The Harvard Automated Processing Pipeline for Electroencephalography (HAPPE): Standardized Processing Software for Developmental and High-Artifact Data

EEG Integrated Platform Lossless (EEG-IP-L) pre-processing pipeline for objective signal quality assessment incorporating data annotation and blind source separation

BACKGROUND

The methods available for pre-processing EEG data are rapidly evolving as researchers gain access to vast computational resources; however, the field currently lacks a set of standardized approaches for data characterization, efficient interactive quality control review procedures, and large-scale automated processing that is compatible with High Performance Computing (HPC) resources.

NEW METHOD

In this paper we describe an infrastructure for the development of standardized procedures for semi and fully automated pre-processing of EEG data. Our pipeline incorporates several methods to isolate cortical signal from noise, maintain maximal information from raw recordings and provide comprehensive quality control and data visualization. In addition, batch processing procedures are integrated to scale up analyses for processing hundreds or thousands of data sets using HPC clusters.

RESULTS

We demonstrate here that by using the EEG Integrated Platform Lossless (EEG-IP-L) pipeline's signal quality annotations, significant increase in data retention is achieved when applying subsequent post-processing ERP segment rejection procedures. Further, we demonstrate that the increase in data retention does not attenuate the ERP signal.

CONCLUSIONS

The EEG-IP-L state provides the infrastructure for an integrated platform that includes long-term data storage, minimal data manipulation and maximal signal retention, and flexibility in post processing strategies.

An Inhomogeneous Point-process Model for the Assessment of the Brain-to-Heart Functional Interplay: a Pilot Study

We propose a novel computational framework for the estimation of functional directional brain-to-heart interplay in an instantaneous fashion. The framework is based on inhomogeneous point-process models for human heartbeat dynamics and employs inverse-Gaussian probability density functions characterizing the timing of R-peak events. The instantaneous estimation of the functional directional coupling is based on the definition of point-process transfer entropy, which is here retrieved from heart rate variability (HRV) and Electroencephalography (EEG) power spectral series gathered from 12 healthy subjects undergoing significant sympathovagal changes induced by a cold-pressor test. Results suggest that EEG oscillations dynamically influence heartbeat dynamics with specific time delays in the 30-60s and 90-120s ranges, and through a functional activity over specific cortical regions.

Modulation of Frontal Oscillatory Power during Blink Suppression in Children: Effects of Premonitory Urge and Reward

There is a dearth of studies examining the underlying mechanisms of blink suppression and the effects of urge and reward, particularly those measuring subsecond electroencephalogram (EEG) brain dynamics. To address these issues, we designed an EEG study to ask 3 questions: 1) How does urge develop? 2) What are EEG-correlates of blink suppression? 3) How does reward change brain dynamics related to urge suppression? This study examined healthy children (N = 26, age 8–12 years) during blink suppression under 3 conditions: blink freely (i.e., no suppression), blink suppressed, and blink suppressed for reward. During suppression conditions, children used a joystick to indicate their subjective urge to blink. Results showed that 1) half of the trials were associated with clearly defined urge time course of ~7 s, which was accompanied by EEG delta (1–4 Hz) power reduction localized at anterior cingulate cortex (ACC); 2) the EEG correlates of blink suppression were found in left prefrontal theta (4–8 Hz) power elevation; and 3) reward improved blink suppression performance while reducing the EEG delta power observed in ACC. We concluded that the empirically supported urge time course and underlying EEG modulations provide a subsecond chronospatial model of the brain dynamics during urge- and reward-mediated blink suppression.

Adjusting ADJUST: Optimizing the ADJUST algorithm for pediatric data using geodesic nets

A major challenge for electroencephalograph (EEG) studies on pediatric populations is that large amounts of data are lost due to artifacts (e.g., movement and blinks). Independent component analysis (ICA) can separate artifactual and neural activity, allowing researchers to remove such artifactual activity and retain a greater percentage of EEG data for analyses. However, manual identification of artifactual components is time-consuming and requires subjective judgment. Automated algorithms, like ADJUST and ICLabel, have been validated on adults, but to our knowledge, no such algorithms have been optimized for pediatric data. Therefore, in an attempt to automate artifact selection for pediatric data collected with geodesic nets, we modified ADJUST's algorithm. Our "adjusted-ADJUST" algorithm was compared to the "original-ADJUST" algorithm and ICLabel in adults, children, and infants on three different performance measures: respective classification agreement with expert coders, the number of trials retained following artifact removal, and the reliability of the EEG signal after preprocessing with each algorithm. Overall, the adjusted-ADJUST algorithm performed better than the original-ADJUST algorithm and no ICA correction with adult and pediatric data. Moreover, in some measures, it performed better than ICLabel for pediatric data. These results indicate that optimizing existing algorithms improves artifact classification and retains more trials, potentially facilitating EEG studies with pediatric populations. Adjusted-ADJUST is freely available under the terms of the GNU General Public License at: https://github.com/ChildDevLab/MADE-EEG-preprocessing-pipeline/tree/master/adjusted_adjust_scripts.

