Clinical outcome prediction with an automated EEG trend, Brain State of the Newborn, after perinatal asphyxia

OBJECTIVE

To evaluate the utility of a fully automated deep learning -based quantitative measure of EEG background, Brain State of the Newborn (BSN), for early prediction of clinical outcome at four years of age.

METHODS

The EEG monitoring data from eighty consecutive newborns was analyzed using the automatically computed BSN trend. BSN levels during the first days of life (a of total 5427 hours) were compared to four clinical outcome categories: favorable, cerebral palsy (CP), CP with epilepsy, and death. The time dependent changes in BSN-based prediction for different outcomes were assessed by positive/negative predictive value (PPV/NPV) and by estimating the area under the receiver operating characteristic curve (AUC).

RESULTS

The BSN values were closely aligned with four visually determined EEG categories (p < 0·001), as well as with respect to clinical milestones of EEG recovery in perinatal Hypoxic Ischemic Encephalopathy (HIE; p < 0·003). Favorable outcome was related to a rapid recovery of the BSN trend, while worse outcomes related to a slow BSN recovery. Outcome predictions with BSN were accurate from 6 to 48 hours of age: For the favorable outcome, the AUC ranged from 95 to 99% (peak at 12 hours), and for the poor outcome the AUC ranged from 96 to 99% (peak at 12 hours). The optimal BSN levels for each PPV/NPV estimate changed substantially during the first 48 hours, ranging from 20 to 80.

CONCLUSIONS

We show that the BSN provides an automated, objective, and continuous measure of brain activity in newborns.

SIGNIFICANCE

The BSN trend discloses the dynamic nature that exists in both cerebral recovery and outcome prediction, supports individualized patient care, rapid stratification and early prognosis.

The Association of Therapeutic Hypothermia With Seizure Burden in Neonates With Hypoxic-Ischemic Encephalopathy

OBJECTIVES

To compare seizure burden between newborn infants treated with therapeutic hypothermia (TH) and those that were not and to compare the need for antiseizure medications (ASM) in a cohort of infants who were diagnosed with neonatal hypoxic-ischemic encephalopathy (HIE).

METHODS

This was a retrospective cohort study on infants born after 35 weeks' gestation, diagnosed with moderate to severe HIE, monitored with amplitude-integrated electroencephalography (aEEG) and eligible for TH. Infants born before the implementation of TH in 2008 were compared with infants born thereafter who received TH. Seizure burden was assessed from aEEG as total time in minutes of seizures activity per hour of recording. Other clinical and demographic data were retrieved from a prospective local database of infants with HIE.

RESULTS

Overall, 149 of 207 infants were included in the study: 112 exposed to TH and 37 not exposed. Cooled infants had a lower seizure burden overall (0.4 vs 2.3 min/h, P < 0.001) and were also less likely to be treated with ASM (74% vs 100%, P < 0.001). In multivariable regression models, not exposed to TH, having a depressed aEEG background, and having higher Apgar scores were associated with higher seizure burden (incidence rate ratio: 4.78 for noncooled infants, P < 0.001); also, not exposed to TH was associated with a higher likelihood of multidrug ASM (odds ratio: 4.83, P < 0.001).

CONCLUSIONS

TH in infants with moderate to severe HIE is associated with significant reduction of seizure burden and ASM therapy.

Sleep State Trend (SST), a bedside measure of neonatal sleep state fluctuations based on single EEG channels

OBJECTIVE

To develop and validate an automated method for bedside monitoring of sleep state fluctuations in neonatal intensive care units.

METHODS

A deep learning-based algorithm was designed and trained using 53 EEG recordings from a long-term (a)EEG monitoring in 30 near-term neonates. The results were validated using an independent dataset from 30 polysomnography recordings. In addition, we constructed Sleep State Trend (SST), a bedside-ready means for visualizing classifier outputs.

RESULTS

The accuracy of quiet sleep detection in the training data was 90%, and the accuracy was comparable (85-86 %) in all bipolar derivations available from the 4-electrode recordings. The algorithm generalized well to a polysomnography dataset, showing 81% overall accuracy despite different signal derivations. SST allowed an intuitive, clear visualization of the classifier output.

CONCLUSIONS

Fluctuations in sleep states can be detected at high fidelity from a single EEG channel, and the results can be visualized as a transparent and intuitive trend in the bedside monitors.

SIGNIFICANCE

The Sleep State Trend (SST) may provide caregivers and clinical studies a real-time view of sleep state fluctuations and its cyclicity.

