Long-range temporal correlations in the EEG bursts of human preterm babies

Quantitative analysis of high-frequency activity in neonatal EEG

OBJECTIVE

To determine the presence and potential utility of independent high-frequency activity recorded from scalp electrodes in the electroencephalogram (EEG) of newborns.

METHODS

We compare interburst intervals and continuous activity at different frequencies for EEGs retrospectively recorded at 256 Hz from 4 newborn groups: 1) 36 preterms (<32 weeks' gestational age, GA); 2) 12 preterms (32-37 weeks' GA); 3) 91 healthy full terms; 4) 15 full terms with hypoxic-ischemic encephalopathy (HIE). At 4 standard frequency bands (delta, 0.5-3 Hz; theta, 3-8 Hz; alpha, 8-15 Hz; beta, 15-30 Hz) and 3 higher-frequency bands (gamma1, 30-48 Hz; gamma2, 52-99 Hz; gamma3, 107-127 Hz), we compared power spectral densities (PSDs), quantitative features, and machine learning model performance. Feature selection and further machine learning methods were performed on one cohort.

RESULTS

We found significant (P < 0.01) differences in PSDs, quantitative analysis, and machine learning modelling at the higher-frequency bands. Machine learning models using only high-frequency features performed best in preterm groups 1 and 2 with a median (95% confidence interval, CI) Matthews correlation coefficient (MCC) of 0.71 (0.12-0.88) and 0.66 (0.36-0.76) respectively. Interburst interval-detector models using both high- and standard-bandwidths produced the highest median MCCs in all four groups. High-frequency features were largely independent of standard-bandwidth features, with only 11/84 (13.1%) of correlations statistically significant. Feature selection methods produced 7 to 9 high-frequency features in the top 20 feature set.

CONCLUSIONS

This is the first study to identify independent high-frequency activity in newborn EEG using in-depth quantitative analysis. Expanding the EEG bandwidths of analysis has the potential to improve both quantitative and machine-learning analysis, particularly in preterm EEG.

Clinical value of cortical bursting in preterm infants with intraventricular haemorrhage

BACKGROUND

In healthy preterm infants, cortical burst rate and temporal dynamics predict important measures such as brain growth. We hypothesised that in preterm infants with germinal matrix-intraventricular haemorrhage (GM-IVH), cortical bursting could provide prognostic information.

AIMS

We determined how cortical bursting was influenced by the injury, and whether this was related to developmental outcome.

STUDY DESIGN

Single-centre retrospective cohort study at University College London Hospitals, UK.

SUBJECTS

33 infants with GM-IVH ≥ grade II (median gestational age: 25 weeks).

OUTCOME MEASURES

We identified 47 EEGs acquired between 24 and 40 weeks corrected gestational age as part of routine clinical care. In a subset of 33 EEGs from 25 infants with asymmetric injury, we used the least-affected hemisphere as an internal comparison. We tested whether cortical burst rate predicted survival without severe impairment (median 2 years follow-up).

RESULTS

In asymmetric injury, cortical burst rate was lower over the worst- than least-affected hemisphere, and bursts over the worst-affected hemisphere were less likely to immediately follow bursts over the least-affected hemisphere than vice versa. Overall, burst rate was lower in cases of GM-IVH with parenchymal involvement, relative to milder structural injury grades. Higher burst rate modestly predicted survival without severe language (AUC 0.673) or motor impairment (AUC 0.667), which was partly mediated by structural injury grade.

CONCLUSIONS

Cortical bursting can index the functional injury after GM-IVH: perturbed burst initiation (rate) and propagation (inter-hemispheric dynamics) likely reflect associated grey matter and white matter damage. Higher cortical burst rate is reassuring for a positive outcome.