Multifractal and Entropy-Based Analysis of Delta Band Neural Activity Reveals Altered Functional Connectivity Dynamics in Schizophrenia

Dynamic functional connectivity (DFC) was established in the past decade as a potent approach to reveal non-trivial, time-varying properties of neural interactions – such as their multifractality or information content –, that otherwise remain hidden from conventional static methods. Several neuropsychiatric disorders were shown to be associated with altered DFC, with schizophrenia (SZ) being one of the most intensely studied among such conditions. Here we analyzed resting-state electroencephalography recordings of 14 SZ patients and 14 age- and gender-matched healthy controls (HC). We reconstructed dynamic functional networks from delta band (0.5–4 Hz) neural activity and captured their spatiotemporal dynamics in various global network topological measures. The acquired network measure time series were made subject to dynamic analyses including multifractal analysis and entropy estimation. Besides group-level comparisons, we built a classifier to explore the potential of DFC features in classifying individual cases. We found stronger delta-band connectivity, as well as increased variance of DFC in SZ patients. Surrogate data testing verified the true multifractal nature of DFC in SZ, with patients expressing stronger long-range autocorrelation and degree of multifractality when compared to controls. Entropy analysis indicated reduced temporal complexity of DFC in SZ. When using these indices as features, an overall cross-validation accuracy surpassing 89% could be achieved in classifying individual cases. Our results imply that dynamic features of DFC such as its multifractal properties and entropy are potent markers of altered neural dynamics in SZ and carry significant potential not only in better understanding its pathophysiology but also in improving its diagnosis. The proposed framework is readily applicable for neuropsychiatric disorders other than schizophrenia.

The Self and Its Prolonged Intrinsic Neural Timescale in Schizophrenia

Schizophrenia (SCZ) can be characterized as a basic self-disorder that is featured by abnormal temporal integration on phenomenological (experience) and psychological (information processing) levels. Temporal integration on the neuronal level can be measured by the brain's intrinsic neural timescale using the autocorrelation window (ACW) and power-law exponent (PLE). Our goal was to relate intrinsic neural timescales (ACW, PLE), as a proxy of temporal integration on the neuronal level, to temporal integration related to self-disorder on psychological (Enfacement illusion task in electroencephalography) and phenomenological (Examination of Anomalous Self-Experience [EASE]) levels. SCZ participants exhibited prolonged ACW and higher PLE during the self-referential task (Enfacement illusion), but not during the non-self-referential task (auditory oddball). The degree of ACW/PLE change during task relative to rest was significantly reduced in self-referential task in SCZ. A moderation model showed that low and high ACW/PLE exerted differential impact on the relationship of self-disorder (EASE) and negative symptoms (PANSS). In sum, we demonstrate abnormal prolongation in intrinsic neural timescale during self-reference in SCZ including its relation to basic self-disorder and negative symptoms. Our results point to abnormal relation of self and temporal integration at the core of SCZ constituting a "common currency" of neuronal, psychological, and phenomenological levels.

Methodological Considerations on EEG Electrical Reference: A Functional Brain-Heart Interplay Study

The growing interest in the study of functional brain-heart interplay (BHI) has motivated the development of novel methodological frameworks for its quantification. While a combination of electroencephalography (EEG) and heartbeat-derived series has been widely used, the role of EEG preprocessing on a BHI quantification is yet unknown. To this extent, here we investigate on four different EEG electrical referencing techniques associated with BHI quantifications over 4-minute resting-state in 15 healthy subjects. BHI methods include the synthetic data generation model, heartbeat-evoked potentials, heartbeat-evoked oscillations, and maximal information coefficient (MIC). EEG signals were offline referenced under the Cz channel, common average, mastoids average, and Laplacian method, and statistical comparisons were performed to assess similarities between references and between BHI techniques. Results show a topographical agreement between BHI estimation methods depending on the specific EEG reference. Major differences between BHI methods occur with the Laplacian reference, while major differences between EEG references are with the MIC analysis. We conclude that the choice of EEG electrical reference may significantly affect a functional BHI quantification.