Newborns at high risk for brain injury: the role of the amplitude-integrated electroencephalography

OBJECTIVE

Amplitude-integrated electroencephalography (aEEG) is a simplified bedside neurophysiology tool that has been implemented in the neonatal intensive care unit and studied in an extensive range of clinical applications in the past decade. This critical review aimed to evaluate a variety of clinical applications of aEEG monitoring in diagnosis, clinical management, and prognosis assessment in critically ill neonates.

SOURCES

The databases of Pubmed, SciELO, Lilacs, and Cochrane, books, and other online resources were consulted, as well as sources of professional experiences.

SUMMARY OF FINDINGS

The clinical use of aEEG to access real-time brain function, background activity, and utility in seizures detection has been described. A critical review was realized considering the authors' professional experience. Newborns with hypoxic-ischemic encephalopathy and seizures screening represent the most common studied population. However, several studies have shown interesting applications on preterm infants, newborns with congenital heart disease, and other clinical situations of high risk of injury to the developing brain.

CONCLUSION

The aEEG has shown to be a useful non-invasive bedside monitor that aids in evaluating brain function, background activity, and cyclicity. aEEG findings have also demonstrated good prognostic value in a group of critically ill neonates. The aEEG seizure diagnosis capability has limitations, which have been already well established. The use of neonatal brain monitoring such as aEEG was shown to give valuable information in several high-risk clinical situations.

Prognostic value of somatosensory-evoked potentials in the newborn with hypoxic-ischemic encephalopathy after the introduction of therapeutic hypothermia

To establish the ability of somatosensory-evoked potentials (SEPs) to detect neurological damage in neonatal patients with hypoxic-ischemic encephalopathy (HIE) treated with therapeutic hypothermia (TH). Retrospective study including 84 neonates ≥ 36 weeks of gestational age with HIE and TH with SEPs performed in the first 14 days of life. SEPs from the median nerve were performed after completion of TH. Either unilateral or bilateral absence of N20, or unilateral or bilateral latency ≥ 36 ms, was considered pathological. All newborns underwent a cerebral resonance imaging (MRI) at between days 7 and 14 of life and a neurodevelopmental evaluation using the Brunet-Lezine test at two years of age; a global Brunet-Lezine test score < 70 was considered unfavorable. The risk of moderate-to-severe alteration on basal ganglia-thalamic (BGT) and/or white matter areas on MRI for pathological SEPs was as follows: odds ratio 95% IC: 23.1 (6.9-76.9), sensitivity 78.6%, specificity 86.3%, positive predictive value 75.9%, and negative predictive value 88%. The BGT and internal capsule were the areas with the greatest risk of lesion with an altered SEPs: odds ratio 95% IC 93.1 (11.1-777.8). The risk of neurodevelopmental impairment for pathological SEPs was odds ratio 95% IC: 38.5 (4.4-335.3), sensitivity 91.7%, specificity 77.8% positive predictive value 52.4%, and negative predictive value 97.2%.

CONCLUSION

The present study demonstrates the good predictive capacity of SEPs performed in the first two weeks of life in newborns with HIE and TH to detect an increased risk of neuroimaging lesions and neurodevelopmental impairment at two years of age.

WHAT IS KNOWN

• Bilateral absence of the N20 cortical component of somatosensory evoked potentials has been associated with poor neurological outcome in neonates with hypoxic-ischemic encephalopathy.

WHAT IS NEW

• This work confirms the predictive capacity of SEPs by adding two important aspects: the value of latency when interpreting SEPs results and the absence of effect of the hypothermia method used on the results of SEPs.

Automated detection of artefacts in neonatal EEG with residual neural networks

BACKGROUND AND OBJECTIVE

To develop a computational algorithm that detects and identifies different artefact types in neonatal electroencephalography (EEG) signals.

METHODS

As part of a larger algorithm, we trained a Residual Deep Neural Network on expert human annotations of EEG recordings from 79 term infants recorded in a neonatal intensive care unit (112 h of 18-channel recording). The network was trained using 10 fold cross validation in Matlab. Artefact types included: device interference, EMG, movement, electrode pop, and non-cortical biological rhythms. Performance was assessed by prediction statistics and further validated on a separate independent dataset of 13 term infants (143 h of 3-channel recording). EEG pre-processing steps, and other post-processing steps such as averaging probability over a temporal window, were also included in the algorithm.