EEG phase-amplitude coupling to stratify encephalopathy severity in the developing brain

BACKGROUND

Neonatal hypoxic ischemic encephalopathy (HIE) is difficult to classify within the narrow therapeutic window of hypothermia. Neurophysiological biomarkers are needed for timely differentiation of encephalopathy severity within the short therapeutic window for initiation of hypothermia therapy.

METHODS

A novel analysis of mean Phase Amplitude Coupling index, PAC_m, of amplitudes high frequencies (12-30 Hz) coupled with phases of low (1,2 Hz) frequencies was calculated from the 6 h EEG recorded during the first day of life. PAC_m values were compared to identify differences between mild versus higher-grade HIE, respectively, for each of the EEG electrodes. A receiver operating characteristic curve was generated to examine the performance of PAC_m.

RESULTS

38 newborns with different HIE grades were enrolled in the first 6 h of life. Threshold PAC_m 0.001 at Fz, O1, O2, P3, and P4 had AUC >0.9 to differentiate HIE severity and predict the persistence of moderate to severe encephalopathy that requires treatment with hypothermia.

CONCLUSION

PAC is a promising biomarker to identify mild from higher severity of HIE after birth.

Prognostic value of neonatal EEG following therapeutic hypothermia in survivors of hypoxic-ischemic encephalopathy

OBJECTIVE

Early prediction of neurological deficits following neonatal hypoxic-ischemic encephalopathy (HIE) may help to target support. Neonatal animal models suggest that recovery following hypoxia-ischemia depends upon cortical bursting. To test whether this holds in human neonates, we correlated the magnitude of cortical bursting during recovery (≥postnatal day 3) with neurodevelopmental outcomes.

METHODS

We identified 41 surviving infants who received therapeutic hypothermia for HIE (classification at hospital discharge: 19 mild, 18 moderate, 4 severe) and had 9-channel electroencephalography (EEG) recordings as part of their routine care. We correlated burst power with Bayley-III cognitive, motor and language scores at median 24 months. To examine whether EEG offered additional prognostic information, we controlled for structural MRI findings.

RESULTS

Higher power of central and occipital cortical bursts predicted worse cognitive and language outcomes, and higher power of central cortical bursts predicted worse motor outcome, all independently of structural MRI findings.

CONCLUSIONS

Clinical EEG after postnatal day 3 may provide additional prognostic information by indexing persistent active mechanisms that either support recovery or exacerbate brain damage, especially in infants with less severe encephalopathy.

SIGNIFICANCE

These findings could allow for the effect of clinical interventions in the neonatal period to be studied instantaneously in the future.

Why Brain Criticality Is Clinically Relevant: A Scoping Review

The past 25 years have seen a strong increase in the number of publications related to criticality in different areas of neuroscience. The potential of criticality to explain various brain properties, including optimal information processing, has made it an increasingly exciting area of investigation for neuroscientists. Recent reviews on this topic, sometimes termed brain criticality, make brief mention of clinical applications of these findings to several neurological disorders such as epilepsy, neurodegenerative disease, and neonatal hypoxia. Other clinicallyrelevant domains - including anesthesia, sleep medicine, developmental-behavioral pediatrics, and psychiatry - are seldom discussed in review papers of brain criticality. Thorough assessments of these application areas and their relevance for clinicians have also yet to be published. In this scoping review, studies of brain criticality involving human data of all ages are evaluated for their current and future clinical relevance. To make the results of these studies understandable to a more clinical audience, a review of the key concepts behind criticality (e.g., phase transitions, long-range temporal correlation, self-organized criticality, power laws, branching processes) precedes the discussion of human clinical studies. Open questions and forthcoming areas of investigation are also considered.

Long-range temporal organisation of limb movement kinematics in human neonates

OBJECTIVE

Movement provides crucial sensorimotor information to the developing brain, evoking somatotopic cortical EEG activity. Indeed, temporal-spatial organisation of these movements, including a diverse repertoire of accelerations and limb combinations (e.g. unilateral progressing to bilateral), predicts positive sensorimotor outcomes. However, in current clinical practice, movements in human neonates are qualitatively characterised only during brief periods (a few minutes) of wakefulness, meaning that the vast majority of sensorimotor experience remains unsampled. Here our objective was to quantitatively characterise the long-range temporal organisation of the full repertoire of newborn movements, over multi-hour recordings.

