Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Feb 1.
Published in final edited form as: J Clin Neurophysiol. 2021 Jun 17;40(2):136–143. doi: 10.1097/WNP.0000000000000865

Theta-alpha variability on admission EEG is associated with outcome in pediatric cerebral malaria

Alexander Andrews 1, Tesfaye Zelleke 2, Dana Harrar 3, Rima Izem 4,5,6, Jiaxiang Gai 7, Douglas Postels 8,9
PMCID: PMC8626528  NIHMSID: NIHMS1690886  PMID: 34669356

Abstract

Introduction

Pediatric cerebral malaria has high rates of mortality and neurological morbidity. Although several biomarkers, including EEG, are associated with survival or morbidity, many are resource intensive or require skilled interpretation for clinical use. Automation of quantitative interpretation of EEG may be preferable in resource-limited settings, where trained interpreters are rare. As currently used quantitative EEG factors do not adequately describe the spectrum of variability seen in studies from children with cerebral malaria, we developed and validated a new quantitative EEG variable, theta-alpha variability (TAV).

Methods

We developed TAV, a new quantitative variable, as a composite of multiple automated EEG outputs. We analyzed EEG records from 194 children (6 months to 14 years old) with cerebral malaria. Independent EEG interpreters performed standard quantitative and qualitative analyses, with the addition of the newly created variable. We assessed the associations of TAV with other quantitative EEG factors, a qualitative assessment of variability, and outcomes.

Results

TAV was not highly correlated with alpha, theta, or delta power, and was not associated with qualitative measures of variability. Children whose EEGs had higher values of TAV had a lower risk of death (OR= 0.934, 95% CI=0.902–0.966) or neurological sequelae (OR=0.960, 95% CI=0.932–0.990) compared to those with lower values. ROC analysis in predicting death at a TAV threshold of 0.244 yielded a sensitivity of 74% and specificity of 70% for an AUC of 0.755.

Conclusion

Theta-alpha variability is independently associated with outcome in pediatric cerebral malaria and can predict death with high sensitivity and specificity. Automated determination of this newly created EEG factor holds promise as a potential method to increase the clinical utility of EEG in resource-limited settings by allowing interventions to be targeted to those at higher risk of death or disability.

INTRODUCTION

Malaria continues to be the most important parasitic disease of humankind. Despite global advances in treatment and prevention, there are 228 million cases annually leading to an estimated 405,000 deaths.1 Cerebral malaria (CM), defined as an otherwise unexplained coma in someone with malaria parasitemia, is one of the most lethal forms of the disease. CM predominantly affects children under the age of 5 years.2 With a case-fatality rate of 17–25%, it is a leading cause of death in children in sub-Saharan Africa.3 One-third of survivors experience neurocognitive sequelae including any combination of epilepsy, motor deficits, cognitive deficits, and behavioral dysregulation.4, 5 Clearly, the disease has a profound public health impact in endemic areas.

In children with CM, multiple admission clinical factors and biomarkers are associated with survival, neurocognitive sequelae in survivors, or both. Modalities to assess these biomarkers include magnetic resonance imaging (MRI), electroencephalogram (EEG), and transcranial Doppler (TCD).6–8 These biomarkers have provided some insight into disease pathophysiology in children with CM, but critical gaps remain in understanding the key steps between malaria infection and adverse outcomes, e.g., death or neurological disability. MRI, EEG, or TCD can prospectively identify children at higher risk of an adverse outcome, allowing enrichment of sample populations in clinical trials of new CM interventions. Biomarkers may also be used to identify survivors at higher risk of neurocognitive sequelae, allowing targeting of resources in limited supply.