RESULTS

The Residual Deep Neural Network showed high accuracy (95%) when distinguishing periods of clean, artefact-free EEG from any kind of artefact, with a median accuracy for individual patient of 91% (IQR: 81%-96%). The accuracy in identifying the five different types of artefacts ranged from 57%-92%, with electrode pop being the hardest to detect and EMG being the easiest. This reflected the proportion of artefact available in the training dataset. Misclassification as clean was low for each artefact type, ranging from 1%-11%. The detection accuracy was lower on the validation set (87%). We used the algorithm to show that EEG channels located near the vertex were the least susceptible to artefact.

CONCLUSION

Artefacts can be accurately and reliably identified in the neonatal EEG using a deep learning algorithm. Artefact detection algorithms can provide continuous bedside quality assessment and support EEG review by clinicians or analysis algorithms.

Python-Embedded Plugin Implementation in ICM+: Novel Tools for Neuromonitoring Time Series Analysis with Examples Using CENTER-TBI Datasets

With the appearance of publicly available, high-resolution, physiological datasets in neurocritical care, like Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI), there is a growing need for tools that could be used by clinical researchers to interrogate this information-rich data. The ICM+ software is widely used for processing data acquired from bedside monitors. Considering the growing popularity of scripting simple-syntax programming languages like Python, particularly among clinical researchers, we have developed an interface in ICM+ that provides a streamlined way of adding Python scripting functionality to the ICM+ calculation engine. The new interface imposes certain requirements on the scripts and needs an accompanying descriptor file that tells ICM+ about the functions implemented, so that they become available to the end user in the same way as native ICM+ functions. ICM+ also now includes a tool that eases the creation of Python functions to be imported. The Python extension works very efficiently, and any user with some degree of experience in scripting can use it to enrich capabilities of ICM+. Depending on the data analysed and calculations performed, Python functions are 15-60% slower than built-in ICM+ functions, which is a more-than-acceptable trade-off for empowering ICM+ with the unlimited analytical freedom offered by extensive Python libraries.

Towards multimodal brain monitoring in asphyxiated newborns with amplitude-integrated EEG and simultaneous somatosensory evoked potentials

BACKGROUND

Somatosensory evoked potentials (SEPs) offer an additional bedside tool for outcome prediction after perinatal asphyxia.

AIMS

To assess the reliability of SEPs recorded with bifrontoparietal amplitude-integrated electroencephalography (aEEG) brain monitoring setup for outcome prediction in asphyxiated newborns undergoing therapeutic hypothermia.

STUDY DESIGN

Retrospective observational single-center study.

SUBJECTS

27 consecutive asphyxiated full- or near-term newborns (25 under hypothermia) that underwent median nerve aEEG-SEPs as part of their clinical evaluation at the neonatal intensive care unit of Helsinki University Hospital.

OUTCOME MEASURES

aEEG-SEP classification (present, absent or unreliable) was compared to classification of SEPs recorded with a full EEG montage (EEG-SEP), and outcome determined from medical records at approximately 12-months-age. Unfavorable outcome included death, cerebral palsy, or severe epilepsy.

RESULTS

The aEEG-SEP and EEG-SEP classifications were concordant in 21 of the 22 newborns with both recordings available. All five newborns with bilaterally absent aEEG-SEPs had absent EEG-SEPs and the four with outcome information available had an unfavorable outcome (one was lost to follow-up). Of the newborns with aEEG-SEPs present, all with follow-up exams available had bilaterally present EEG-SEPs and a favorable outcome (one was lost to follow-up). One newborn with unilaterally absent aEEG-SEP at 25 h of age had bilaterally present EEG-SEPs on the next day, and a favorable outcome.

CONCLUSIONS

aEEG-SEPs recorded during therapeutic hypothermia on the first postnatal days are reliable for assessing brain injury severity. Adding SEP into routine aEEG brain monitoring offers an additional tool for very early outcome prediction after birth asphyxia.

Effect of morphine on cerebral activity of extremely premature infants

BACKGROUND

Morphine is widely used for sedation in premature infants and may have long-term effects on neurodevelopmental outcome.

AIMS

To assess its effect on cerebral activity of extremely preterm infants.

METHODS

Retrospective study in infants less than 28 weeks of gestational age (GA), treated with morphine and monitored with amplitude integrated EEG (aEEG), that were assessed at baseline and 6 consecutive 2 h periods. At each period, the background activity score and presence of cycling were determined.

