation of the artery. Post-operative neurologic exam was non-focal. He was sedated and placed on normothermia protocol in the ICU to prevent cerebral infarction. Visual EEG review initially showed gradually increasing intermittent left temporal slowing. Approximately 90 min later, acute persistent left hemispheric attenuation was seen, most prominent over the centro-posterior region, followed by a gradual increase in left hemispheric delta slowing. Subsequent neurologic exam off sedation revealed a new right hemiparesis; brain MRI showed an acute left middle cerebral artery infarct. QEEG showed immediate prominent and persistent decreases in delta, theta, and alpha power, as well as large reductions in the amplitude-integrated EEG (aEEG) over the left hemisphere. There was also a slight decline in beta power over the left. A delayed decrease in the ADR over the left side was seen approximately 20 min later. Visual EEG inspection in this case revealed acute lasting focal left hemispheric attenuation, with subsequent increased slowing after the time of presumed infarct. Interestingly, QEEG analysis showed simultaneous decreases in absolute power values and aEEG prior to change in the ADR. An automatic alarm system monitoring multiple QEEG parameters could therefore be helpful in monitoring for cerebral ischemia in the ICU, particularly when a trained neurophysiologist is not available.