Brain MRI is the most widely studied of the predictive biomarkers in pediatric CM.8–11 Children with highly increased brain volume (decreased sulci and cisterns, blurring of the gray-white junction, without or with uncal herniation), have a 14 times higher odds of death (95% CI= 4.5–53.4) compared to children who lack these findings.8 Compared to brain MRI, the usefulness of EEG as a biomarker to predict outcome in CM is less established. A single previously published study of EEG in CM showed that multiple qualitative EEG factors are independently associated with death (slower background frequencies, lower average voltages, focal slowing, and lack of reactivity) and long-term neurologic sequelae (presence of electrographic seizures).6

Qualitative EEG interpretation requires the expertise of an electroencephalographer, whose specialized skills are in short supply in many malaria-endemic areas. If automated methods of EEG interpretation using EEG markers associated with outcome were developed and validated, the clinical application of EEG in the care of children with CM may be broadened. To date, quantitative EEG interpretation methods have not been widely utilized in malaria endemic areas.

As clinician-neurologists providing EEG interpretation to a CM treatment research site in Africa, we noted both the complexity of EEG waveforms (with mixtures of several sub-frequencies) and a broad range of spontaneous variability. In qualitative interpretation, these studies were categorized as either “variable” or “not variable”, a dichotomy widely employed in clinical practice, but one that may be insensitive to important differences in pathophysiologic state. During quantitative analysis of these EEG records, we found no standard, validated measure of variability in currently available interpretation software. Published medical literature did not provide guidance to a measure of variability assessing multiple sub-frequencies (delta, theta, alpha) in quantitative EEG interpretation.12–16 Analysis of variability of a single frequency band in EEG does not take into account the complexity of waveforms seen in comatose children with CM. In response and in order to better describe the range of variability we observed in EEG studies from children with CM, we created and validated a quantitative measure of EEG variability, termed “theta-alpha variability” (TAV), to represent the continuum of observations in affected children.

We aimed to investigate whether TAV is independently associated with outcome in pediatric CM. Quantifying the fast frequency component of single frequency bands in the EEG is potentially useful in predicting outcome in other critical illnesses. Variability in the 8–12 Hz alpha rhythm is associated with better outcome in severe traumatic brain injury.17 Lower variability of relative alpha power on an expanded spectrum (6–14Hz) may differentiate comatose from non-comatose patients in clinical conditions of sepsis.18 In patients with subarachnoid hemorrhage, relative alpha variability decreases with vasospasm and improves with its resolution.19 Finally, regional decline in the alpha-delta ratio is associated with early regional ischemia in aneurysmal subarachnoid hemorrhage.20 These methods have not been evaluated in children, nor in CM.

METHODS

Patient Demographics

We analyzed the EEGs of 194 children between 6 months and 14 years old with a clinical diagnosis of CM admitted to Queen Elizabeth Central Hospital (Blantyre, Malawi) between January 2012 and June 2017. Participants were enrolled in a long-standing study of CM pathogenesis. The parent study was approved by Institutional Review Boards of the University of Malawi College of Medicine (Malawi) and Michigan State University (USA).

Standard diagnostic criteria for the clinical diagnosis of CM included: coma (Blantyre coma score (a standard measure of the depth of coma used across Africa) ≤ 2); asexual forms of malaria parasites on peripheral blood smear; and no other explanation for mental status (e.g., hypoglycemia, meningitis, trauma, or post-ictal state).8 EEG was performed within 4 hours of admission, after clinical stabilization and the start of intravenous antimalarials.

Clinical seizures witnessed following hospital admission were generally treated with benzodiazepine monotherapy if limited to less than 5 minutes duration. Children with longer duration or recurrent seizures additionally received a loading dose of intravenous phenobarbital 18 mg/kg followed by maintenance therapy. Intravenous phenytoin was administered in patients whose seizures did not respond to phenobarbital. All patients received intense supportive therapy and current standard of care therapies for children with CM. At hospital discharge, children who survived were classified as being without or with neurological sequelae based on neurological and physical examinations and caretaker interview.