RESULTS

A total of 140 infants were included. Background aEEG activity score worsened in 40% of the infants by 6 h after morphine initiation and cycling, initially present in 70% of the infants, could be detected in only 10% after 10 h. On multivariable analysis, adjusted for baseline aEEG score and GA, the background activity decreased by 0.85 (95% confidence interval (CI): 0.74-0.97) immediately after morphine initiation, 0.72 (CI: 0.63-0.83) at 2 h later and on average by 0.64 (CI: 0.57-0.71) at each subsequent 2 h periods. The relative risk of cycling loss at 2 h and 4 h after morphine administration was 1.58 (CI: 1.15-2.16) and 3.37 (CI: 2.23-5.08), respectively.

CONCLUSIONS

Continuous infusion of morphine in extremely preterm infants significantly depresses their cerebral activity soon after its commencement.

Use of 2.1 MHz MRI scanner for brain imaging and its preliminary results in stroke

Cerebral stroke greatly contributes to death and disability rates in China and the whole world. Effective non-invasive imaging device for bedside monitoring of stroke is critically needed in clinically. This study developed a lightweight (350 kg) and low-footprint magnetic resonance imaging (MRI) system for brain imaging. Static magnetic field was built using an H-typed permanent magnet, which has 50.9 mT magnetic field strength (corresponding to 2.167 MHz proton Larmor frequency). Biplanar gradient coils were designed using the target field method based on dipole equivalent. Radio-frequency coils were optimized by particle swarm optimization. The 2 MHz MRI system was deployed in the Department of Neurology of hospital to test its performance in stroke imaging detection. Gradient recall echo and fast spin echo sequences were utilized to acquire T1- and T2-weighted MR images, respectively. Brain images of a healthy volunteer, a patient with hemorrhagic stroke, a patient of ischemic stroke, and a patient with ischemic stroke and images from 17-day long-term monitoring of hemorrhagic stroke were obtained with a 1.5 mm * 2.0 mm spatial resolution in plane, and a 10 mm thickness.

The use of automated pupillometry to assess cerebral autoregulation: a retrospective study

BACKGROUND

Critically ill patients are at high risk of developing neurological complications. Among all the potential aetiologies, brain hypoperfusion has been advocated as one of the potential mechanisms. Impairment of cerebral autoregulation (CAR) can result in brain hypoperfusion. However, assessment of CAR is difficult at bedside. We aimed to evaluate whether the automated pupillometer might be able to detect impaired CAR in critically ill patients.

METHODS

We included 92 patients in this retrospective observational study; 52 were septic. CAR was assessed using the Mxa index, which is the correlation index between continuous recording of cerebral blood flow velocities using the transcranial Doppler and invasive arterial blood pressure over 8 ± 2 min. Impaired CAR was defined as an Mxa > 0.3. Automated pupillometer (Neuroptics, Irvine, CA, USA) was used to assess the pupillary light reflex concomitantly to the CAR assessment.

RESULTS

The median Mxa was 0.33 in the whole cohort (0.33 in septic patients and 0.31 in the non-septic patients; p = 0.77). A total of 51 (55%) patients showed impaired CAR, 28 (54%) in the septic group and 23 (58%) in the non-septic group. We found a statistically significant although weak correlation between Mxa and the Neurologic Pupil Index (r 2 = 0.04; p = 0.048) in the whole cohort as in septic patients (r 2 = 0.11; p = 0.026); no correlation was observed in non-septic patients and for other pupillometry-derived variables.

CONCLUSIONS

Automated pupillometry cannot predict CAR indices such as Mxa in a heterogeneous population of critically ill patients.

Probabilistic graphical model identifies clusters of EEG patterns in recordings from neonates

OBJECTIVES

In this paper we introduce a novel method for the evaluation of neonatal brain function via multivariate EEG (electroencephalography) signal processing and embedding into a probabilistic graph, the so called Chow-Liu tree.

METHODS

Using 28 EEG recordings of preterm and term neonate infants the complex features of the EEG signals were constructed in the form of a Chow-Liu tree. The trees were embedded into a 3 dimensional Euclidean space. Clustering of specific EEG patterns was done by complete linkage algorithm.

RESULTS

Our analytic tool was able to build clusters of patients with pathological EEG findings. In particular, we were able to make a visual proof on a 3d multidimensional scaling coordinate system with a good performance. The distances (graph edit distance) between Chow-Liu trees of different infants were proportional to the clinical findings of corresponding infants.

CONCLUSION

Our method may provide a basis for the future development of a diagnostic/prognostic non-invasive brain monitoring tool which will be able to differentiate between a variety of complex clinical findings.