METHODS

We monitored motor activity across 2-4 h in 11 healthy newborn infants (median 1 day old), who wore limb sensors containing synchronised tri-axial accelerometers and gyroscopes. Movements were identified using acceleration and angular velocity, and their organisation across the recording was characterised using cluster analysis and spectral estimation.

RESULTS

Movement occurrence was periodic, with a 1-hour cycle. Peaks in movement occurrence were associated with higher acceleration, and a higher proportion of movements being bilateral.

CONCLUSIONS

Neonatal movement occurrence is cyclical, with periods consistent with sleep-wake behavioural architecture. Movement kinematics are organised by these fluctuations in movement occurrence. Recordings that exceed 1-hour are necessary to capture the long-range temporal organisation of the full repertoire of newborn limb movements.

SIGNIFICANCE

Future work should investigate the prognostic value of combining these movement recordings with synchronised EEG, in at-risk infants.

Temporal ordering of input modulates connectivity formation in a developmental neuronal network model of the cortex

Preterm infant brain activity is discontinuous; bursts of activity recorded using EEG (electroencephalography), thought to be driven by subcortical regions, display scale free properties and exhibit a complex temporal ordering known as long-range temporal correlations (LRTCs). During brain development, activity-dependent mechanisms are essential for synaptic connectivity formation, and abolishing burst activity in animal models leads to weak disorganised synaptic connectivity. Moreover, synaptic pruning shares similar mechanisms to spike-timing dependent plasticity (STDP), suggesting that the timing of activity may play a critical role in connectivity formation. We investigated, in a computational model of leaky integrate-and-fire neurones, whether the temporal ordering of burst activity within an external driving input could modulate connectivity formation in the network. Connectivity evolved across the course of simulations using an approach analogous to STDP, from networks with initial random connectivity. Small-world connectivity and hub neurones emerged in the network structure—characteristic properties of mature brain networks. Notably, driving the network with an external input which exhibited LRTCs in the temporal ordering of burst activity facilitated the emergence of these network properties, increasing the speed with which they emerged compared with when the network was driven by the same input with the bursts randomly ordered in time. Moreover, the emergence of small-world properties was dependent on the strength of the LRTCs. These results suggest that the temporal ordering of burst activity could play an important role in synaptic connectivity formation and the emergence of small-world topology in the developing brain.

Behavioural discrimination of noxious stimuli in infants is dependent on brain maturation

Changes in facial expression are an essential form of social communication and in nonverbal infants are often used to alert care providers to pain-related distress. However, studies of early human brain development suggest that premature infants aged less than 34 weeks' gestation do not display discriminative brain activity patterns to equally salient noxious and innocuous events. Here we examine the development of facial expression in 105 infants, aged between 28 and 42 weeks' gestation. We show that the presence of facial expression change after noxious and innocuous stimulation is age-dependent and that discriminative facial expressions emerge from approximately 33 weeks' gestation. In a subset of 49 infants, we also recorded EEG brain activity and demonstrated that the temporal emergence of facial discrimination mirrors the developmental profile of the brain's ability to generate discriminative responses. Furthermore, within individual infants, the ability to display discriminative facial expressions is significantly related to brain response maturity. These data demonstrate that the emergence of behavioural discrimination in early human life corresponds to our brain's ability to discriminate noxious and innocuous events and raises fundamental questions as to how best to interpret infant behaviours when measuring and treating pain in premature infants.

Detecting bursts in the EEG of very and extremely premature infants using a multi-feature approach

•

Machine learning approach enables accurate detection of bursts in preterm EEG.

•

Features of amplitude and spectral shape capture discriminating information.