Quantifying Functional Links between Brain and Heartbeat Dynamics in the Multifractal Domain: a Preliminary Analysis

Quantification of brain-heart interplay (BHI) has mainly been performed in the time and frequency domains. However, such functional interactions are likely to involve nonlinear dynamics associated with the two systems. To this extent, in this preliminary study we investigate the functional coupling between multifractal properties of Electroencephalography (EEG) and Heart Rate Variability (HRV) series using a channel- and time scale-wise maximal information coefficient analysis. Experimental results were gathered from 24 healthy volunteers undergoing a resting state and a cold-pressure test, and suggest that significant changes between the two experimental conditions might be associated with nonlinear quantifiers of the multifractal spectrum. Particularly, major brain-heart functional coupling was associated with the secondorder cumulant of the multifractal spectrum. We conclude that a functional nonlinear relationship between brain- and heartbeat-related multifractal sprectra exist, with higher values associated with the resting state.

Evoked Potentials and EEG Analysis in Rett Syndrome and Related Developmental Encephalopathies: Towards a Biomarker for Translational Research

Rett syndrome is a debilitating neurodevelopmental disorder for which no disease-modifying treatment is available. Fortunately, advances in our understanding of the genetics and pathophysiology of Rett syndrome has led to the development of promising new therapeutics for the condition. Several of these therapeutics are currently being tested in clinical trials with others likely to progress to clinical trials in the coming years. The failure of recent clinical trials for Rett syndrome and other neurodevelopmental disorders has highlighted the need for electrophysiological or other objective biological markers of treatment response to support the success of clinical trials moving forward. The purpose of this review is to describe the existing studies of electroencephalography (EEG) and evoked potentials (EPs) in Rett syndrome and discuss the open questions that must be addressed before the field can adopt these measures as surrogate endpoints in clinical trials. In addition to summarizing the human work on Rett syndrome, we also describe relevant studies with animal models and the limited research that has been carried out on Rett-related disorders, particularly methyl-CpG binding protein 2 (MECP2) duplication syndrome, CDKL5 deficiency disorder, and FOXG1 disorder.

EEG-IP: an international infant EEG data integration platform for the study of risk and resilience in autism and related conditions

BACKGROUND

Establishing reliable predictive and diganostic biomarkers of autism would enhance early identification and facilitate targeted intervention during periods of greatest plasticity in early brain development. High impact research on biomarkers is currently limited by relatively small sample sizes and the complexity of the autism phenotype.

METHODS

EEG-IP is an International Infant EEG Data Integration Platform developed to advance biomarker discovery by enhancing the large scale integration of multi-site data. Currently, this is the largest multi-site standardized dataset of infant EEG data.

RESULTS

First, multi-site data from longitudinal cohort studies of infants at risk for autism was pooled in a common repository with 1382 EEG longitudinal recordings, linked behavioral data, from 432 infants between 3- to 36-months of age. Second, to address challenges of limited comparability across independent recordings, EEG-IP applied the Brain Imaging Data Structure (BIDS)-EEG standard, resulting in a harmonized, extendable, and integrated data state. Finally, the pooled and harmonized raw data was preprocessed using a common signal processing pipeline that maximizes signal isolation and minimizes data reduction. With EEG-IP, we produced a fully standardized data set, of the pooled, harmonized, and pre-processed EEG data from multiple sites.

CONCLUSIONS

Implementing these integrated solutions for the first time with infant data has demonstrated success and challenges in generating a standardized multi-site data state. The challenges relate to annotation of signal sources, time, and ICA analysis during pre-processing. A number of future opportunities also emerge, including validation of analytic pipelines that can replicate existing findings and/or test novel hypotheses.