SIGNIFICANCE

This model addresses relevant issues in neonatology and neuropediatrics in terms of identification of possible clinical factors which interfere with normal brain development and will allow fast unbiased recognition of infants with specific pathological EEG findings.

Effect of Mild Hypocapnia on Critical Closing Pressure and Other Mechanoelastic Parameters of the Cerebrospinal System

OBJECTIVE

Brain arterial critical closing pressure (CrCP) has been studied in several diseases such as traumatic brain injury (TBI), subarachnoid haemorrhage, hydrocephalus, and in various physiological scenarios: intracranial hypertension, decreased cerebral perfusion pressure, hypercapnia, etc. Little or nothing so far has been demonstrated to characterise change in CrCP during mild hypocapnia.

METHOD

We retrospectively analysed recordings of intracranial pressure (ICP), arterial blood pressure (ABP) and blood flow velocity from 27 severe TBI patients (mean 39.5 ± 3.4 years, 6 women) in whom a ventilation increase (20% increase in respiratory minute volume) was performed over 50 min as part of a standard clinical CO₂ reactivity test. CrCP was calculated using the Windkessel model of cerebral arterial flow. Arteriolar wall tension (WT) was calculated as a difference between CrCP and ICP. The compartmental compliances arterial (C _a ) and cerebrospinal fluid space (C _i ) were also evaluated.

RESULTS

During hypocapnia, ICP decreased from 17±6.8 to 13.2±6.6 mmHg (p < 0.000001). Wall tension increased from 14.5 ± 9.9 to 21.7±9.1 mmHg (p < 0.0002). CrCP, being a sum of WT + ICP, changed significantly from 31.5 ± 11.9 mmHg to 34.9±11.1 mmHg (p < 0.002), and the closing margin (ABP-CrCP) remained constant at an average value of 60 mmHg. C _a decreased significantly during hypocapnia by 30% (p < 0.00001) and C _i increased by 26% (p < 0.003).

CONCLUSION

During hypocapnia in TBI patients, ICP decreases and WT increases. CrCP increases slightly as the rise in wall tension outweighs the decrease in ICP. The closing margin remained unchanged, suggesting that the risk of hypocapnia-induced ischemia might not be increased.

Evaluation of SEPs in asphyxiated newborns using a 4-electrode aEEG brain monitoring set-up

•

Neonatal SEPs can be reliably detected using the 4-electrode aEEG monitoring setup.

•

SEP is discernible in most fullterm newborns with 300 averages.

•

Recording SEPs jointly with aEEG facilitates SEP assessment after birth asphyxia.

Objective

To evaluate the reliability of recording cortical somatosensory evoked potentials (SEPs) in asphyxiated newborns using the 4-electrode setup applied in routine long-term amplitude-integrated EEG (aEEG) brain monitoring and to assess the number of averages needed for reliably detecting the cortical responses.

Methods

We evaluated median nerve SEPs in 50 asphyxiated full-term newborns. The SEP interpretation (present or absent) from the original recordings with 21-electrodes and approximately 600 trials served as the reference. This was compared to SEP classification (absent, present, or unreliable) based on a reduced (300 or 150) number of averages, and to classification based on only four electrodes (F3, P3, F4, P4).

Results

Compared to the original classification, cortical SEPs were uniformly interpreted as present or absent in all 50 newborns with the 4-electrode setup and 600 averages. Reducing number of averages to 300 still resulted in correct SEP interpretation in 49/50 newborns with 21-electrode setup, and 46/50 newborns with 4-electrode setup.

Conclusions

Evaluation of early cortical neonatal SEPs is reliable from the 4-electrode setup commonly used in aEEG monitoring. SEP is discernible in most newborns with 300 averages.

Significance

Adding SEP into routine aEEG monitoring offers an additional tool for early neonatal neurophysiological evaluation.

Amplitude-integrated electroencephalography for seizure detection in newborn infants

The amplitude-integrated electroencephalogram (aEEG) is a filtered and compressed EEG trend that can be used for long-term monitoring of brain function in patients of all ages. aEEG is increasingly used in neonatal intensive care units since several studies have shown its utility in high-risk newborn infants. Main indications for aEEG monitoring include early evaluation of brain function after perinatal asphyxia and seizure detection. The aEEG is usually recorded from one or two channels derived from parietal, central, or frontal leads. Although the aEEG is very useful for identifying high-risk infants and infants with seizures, the compressed trend has limitations with regards to detection of individual seizures. However, modern monitors also display the corresponding EEG (aEEG/EEG), which increases the probability of detecting single brief seizures. For improved evaluation of electrocortical brain activity the aEEG/EEG should be assessed together with repeated conventional EEGs or multi-channel EEG monitoring in a multi-disciplinary team.