EEG Acquisition and Processing

Electrodes were applied by trained technicians using the standard 10–20 system and recorded on a CeeGraph digital machine (Biologic/Natus, Pleasanton, California) through 2016 and on an XLTEC 32 channel machine (Natus Medical Corporation, Pleasanton, California) in 2017, both with a sampling frequency of 200 Hz. Recordings were at least 30 minutes in duration. Interpretations were performed after hospital discharge by trained electrophysiologists (AA or TZ) using Persyst 13 (Persyst Corporation, Prescott, Arizona). EEG interpreters were blinded to participant outcome.

Under conditions of artifact reduction, EEG waveforms were subject to Fast Fourier Transformation to yield the following values: absolute power (μV2) of frequency bands of interest, including delta (0.5–3.9Hz), theta (4.0–7.9Hz), and alpha (8.0–12.9Hz); total power of all frequency bands; relative spectral power of frequency bands of interest (a quotient of the power of the frequency band of interest with total power as divisor); and power ratios of frequency bands of interest (a quotient of the power of the frequency band of interest with delta power as divisor). Measured power ratios included the low-delta to delta ratio (0.5–1.9Hz/05.−3.9Hz), theta-plus-alpha to delta ratio (4.0–12.9Hz/0.5–3.9Hz), and alpha to delta ratio (8.0–12.9Hz/0.5–3.9Hz). Samples were derived using a 2-second time window in the Persyst trend analysis panel (Figure 1). The raw EEG in these segments was inspected and confirmed to be devoid of artifact.

Figure 1.

Figure 1

Open in a new tab

Quantitative EEG panel with variables of interest

Quantitative EEG panel of a child with cerebral malaria in Persyst 13. Panel depicts variables of interest: total power (uV2), absolute delta power, absolute theta power, absolute alpha power, and the theta-plus-alpha to delta ratio.

Calculation of Theta-Alpha Variability (TAV)

We calculated TAV using a 2-second time window placed by visual inspection at the maximum and minimum theta-plus-alpha to delta ratio (TADR) values as represented in the quantitative EEG panel extracted from the entire length of the study. A 2-second window was chosen due to the short duration of time in which theta and alpha prevalence was seen to increase in recordings comprised overwhelmingly of continuous delta activity. We then calculated TAV as (TADRmaximum – TADRminimum)/(TADRmaximum + TADRminimum) (Figures 2,3).

Figure 2.

Figure 2

Figure 2

Figure 2

Figure 2

Figure 2

Open in a new tab

Determination of theta-alpha variability in a survivor with no neurologic sequelae

Quantitative EEG panel used to visually select TADR minimum (a) and maximum (c). The x-axis marks time in the tracing and y-axis value provides the TADR. Raw EEG (b, d) at the same time points. TAV is the percent change (e) in the TADR. This participant survived with no neurologic sequelae.

Figure 3.

Figure 3

Figure 3

Figure 3

Figure 3

Figure 3

Open in a new tab

Determination of theta-alpha variability in a patient who died

Example of TADR minimum (a) and maximum (c) within the Persyst 13 panel with raw EEG (b, d) at the same time points and calculation (e) of TAV.

Additional standard quantitative EEG variables of total power, relative power, and power ratios were calculated as the mean of five 2-second samples. Concurrent values for these variables were taken at the time points of TADR maximum and minimum then combined as an average with three additional 2-second samples taken consistently at the 5, 10, and 15-minute marks of the tracing, segments which were also confirmed to be devoid of artifact on the raw EEG. We used this sampling method to allow reproduction by rote using Persyst Systems software.

Any segment of EEG that included sporadic or periodic epileptiform discharges, non-cerebral artifact, and/or electrographic seizure activity was excluded from the analysis. When electrographic seizure activity was present, five quantitative samples of the EEG were selected at any time point up to EEG onset of seizure. No samples were taken from the ictal or post-ictal recording.

Statistical Analyses

We summarized patient demographic characteristics per outcome group (death, survival with neurological sequelae, survival without neurological sequelae) using means and standard deviation for continuous variables or counts and frequencies for categorical variables. We compared these characteristics between the three outcome groups using F-tests for continuous characteristics, or chi square tests for categorical variables.