•

Improves reliability of estimates of inter-burst intervals.

Aim: To develop a method that segments preterm EEG into bursts and inter-bursts by extracting and combining multiple EEG features. Methods: Two EEG experts annotated bursts in individual EEG channels for 36 preterm infants with gestational age < 30 weeks. The feature set included spectral, amplitude, and frequency-weighted energy features. Using a consensus annotation, feature selection removed redundant features and a support vector machine combined features. Area under the receiver operator characteristic (AUC) and Cohen’s kappa (κ) evaluated performance within a cross-validation procedure. Results: The proposed channel-independent method improves AUC by 4–5% over existing methods (p < 0.001, n=36), with median (95% confidence interval) AUC of 0.989 (0.973–0.997) and sensitivity–specificity of 95.8–94.4%. Agreement rates between the detector and experts’ annotations, κ=0.72 (0.36–0.83) and κ=0.65 (0.32–0.81), are comparable to inter-rater agreement, κ=0.60 (0.21–0.74). Conclusions: Automating the visual identification of bursts in preterm EEG is achievable with a high level of accuracy. Multiple features, combined using a data-driven approach, improves on existing single-feature methods.

Early biomarkers of brain injury and cerebral hypo- and hyperoxia in the SafeBoosC II trial

BACKGROUND

The randomized clinical trial, SafeBoosC II, examined the effect of monitoring of cerebral oxygenation by near-infrared spectroscopy combined with a guideline on treatment when cerebral oxygenation was out of the target range. Data on cerebral oxygenation was collected in both the intervention and the control group. The primary outcome was the reduction in the burden of cerebral hypo- and hyperoxia between the two groups. In this study we describe the associations between the burden of cerebral hypo- and hyperoxia, regardless of allocation to intervention or control group, and the biomarkers of brain injury from birth till term equivalent age that was collected as secondary and explorative outcomes in the SafeBoosC II trial.

METHODS

Cerebral oxygenation was continuously monitored during the first 72h of life in 166 extremely preterm infants. Cranial ultrasound was performed at day 1,4,7,14, and 35 and at term. Electroencephalogram (EEG) was recorded at 64h. Blood-samples taken at 6 and 64 hours were analysed for the brain injury biomarkers; S100beta, brain-fatty-acid-binding-protein, and neuroketal. All analyses were conducted post hoc.

RESULTS

Significantly more infants with a cerebral burden of hypoxia within the 4th quartile versus infants within quartile 1-3 were diagnosed with severe intracranial haemorrhage (11/39 versus 11/117, p = 0.003), had low burst rate on EEG (12/28 versus 21/103, p = 0.015), or died (14/41 versus 18/123, p = 0.006), whereas none of these events were significantly associated with cerebral hyperoxia. The blood biomarkers were not significantly associated with the burden of cerebral hypo- or hyperoxia.

CONCLUSIONS

The explorative analysis showed that early burden of cerebral hypoxia, but not hyperoxia was significantly associated with low brain electrical activity and severe intracranial haemorrhage while none of the three blood biomarkers were associated with the burden of either cerebral hypo- or hyperoxia.

Extraction of features from sleep EEG for Bayesian assessment of brain development

Brain development can be evaluated by experts analysing age-related patterns in sleep electroencephalograms (EEG). Natural variations in the patterns, noise, and artefacts affect the evaluation accuracy as well as experts’ agreement. The knowledge of predictive posterior distribution allows experts to estimate confidence intervals within which decisions are distributed. Bayesian approach to probabilistic inference has provided accurate estimates of intervals of interest. In this paper we propose a new feature extraction technique for Bayesian assessment and estimation of predictive distribution in a case of newborn brain development assessment. The new EEG features are verified within the Bayesian framework on a large EEG data set including 1,100 recordings made from newborns in 10 age groups. The proposed features are highly correlated with brain maturation and their use increases the assessment accuracy.