Occurrence of CPPopt Values in Uncorrelated ICP and ABP Time Series

OBJECTIVES

Optimal cerebral perfusion pressure (CPPopt) is a concept that uses the pressure reactivity (PRx)-CPP relationship over a given period to find a value of CPP at which PRx shows best autoregulation. It has been proposed that this relationship be modelled by a U-shaped curve, where the minimum is interpreted as being the CPP value that corresponds to the strongest autoregulation. Owing to the nature of the calculation and the signals involved in it, the occurrence of CPPopt curves generated by non-physiological variations of intracranial pressure (ICP) and arterial blood pressure (ABP), termed here "false positives", is possible. Such random occurrences would artificially increase the yield of CPPopt values and decrease the reliability of the methodology.In this work, we studied the probability of the random occurrence of false-positives and we compared the effect of the parameters used for CPPopt calculation on this probability.

MATERIALS AND METHODS

To simulate the occurrence of false-positives, uncorrelated ICP and ABP time series were generated by destroying the relationship between the waves in real recordings. The CPPopt algorithm was then applied to these new series and the number of false-positives was counted for different values of the algorithm's parameters.

RESULTS

The percentage of CPPopt curves generated from uncorrelated data was demonstrated to be 11.5%.

CONCLUSION

This value can be minimised by tuning some of the calculation parameters, such as increasing the calculation window and increasing the minimum PRx span accepted on the curve.

Neonatal somatosensory evoked potentials persist during hypothermia

AIM

Treatment with therapeutic hypothermia has challenged the use of amplitude-integrated electroencephalography in predicting outcomes after perinatal asphyxia. In this study, we assessed the feasibility and gain of somatosensory evoked potentials (SEP) during hypothermia.

METHODS

This retrospective study comprised neonates from 35 + 6 to 42 + 2 gestational weeks and treated for asphyxia or hypoxic-ischaemic encephalopathy at Helsinki University Hospital between 14 February 2007 and 23 December 2009. This period was partly before the introduction of routine therapeutic hypothermia, which enabled us to include normothermic neonates who would these days receive hypothermia treatment. We analysed SEPs from 47 asphyxiated neonates and compared the results between 23 normothermic and 24 hypothermic neonates.

RESULTS

Our data showed that hypothermia led to SEP latencies lengthening by a few milliseconds, but the essential gain for predicting outcomes by SEPs was preserved during hypothermia. Of the 24 hypothermic neonates, bilaterally absent SEPs were associated with poor outcome in 2/2 neonates, normal SEPs were associated with good outcomes in 13/15 neonates and 5/7 neonates with unilaterally absent or grossly delayed SEPs had a poor outcome.

CONCLUSION

Our findings indicated that SEPs were a reliable tool for evaluating the somatosensory system in asphyxiated neonates in both normothermic and hypothermic conditions.

Evoked potentials recorded during routine EEG predict outcome after perinatal asphyxia

OBJECTIVE

To evaluate the added value of somatosensory (SEPs) and visual evoked potentials (VEPs) recorded simultaneously with routine EEG in early outcome prediction of newborns with hypoxic-ischemic encephalopathy under modern intensive care.

METHODS

We simultaneously recorded multichannel EEG, median nerve SEPs, and flash VEPs during the first few postnatal days in 50 term newborns with hypoxic-ischemic encephalopathy. EEG background was scored into five grades and the worst two grades were considered to indicate poor cerebral recovery. Evoked potentials were classified as absent or present. Clinical outcome was determined from the medical records at a median age of 21months. Unfavorable outcome included cerebral palsy, severe mental retardation, severe epilepsy, or death.

RESULTS

The accuracy of outcome prediction was 98% with SEPs compared to 90% with EEG. EEG alone always predicted unfavorable outcome when it was inactive (n=9), and favorable outcome when it was normal or only mildly abnormal (n=17). However, newborns with moderate or severe EEG background abnormality could have either favorable or unfavorable outcome, which was correctly predicted by SEP in all but one newborn (accuracy in this subgroup 96%). Absent VEPs were always associated with an inactive EEG, and an unfavorable outcome. However, presence of VEPs did not guarantee a favorable outcome.

CONCLUSIONS

SEPs accurately predict clinical outcomes in newborns with hypoxic-ischemic encephalopathy and improve the EEG-based prediction particularly in those newborns with severely or moderately abnormal EEG findings.