We calculated Pearson correlation coefficients to investigate linear associations between TAV and other quantitative EEG variables of interest. To determine whether TAV was associated with clinical outcome, we used multinomial regression with logit link adjusting for age, admission glucose, and Blantyre coma score to estimate odds ratios (OR) and 95% confidence intervals (Wald). Ratios in addition to the TADR, given that they hold component parts in common, were not included in the final analysis due to concern for overfitting in multivariate regression. The visual minimum and maximum were sampled only for TADR. To aid with interpretation of the OR coefficients, we standardized the scale of each quantitative EEG variable in the model by subtracting the mean and dividing by two standard deviations. We compared the mean TAV and the standard dichotomous qualitative variability measure (variable versus not variable) using an independent group t-test to assess the association between the two measures. Finally, to evaluate the newly created variable’s predictive power on survival, we calculated the sensitivity and specificity of TAV for the outcome of death for varying thresholds.

RESULTS

EEG tracings, as described above, were available for review from 194 children admitted during the study period. Six patients had electrographic seizures during their EEG recording. Those who were younger or who had lower Blantyre Coma Scores on admission were significantly more likely to die (Table 1). Admission blood glucose, platelet count, and hematocrit were not associated with outcome.

Table 1:

Demographic characteristics of included participants (n=194) stratified by outcome

Admission factor Died (n=28) Survived with neurological sequelae (n=23) Survived without neurological sequelae (n=143) P value for difference between groups
Age (months) 39.1 (22.9) 44.0 (32.6) 55.2 (32.0) 0.0037
Glucose (mmol/L) 5.1 (3.5) 6.6 (3.5) 6.2 (2.3) 0.0960
Platelet count (× 103) 125 (209) 239 (439) 136 (304) 0.3121
Hematocrit (%) 23.3 (11.1) 19.5 (8.8) 22.1 (8.8) 0.3087
Blantyre Coma Score
0 7 (25%) 5 (21.7%) 7 (4.9%) 0.0006
1 15 (53.6%) 8 (34.8%) 60 (42.0%)
2 6 (21.4%) 10 (43.5%) 76 (53.1%)
Open in a new tab

Table 1: Variables expressed as mean (SD) for continuous variables, counts (percent) for categorical variables. P-values estimated using ANOVA for continuous variables, chi square for categorical; SD= standard deviation. Blantyre Coma Score is assessed in response to painful stimulus and has 3 elements: Motor (2 localizes painful stimulus, 1 withdraws, 0 no response), Verbal (2 cry or appropriate speech to painful stimulus, 1 moaning or groaning, 0 no response), and Eye Movement (1 follows moving visual stimulus, 0 does not follow; only scored if spontaneous eye opening is present).

TAV was not closely correlated with absolute alpha, theta, or delta power (Figure 4). Visual comparison of pairwise scatterplots between TAV and other quantitative EEG measures confirmed little correlation. The maximum Pearson correlation coefficient between TAV and other commonly used quantitative EEG factors was 0.17.

Figure 4.

Figure 4

Open in a new tab

Correlation between theta-alpha variability and other quantitative EEG measures

Descending diagonal (upper left to lower right) boxes contain histogram distribution of theta-alpha variability (Ѳ-α var.), absolute delta power (δ power), absolute theta power (Ѳ power), and absolute alpha power values (α power) in the EEGs analyzed. In the lower left, scatterplots with local regression (Loess) are found in boxes where correlating variables (horizontal and vertical) intersect. In the upper right, Pearson’s correlation coefficients for each pair of variables are noted. The maximum Pearson correlation coefficient between TAV and other commonly used quantitative EEG factors was 0.17, reflecting the independent nature of this novel variable.