The SafeBoosC II randomized trial: treatment guided by near-infrared spectroscopy reduces cerebral hypoxia without changing early biomarkers of brain injury

BACKGROUND

The SafeBoosC phase II multicentre randomized clinical trial investigated the benefits and harms of monitoring cerebral oxygenation by near-infrared spectroscopy (NIRS) combined with an evidence-based treatment guideline vs. no NIRS data and treatment as usual in the control group during the first 72 h of life. The trial demonstrated a significant reduction in the burden of cerebral hypoxia in the experimental group. We now report the blindly assessed and analyzed treatment effects on electroencephalographic (EEG) outcomes (burst rate and spectral edge frequency 95% (SEF95)) and blood biomarkers of brain injury (S100β, brain fatty acid-binding protein, and neuroketal).

METHODS

One hundred and sixty-six extremely preterm infants were randomized to either experimental or control group. EEG was recorded at 64 h of age and blood samples were collected at 6 and 64 h of age.

RESULTS

One hundred and thirty-three EEGs were evaluated. The two groups did not differ regarding burst rates (experimental 7.2 vs. control 7.7 burst/min) or SEF95 (experimental 18.1 vs. control 18.0 Hz). The two groups did not differ regarding blood S100β, brain fatty acid-binding protein, and neuroketal concentrations at 6 and 64 h (n = 123 participants).

CONCLUSION

Treatment guided by NIRS reduced the cerebral burden of hypoxia without affecting EEG or the selected blood biomarkers.

Structural damage in early preterm brain changes the electric resting state networks

A robust functional bimodality is found in the long-range spatial correlations of newborn cortical activity, and it likely provides the developmentally crucial functional coordination during the initial growth of brain networks. This study searched for possible acute effects on this large scale cortical coordination after acute structural brain lesion in early preterm infants. EEG recordings were obtained from preterm infants without (n=11) and with (n=6) haemorrhagic brain lesion detected in their routine ultrasound exam. The spatial cortical correlations in band-specific amplitudes were examined within two amplitude regimes, high and low amplitude periods, respectively. Technical validation of our analytical approach showed that bimodality of this kind is a genuine physiological characteristic of each brain network. It was not observed in datasets created from uniform noise, neither is it found between randomly paired signals. Hence, the observed bimodality arises from specific interactions between cortical regions. We found that significant long-range amplitude correlations are found in most signal pairs in both groups at high amplitudes, but the correlations are generally weaker in newborns with brain lesions. The group difference is larger during high mode, however the difference did not have any statistically apparent topology. Graph theoretical analysis confirmed a significantly larger weight dispersion in the newborns with brain lesion. Comparison of graph measures to a child's performance at two years showed that lower clustering coefficient and weight dispersion were both correlated to better neurodevelopmental outcomes. Our findings suggest that the common preterm brain haemorrhage causes diffuse changes in the functional long-range cortical correlations. It has been recently recognized that the high mode network activity is crucial for early brain development. The present observations may hence offer a mechanistic link between early lesion and the later emergence of complex neurocognitive sequelae.

Objective differentiation of neonatal EEG background grades using detrended fluctuation analysis

A quantitative and objective assessment of background electroencephalograph (EEG) in sick neonates remains an everyday clinical challenge. We studied whether long range temporal correlations quantified by detrended fluctuation analysis (DFA) could be used in the neonatal EEG to distinguish different grades of abnormality in the background EEG activity. Long-term EEG records of 34 neonates were collected after perinatal asphyxia, and their background was scored in 1 h epochs (8 h in each neonate) as mild, moderate or severe. We applied DFA on 15 min long, non-overlapping EEG epochs (n = 1088) filtered from 3 to 8 Hz. Our formal feasibility study suggested that DFA exponent can be reliably assessed in only part of the EEG epochs, and in only relatively short time scales (10-60 s), while it becomes ambiguous if longer time scales are considered. This prompted further exploration whether paradigm used for quantifying multifractal DFA (MF-DFA) could be applied in a more efficient way, and whether metrics from MF-DFA paradigm could yield useful benchmark with existing clinical EEG gradings. Comparison of MF-DFA metrics showed a significant difference between three visually assessed background EEG grades. MF-DFA parameters were also significantly correlated to interburst intervals quantified with our previously developed automated detector. Finally, we piloted a monitoring application of MF-DFA metrics and showed their evolution during patient recovery from asphyxia. Our exploratory study showed that neonatal EEG can be quantified using multifractal metrics, which might offer a suitable parameter to quantify the grade of EEG background, or to monitor changes in brain state that take place during long-term brain monitoring.