SIGNIFICANCE

SEPs should be added to routine EEG recordings for early bedside assessment of newborns with hypoxic-ischemic encephalopathy.

A testbed for optimizing electrodes embedded in the skull or in artificial skull replacement pieces used after injury

BACKGROUND

Custom-fitted skull replacement pieces are often used after a head injury or surgery to replace damaged bone. Chronic brain recordings are beneficial after injury/surgery for monitoring brain health and seizure development. Embedding electrodes directly in these artificial skull replacement pieces would be a novel, low-risk way to perform chronic brain monitoring in these patients. Similarly, embedding electrodes directly in healthy skull would be a viable minimally-invasive option for many other neuroscience and neurotechnology applications requiring chronic brain recordings.

NEW METHOD

We demonstrate a preclinical testbed that can be used for refining electrode designs embedded in artificial skull replacement pieces or for embedding directly into the skull itself. Options are explored to increase the surface area of the contacts without increasing recording contact diameter to maximize recording resolution.

RESULTS

Embedding electrodes in real or artificial skull allows one to lower electrode impedance without increasing the recording contact diameter by making use of conductive channels that extend into the skull. The higher density of small contacts embedded in the artificial skull in this testbed enables one to optimize electrode spacing for use in real bone.

COMPARISON WITH EXISTING METHODS

For brain monitoring applications, skull-embedded electrodes fill a gap between electroencephalograms recorded on the scalp surface and the more invasive epidural or subdural electrode sheets.

CONCLUSIONS

Embedding electrodes into the skull or in skull replacement pieces may provide a safe, convenient, minimally-invasive alternative for chronic brain monitoring. The manufacturing methods described here will facilitate further testing of skull-embedded electrodes in animal models.

The use and yield of continuous EEG in critically ill patients: A comparative study of three centers

OBJECTIVE

Continuous EEG (cEEG) monitoring of critically ill patients has gained widespread use, but there is substantial reported variability in its use. We analyzed cEEG and antiseizure drug (ASD) usage at three high volume centers.

METHODS

We utilized a multicenter cEEG database used daily as a clinical reporting tool in three tertiary care sites (Emory Hospital, Brigham and Women's Hospital and Yale - New Haven Hospital). We compared the cEEG usage patterns, seizure frequency, detection of rhythmic/periodic patterns (RPP), and ASD use between the sites.

RESULTS

5792 cEEG sessions were analyzed. Indication for cEEG monitoring and recording duration were similar between the sites. Seizures detection rate was nearly identical between the three sites, ranging between 12.3% and 13.6%. Median time to first seizure and detection rate of RPPs were similar. There were significant differences in doses of levetiracetam, valproic acid, and lacosamide used between the three sites.

CONCLUSIONS

There was remarkable uniformity in seizure detection rates within three high volume centers. In contrast, dose of ASD used frequently differed between the three sites.

SIGNIFICANCE

These large volume data are in line with recent guidelines regarding cEEG use. Difference in ASD use suggests discrepancies in how cEEG results influence patient management.

Brain Oxygenation Monitoring

A mismatch between cerebral oxygen supply and demand can lead to cerebral hypoxia/ischemia and deleterious outcomes. Cerebral oxygenation monitoring is an important aspect of multimodality neuromonitoring. It is increasingly deployed whenever intracranial pressure monitoring is indicated. Although there is a large body of evidence demonstrating an association between cerebral hypoxia/ischemia and poor outcomes, it remains to be determined whether restoring cerebral oxygenation leads to improved outcomes. Randomized prospective studies are required to address uncertainties about cerebral oxygenation monitoring and management. This article describes the different methods of monitoring cerebral oxygenation, their indications, evidence base, limitations, and future perspectives.

Comparison of four near-infrared spectroscopy devices shows that they are only suitable for monitoring cerebral oxygenation trends in preterm infants

AIM

Measuring cerebral oxygenation using near-infrared spectroscopy (NIRS) has taken on an increasingly important role in the field of neonatology. Several companies have already developed commercial devices, and more publications are reporting absolute boundary values or percentiles for neonates. We compared four commercially used devices to discover whether they provided consistent results in the same patients.

METHODS

We recruited nine preterm infants and tested them for 2 h, using sensors from two different devices. The measurements were carried out six times on each child, so that all four devices were compared with each other. A total of 54 measurements were conducted. The following devices were compared: the NIRO 200 (Hamamatsu Photonics K.K), the INVOS 5100c (Somanetics), the Fore-Sight (CAS Med.) and the SenSmart X-100 (NONIN).