When controlling for demographic and clinical factors known to be associated with outcome (age and Blantyre coma score), TAV was independently associated with both mortality and neurological morbidity in survivors. Children whose EEGs had higher values of TAV had a lower risk of death (OR= 0.934, 95% CI=0.902–0.966) or neurological sequelae (OR=0.960, 95% CI=0.932–0.990) compared to those with lower values. At a TAV threshold of 0.244, the sensitivity was 74% and specificity was 70% for an AUC of 0.755 in predicting outcome of death.

The continuous values of TAV were not associated with the qualitative EEG dichotomous factor of EEG variability (variable vs. not variable) (p= 0.0659). Consequently, a single cut point of TAV cannot be chosen to reflect the qualitative EEG factor of variable vs. not variable.

Examples of raw EEG data matched with the quantitative EEG panel and clinical outcome are included in Figures 2 and 3.

DISCUSSION

We formulated a new quantitative EEG variable, TAV, defined as the change in TADR over time. Power in the theta frequency band was included in our expanded definition of fast frequencies since theta activity is common in the normal waking EEG of young children. Quantification of power in the alpha frequency band has precedent in the intensive care unit EEG literature.17–19 Power of the beta frequency band (13–20Hz) was not included as a variable of interest given the possibility of benzodiazepine administration in some patients.

Whether or not anti-seizure medication was administered to patients prior to EEG recording was not easily discernable. Patient caregivers were asked at the time of admission whether the child had received any medication, including anti-seizure medications and antibiotics, prior to arrival at Queen Elizabeth Central Hospital. There are no reliable medical records- either paper or electronic- available for consultation in this setting. Information obtained verbally is subject to recall bias and was, therefore, not used in our study. Not taking prior anti-seizure medications into account when performing our analyses was also by design. Any appreciable effect of anti-seizure medications on TAV would have produced non-differential misclassification, biasing results towards the null. The association of TAV and outcome was robust, despite possible EEG changes from pre-hospital administration of benzodiazepines.

Of the additional power ratios measured in our study, the low-delta to delta ratio was not included in the final analysis as it did not capture segments of EEG that constituted a variable record on qualitative interpretation. The alpha to delta ratio was not chosen for final analysis since participants under the age of 9 years lack a full representation of alpha activity in the normal waking background.

In addition to its utility in mortality prediction, association of TAV and adverse outcomes in pediatric CM may shed light into disease pathogenesis. Highly increased brain volume (blurring of the gray-white junction, obliteration of CSF spaces with or without uncal herniation) on brain MRI is strongly associated with mortality in children with CM.8 It is assumed that the MRI features of highly increased brain volume are due to increases in intracranial pressure (ICP) but ICP has not been directly measured in children with these MRI findings due to risks of infection and resource limitations. As TAV may be a measure of the brain’s spontaneous ability to mount alpha and theta-range activity, our results support the potential link between paucity of fast frequencies on EEG and increased ICP.21 Furthermore, our results contribute to the growing body of evidence that a key pre-morbid pathophysiological mechanism responsible for mortality in children with CM is increased ICP.

Though qualitative EEG factors have been demonstrated to be associated with outcome in pediatric CM, study interpretation requires the expertise of a neurologist. In contrast, determination of values of TAV from a processed EEG requires comparatively little formal training. Technologists might be trained to distinguish non-ictal from potentially ictal segments before further processing the raw EEG data. The selection of an appropriate maximum and minimum of the TADR in our study was by visual inspection, increasing the possibility of inter-rater variation. The degree to which measurement of TAV is replicable across users should be determined. Moreover, the determination of a value for TAV relies on rules that may be followed by rote. Given the relative ease of selecting a visual minimum and maximum TADR, we feel the likelihood of substantial interrater variance is low and the prospects for automation favorable.

Decreasing the need for real time human EEG interpretation in resource poor contexts may increase the clinical usefulness of this technology. In CM (and other disease states) early identification of children at high risk of adverse outcomes allows enrichment of samples for clinical trials (enrolling children at higher risk of death, for example) or focusing limited early intervention services (e.g., physical therapy) to those identified at higher risk of neurological sequelae. Certainly, combinations of biomarkers (MRI with EEG, EEG with TCD) may be more valuable than a single modality in isolation.