Day-to-Day Test-Retest Reliability of EEG Profiles in Children With Autism Spectrum Disorder and Typical Development

Biomarker development is currently a high priority in neurodevelopmental disorder research. For many types of biomarkers (particularly biomarkers of diagnosis), reliability over short periods is critically important. In the field of autism spectrum disorder (ASD), resting electroencephalography (EEG) power spectral densities (PSD) are well-studied for their potential as biomarkers. Classically, such data have been decomposed into pre-specified frequency bands (e.g., delta, theta, alpha, beta, and gamma). Recent technical advances, such as the Fitting Oscillations and One-Over-F (FOOOF) algorithm, allow for targeted characterization of the features that naturally emerge within an EEG PSD, permitting a more detailed characterization of the frequency band-agnostic shape of each individual's EEG PSD. Here, using two resting EEGs collected a median of 6 days apart from 22 children with ASD and 25 typically developing (TD) controls during the Feasibility Visit of the Autism Biomarkers Consortium for Clinical Trials, we estimate test-retest reliability based on the characterization of the PSD shape in two ways: (1) Using the FOOOF algorithm we estimate six parameters (offset, slope, number of peaks, and amplitude, center frequency and bandwidth of the largest alpha peak) that characterize the shape of the EEG PSD; and (2) using nonparametric functional data analyses, we decompose the shape of the EEG PSD into a reduced set of basis functions that characterize individual power spectrum shapes. We show that individuals exhibit idiosyncratic PSD signatures that are stable over recording sessions using both characterizations. Our data show that EEG activity from a brief 2-min recording provides an efficient window into characterizing brain activity at the single-subject level with desirable psychometric characteristics that persist across different analytical decomposition methods. This is a necessary step towards analytical validation of biomarkers based on the EEG PSD and provides insights into parameters of the PSD that offer short-term reliability (and thus promise as potential biomarkers of trait or diagnosis) vs. those that are more variable over the short term (and thus may index state or other rapidly dynamic measures of brain function). Future research should address the longer-term stability of the PSD, for purposes such as monitoring development or response to treatment.

Microstates-based resting frontal alpha asymmetry approach for understanding affect and approach/withdrawal behavior

The role of resting frontal alpha-asymmetry in explaining neural-mechanisms of affect and approach/withdrawal behavior is still debatable. The present study explores the ability of the quasi-stable resting EEG asymmetry information and the associated neurovascular synchronization/desynchronization in bringing more insight into the understanding of neural-mechanisms of affect and approach/withdrawal behavior. For this purpose, a novel frontal alpha-asymmetry based on microstates, that assess quasi-stable EEG scalp topography information, is proposed and compared against standard frontal-asymmetry. Both proposed and standard frontal alpha-asymmetries were estimated from thirty-nine healthy volunteers resting-EEG simultaneously acquired with resting-fMRI. Further, neurovascular mechanisms of these asymmetry measures were estimated through EEG-informed fMRI. Subsequently, the Hemodynamic Lateralization Index (HLI) of the neural-underpinnings of both asymmetry measures was assessed. Finally, the robust correlation of both asymmetry-measures and their HLI’s with PANAS, BIS/BAS was carried out. The standard resting frontal-asymmetry and its HLI yielded no significant correlation with any psychological-measures. However, the microstate resting frontal-asymmetry correlated significantly with negative affect and its neural underpinning’s HLI significantly correlated with Positive/Negative affect and BIS/BAS measures. Finally, alpha-BOLD desynchronization was observed in neural-underpinning whose HLI correlated significantly with negative affect and BIS. Hence, the proposed resting microstate-frontal asymmetry better assesses the neural-mechanisms of affect, approach/withdrawal behavior.

Maternal antenatal anxiety and electrophysiological functioning amongst a sub-set of preschoolers participating in the GUSTO cohort

BACKGROUND

Antenatal maternal anxiety is a risk for offspring psychological and cognitive difficulties. The preschool years represent an important time for brain development, and so may be a window for intervention. However, electrophysiological investigations of maternal anxiety and preschoolers' brain functioning are lacking. We ask whether anxiety symptoms predict neurophysiology, and consider timing specificity (26-weeks antenatal or 24-months postnatal), form of insult (anxiety symptoms, per se, or also depression symptoms), and offspring gender.

METHODS

The sample consisted of a subset of 71 mothers and their 3 year old children taking part in the prospective birth cohort, GUSTO. Mothers provided antenatal (26 weeks) and postnatal (2 years) anxiety and depressive symptomatology data, respectively via the "State Trait Anxiety Questionnaire" and the "Edinburgh Postpartum Depression Scale." Offspring provided electrophysiological data, obtained while they indicated the emotional expression of actors whose facial expressions remained consistent throughout a pre-switch block, but were reversed at "post-switch."

RESULTS

Three electrophysiological components linked to different information processing stages were identified. The two earliest occurring components (i.e., the N1 and P2) differed across blocks. During post-switch, both were significantly predicted by maternal anxiety, after controlling for pre-switch neurophysiology. Similar results were observed with depression. Antenatal mental health remained a significant predictor after controlling for postnatal mental health.