RESULTS

The cerebral tissue oxygenation data yielded by the individual devices differed significantly from each other, ranging from a minimum difference of 2.93% to a maximum difference of 12.66%.

CONCLUSION

The commercially available NIRS devices showed highly significant differences in local cerebral tissue oxygenation levels, to the extent that the industry cannot agree on uniform and reproducible standards. Therefore, NIRS should only be used for trend measurements in preterm infants.

Line length as a robust method to detect high-activity events: automated burst detection in premature EEG recordings

OBJECTIVE

EEG is a valuable tool for evaluation of brain maturation in preterm babies. Preterm EEG constitutes of high voltage burst activities and more suppressed episodes, called interburst intervals (IBIs). Evolution of background characteristics provides information on brain maturation and helps in prediction of neurological outcome. The aim is to develop a method for automated burst detection.

METHODS

Thirteen polysomnography recordings were used, collected at preterm postmenstrual age of 31.4 (26.1-34.4)weeks. We developed a burst detection algorithm based on the feature line length and compared it with manual scorings of clinical experts and other published methods.

RESULTS

The line length-based algorithm is robust (84.27% accuracy, 84.00% sensitivity, 85.70% specificity). It is not critically dependent on the number of measurement channels, because two channels still provide 82% accuracy. Furthermore, it approximates well clinically relevant features, such as median IBI duration 5.45 (4.00-7.11)s, maximum IBI duration 14.02 (8.73-18.80)s and burst percentage 48.89 (35.45-60.12)%, with a median deviation of respectively 0.65s, 1.96s and 6.55%.

CONCLUSION

Automated assessment of long-term preterm EEG is possible and its use will optimize EEG interpretation in the NICU.

SIGNIFICANCE

This study takes a first step towards fully automatic analysis of the preterm brain.

Smartphones as pocketable labs: visions for mobile brain imaging and neurofeedback

Mobile brain imaging solutions, such as the Smartphone Brain Scanner, which combines low cost wireless EEG sensors with open source software for real-time neuroimaging, may transform neuroscience experimental paradigms. Normally subject to the physical constraints in labs, neuroscience experimental paradigms can be transformed into dynamic environments allowing for the capturing of brain signals in everyday contexts. Using smartphones or tablets to access text or images may enable experimental design capable of tracing emotional responses when shopping or consuming media, incorporating sensorimotor responses reflecting our actions into brain machine interfaces, and facilitating neurofeedback training over extended periods. Even though the quality of consumer neuroheadsets is still lower than laboratory equipment and susceptible to environmental noise, we show that mobile neuroimaging solutions, like the Smartphone Brain Scanner, complemented by 3D reconstruction or source separation techniques may support a range of neuroimaging applications and thus become a valuable addition to high-end neuroimaging solutions.

Density-modulated t's array, a new technique of processed electroencephalogram, for monitoring the effects of midazolam and nitrous oxide during spinal anesthesia

We have investigated the utility of a new electroencephalogram (EEG) processing system, density-modulated t's array (DTA), which we have installed in a laptop personal computer together with density-modulated spectral array (DSA) for clinical monitoring. Ten patients scheduled for orthopedic operations on the lower extremities were anesthetized with 0.5% bupivacaine intrathecally, 50% nitrous oxide in oxygen by mask, and midazolam at a dose of 0.1 mg/kg intravenously. Immediately following the administration of the drugs, the power at the frequencies between 15 and 20 Hz increased. However, the power at these higher frequencies disappeared gradually and the power in the delta band and the smaller one in the alpha band became predominant. This pattern of dominant-band shift on the DSA and DTA was observed in all the patients. In three of the patients, the sedation level remained stable as judged by the absence of body movement, quiet, regular breathing and stable hemodynamics as well as steady EEG frequency distribution throughout the operations. They awoke from anesthesia rapidly on withdrawal of nitrous oxide, with return of the power at the higher frequencies. In the other seven patients, the power at the higher frequencies suddenly reappeared on the DSA and DTA during operation and slight movements of the head and upper limbs were observed with rises in blood pressure and heart rate. In three of these seven patients, the EEG change notably preceded the physiological activities by a few minutes. On the DTA, the occurrence of any significant clinical phenomenon was displayed in a color representing at value greater than ±3. The DTA, testing power changes in the EEG at each 1-Hz interval for significant difference, permits the visual and quantitative assessment of EEG changes.