Comparing the clinical usefulness of quantitative and qualitative EEG in CM is necessary in order to determine what information might be lost when quantitative methods are used in lieu of qualitative ones. Validating the usefulness of TAV in disease states other than CM is important in formulating conclusions about its potential pathophysiologic meaning in this disease state.

TAV is a newly developed quantitative EEG factor that is independently associated with both death and neurological morbidity in children with CM. The measure is able to predict death with good sensitivity and specificity. Further research in automation of EEG interpretation for use in resource-limited settings is warranted. Removing the need for interpretation by trained neurologists may increase the clinical usefulness of EEG studies in malaria endemic areas.

Acknowledgements

The authors thank the staff of the Blantyre Malaria Project (Blantyre, Malawi) whose clinical care of participants allowed this study to take place. We thank Persyst Corporation for donation of EEG interpretation software for this study’s use. Drs. Terrie Taylor, Susan Pierce, Lou Miller, and Taha Gholipour graciously provided critical feedback during manuscript development.

Conflicts of Interest and Source of Funding: Drs. Postels and Zelleke received grant support for completion of this project through NIH/ NIAID through the Intergovernmental Personnel Act. Dr. Postels receives grant funding unrelated to this project through NIH/ NIAID. Dr. Izem and Mr. Gai’s work in the Division of Biostatistics and Study Methodology was supported by Award Number UL1TR001876 from the NIH National Center for Advancing Translational Sciences. The remaining authors have no conflicts of interest or sources of funding to declare.

Contributor Information

Alexander Andrews, Department of Pediatrics, MedStar Georgetown University Hospital.

Tesfaye Zelleke, Department of Neurology, The George Washington University School of Medicine, Children’s National Hospital.

Dana Harrar, Department of Neurology, The George Washington University School of Medicine, Children’s National Hospital.

Rima Izem, Division of Biostatistics and Study Methodology, Children’s National Research Institute; Division of Epidemiology, The George Washington University School of Public Health; Department of Pediatrics, The George Washington University School of Medicine.

Jiaxiang Gai, Division of Biostatistics and Study Methodology, Children’s National Research Institute.

Douglas Postels, Department of Neurology, The George Washington University School of Medicine, Children’s National Hospital; Blantyre Malaria Project; University of Malawi College of Medicine; Blantyre, Malawi.