Identification of Criticality in Neuronal Avalanches: II. A Theoretical and Empirical Investigation of the Driven Case

The observation of apparent power laws in neuronal systems has led to the suggestion that the brain is at, or close to, a critical state and may be a self-organised critical system. Within the framework of self-organised criticality a separation of timescales is thought to be crucial for the observation of power-law dynamics and computational models are often constructed with this property. However, this is not necessarily a characteristic of physiological neural networks-external input does not only occur when the network is at rest/a steady state. In this paper we study a simple neuronal network model driven by a continuous external input (i.e. the model does not have an explicit separation of timescales from seeding the system only when in the quiescent state) and analytically tuned to operate in the region of a critical state (it reaches the critical regime exactly in the absence of input-the case studied in the companion paper to this article). The system displays avalanche dynamics in the form of cascades of neuronal firing separated by periods of silence. We observe partial scale-free behaviour in the distribution of avalanche size for low levels of external input. We analytically derive the distributions of waiting times and investigate their temporal behaviour in relation to different levels of external input, showing that the system's dynamics can exhibit partial long-range temporal correlations. We further show that as the system approaches the critical state by two alternative 'routes', different markers of criticality (partial scale-free behaviour and long-range temporal correlations) are displayed. This suggests that signatures of criticality exhibited by a particular system in close proximity to a critical state are dependent on the region in parameter space at which the system (currently) resides.

Identification of Criticality in Neuronal Avalanches: I. A Theoretical Investigation of the Non-driven Case

In this paper, we study a simple model of a purely excitatory neural network that, by construction, operates at a critical point. This model allows us to consider various markers of criticality and illustrate how they should perform in a finite-size system. By calculating the exact distribution of avalanche sizes, we are able to show that, over a limited range of avalanche sizes which we precisely identify, the distribution has scale free properties but is not a power law. This suggests that it would be inappropriate to dismiss a system as not being critical purely based on an inability to rigorously fit a power law distribution as has been recently advocated. In assessing whether a system, especially a finite-size one, is critical it is thus important to consider other possible markers. We illustrate one of these by showing the divergence of susceptibility as the critical point of the system is approached. Finally, we provide evidence that power laws may underlie other observables of the system that may be more amenable to robust experimental assessment.

Adaptive time-varying detrended fluctuation analysis

Detrended fluctuation analysis (DFA) is a technique commonly used to assess and quantify the presence of long-range temporal correlations (LRTCs) in neurophysiological time series. Convergence of the method is asymptotic only and therefore its application assumes a constant scaling exponent. However, most neurophysiological data are likely to involve either spontaneous or experimentally induced scaling exponent changes. We present a novel extension of the DFA method that permits the characterisation of time-varying scaling exponents. The effectiveness of the methodology in recovering known changes in scaling exponents is demonstrated through its application to synthetic data. The dependence of the method on its free parameters is systematically explored. Finally, application of the methodology to neurophysiological data demonstrates that it provides experimenters with a way to identify previously un-recognised changes in the scaling exponent in the data. We suggest that this methodology will make it possible to go beyond a simple demonstration of the presence of scaling to an appreciation of how it may vary in response to either intrinsic changes or experimental perturbations.