References

  • 1.World malaria report 2019. Geneva: World Health Organization; 2019. Available at https://www.who.int/publications/i/item/9789241565721. Accessed 25 September 2020. [Google Scholar]
  • 2.Idro R, Marsh K, John CC, Newton CR. Cerebral Malaria; Mechanisms Of Brain Injury And Strategies For Improved Neuro-Cognitive Outcome. Pediatr Res 2010;68(4):267–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Snow RW CM, Newton CR, Steketee RW, Craig MH. The Public Health Burden of Plasmodium falciparum Malaria in Africa: Deriving the Numbers. Working Paper No. 11, Disease Control Priorities Project. Bethseda, Maryland: Fogarty International Center. 2003. [Google Scholar]
  • 4.Birbeck GL, Molyneux ME, Kaplan PW, et al. Blantyre Malaria Project Epilepsy Study (BMPES) of neurological outcomes in retinopathy-positive paediatric cerebral malaria survivors: a prospective cohort study. Lancet Neurol 2010;9(12):1173–1181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Postels DG, Taylor TE, Molyneux M, et al. Neurologic outcomes in retinopathy-negative cerebral malaria survivors. Neurology. 2012;79(12):1268–1272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Postels DG, Wu X, Li C, et al. Admission EEG findings in diverse paediatric cerebral malaria populations predict outcomes. Malar J 2018;17:208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.O’Brien NF, Mutatshi Taty T, Moore-Clingenpeel M, et al. Transcranial Doppler Ultrasonography Provides Insights into Neurovascular Changes in Children with Cerebral Malaria. J Pediatr 2018;203:116–124. [DOI] [PubMed] [Google Scholar]
  • 8.Seydel KB, Kampondeni SD, Valim C, et al. Brain swelling and death in children with cerebral malaria. N Engl J Med 2015;372:1126–1137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Postels DG, Li C, Birbeck GL, et al. Brain MRI of children with retinopathy-negative cerebral malaria. Am J Trop Med Hyg 2014;91:943–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Potchen MJ, Birbeck GL, Demarco JK, et al. Neuroimaging findings in children with retinopathy-confirmed cerebral malaria. Eur J Radiol 2010;74:262–268. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Potchen MJ, Kampondeni SD, Seydel KB, et al. Acute brain MRI findings in 120 Malawian children with cerebral malaria: new insights into an ancient disease. AJNR Am J Neuroradiol 2012;33(9):1740–1746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Lewine JD, Plis S, Ulloa A, et al. Quantitative EEG Biomarkers for Mild Traumatic Brain Injury. J Clin Neurophysiol 2019;36:298–305. [DOI] [PubMed] [Google Scholar]
  • 13.Chen Y, Xu W, Wang L, et al. Transcranial Doppler combined with quantitative EEG brain function monitoring and outcome prediction in patients with severe acute intracerebral hemorrhage. Crit Care. 2018;22(1):36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Tolonen A, Särkelä MOK, Takala RSK, et al. Quantitative EEG Parameters for Prediction of Outcome in Severe Traumatic Brain Injury: Development Study. Clinical EEG and Neuroscience. 2018;49(4):248–257. [DOI] [PubMed] [Google Scholar]
  • 15.Vakulin A, D’Rozario A, Kim JW, et al. Quantitative sleep EEG and polysomnographic predictors of driving simulator performance in obstructive sleep apnea. J Clin Neurophysiol 2016;127:1428–1435. [DOI] [PubMed] [Google Scholar]
  • 16.Höller Y, Trinka E, Kalss G, Schiepek G, Michaelis R. Correlation of EEG spectra, connectivity, and information theoretical biomarkers with psychological states in the epilepsy monitoring unit - A pilot study. Epilepsy Behav 2019;99:106485. [DOI] [PubMed] [Google Scholar]
  • 17.Vespa PM, Boscardin WJ, Hovda DA, et al. Early and persistent impaired percent alpha variability on continuous electroencephalography monitoring as predictive of poor outcome after traumatic brain injury. J Neurosurg 2002;97:84–92. [DOI] [PubMed] [Google Scholar]
  • 18.Admiraal MM, Gilmore EJ, Van Putten M, Zaveri HP, Hirsch LJ, Gaspard N. Disruption of Brain-Heart Coupling in Sepsis. J Clin Neurophysiol 2017;34:413–420. [DOI] [PubMed] [Google Scholar]
  • 19.Vespa PM, Nuwer MR, Juhász C, et al. Early detection of vasospasm after acute subarachnoid hemorrhage using continuous EEG ICU monitoring. Electroencephalogr Clin Neurophysiol 1997;103:607–615. [DOI] [PubMed] [Google Scholar]
  • 20.Claassen J, Hirsch LJ, Kreiter KT, et al. Quantitative continuous EEG for detecting delayed cerebral ischemia in patients with poor-grade subarachnoid hemorrhage. Clin Neurophysiol. 2004;115:2699–2710. [DOI] [PubMed] [Google Scholar]
  • 21.Sheikh ZB, Maciel CB, Dhakar MB, Hirsch LJ, Gilmore EJ. Nonepileptic electroencephalographic correlates of episodic increases in intracranial pressure. J Clin Neurophysiol Epub 2020 Jul 20. [DOI] [PubMed] [Google Scholar]

RESOURCES