CONCLUSION

In combination with past work, these findings suggest the importance of reducing symptoms in women prior to and during pregnancy, and offering support to offspring early in development.

Use of longitudinal EEG measures in estimating language development in infants with and without familial risk for autism spectrum disorder

Language development in children with autism spectrum disorder (ASD) varies greatly among affected individuals and is a strong predictor of later outcomes. Younger siblings of children with ASD have increased risk of ASD, but also language delay. Identifying neural markers of language outcomes in infant siblings could facilitate earlier intervention and improved outcomes. This study aimed to determine whether EEG measures from the first 2-years of life can explain heterogeneity in language development in children at low- and high-risk for ASD, and to determine whether associations between EEG measures and language development are different depending on ASD risk status or later ASD diagnosis. In this prospective longitudinal study EEG measures collected between 3-24 months were used in a multivariate linear regression model to estimate participants' 24-month language development. Individual baseline longitudinal EEG measures included (1) the slope of EEG power across 3-12 months or 3-24 months of life for 6 canonical frequency bands, (2) estimated EEG power at age 6-months for the same frequency bands, and (3) terms representing the interaction between ASD risk status and EEG power measures. Modeled 24-month language scores using EEG data from either the first 2-years (Pearson R = 0.70, 95% CI 0.595-0.783, P=1x10^-18) or the first year of life (Pearson R=0.66, 95% CI 0.540-0.761, P=2.5x10^-14) were highly correlated with observed scores. All models included significant interaction effects of risk on EEG measures, suggesting that EEG-language associations are different depending on risk status, and that different brain mechanisms effect language development in low-versus high-risk infants.

Neurophysiologic Characterization of Resting State Connectivity Abnormalities in Schizophrenia Patients

Background: Patients with schizophrenia show abnormal spontaneous oscillatory activity in scalp-level electroencephalographic (EEG) responses across multiple frequency bands. While oscillations play an essential role in the transmission of information across neural networks, few studies have assessed the frequency-specific dynamics across cortical source networks at rest. Identification of the neural sources and their dynamic interactions may improve our understanding of core pathophysiologic abnormalities associated with the neuropsychiatric disorders. Methods: A novel multivector autoregressive modeling approach for assessing effective connectivity among cortical sources was developed and applied to resting-state EEG recordings obtained from n = 139 schizophrenia patients and n = 126 healthy comparison subjects. Results: Two primary abnormalities in resting-state networks were detected in schizophrenia patients. The first network involved the middle frontal and fusiform gyri and a region near the calcarine sulcus. The second network involved the cingulate gyrus and the Rolandic operculum (a region that includes the auditory cortex). Conclusions: Schizophrenia patients show widespread patterns of hyper-connectivity across a distributed network of the frontal, temporal, and occipital brain regions. Results highlight a novel approach for characterizing alterations in connectivity in the neuropsychiatric patient populations. Further mechanistic characterization of network functioning is needed to clarify the pathophysiology of neuropsychiatric and neurological diseases.

Using the N400 event-related potential to study word learning from context in children from low- and higher-socioeconomic status homes

Children from low-socioeconomic status (SES) homes have significantly smaller vocabularies than their higher-SES peers, a gap that increases over the course of the school years. One reason for the increase in this vocabulary gap during the school years is that children from low-SES homes learn fewer words from context than their higher-SES peers. To better understand how the process of word learning from context might differ in children related to SES, we investigated changes in the N400 event-related potential (ERP) as children from low- and higher-SES homes learned new words using only the surrounding linguistic context. There were no differences in the N400 response to known words related to SES. In response to the target word being learned, children from higher-SES homes, like adults in previous studies, exhibited an attenuation of the N400 across exposures as they attached meaning to it. Children from low-SES homes did not show this same attenuation. These findings support previous work showing that children from low-SES homes may have differences or more variability in the neural components supporting language processing, and they extend previous work to illustrate how this variability may relate to word learning and, ultimately, vocabulary growth.

A Three-Class Classification of Cognitive Workload Based on EEG Spectral Data

Evaluation of cognitive workload finds its application in many areas, from educational program assessment through professional driver health examination to monitoring the mental state of people carrying out jobs of high responsibility, such as pilots or airline traffic dispatchers. Estimation of multilevel cognitive workload is a task usually realized in a subject-dependent way, while the present research is focused on developing the procedure of subject-independent evaluation of cognitive workload level. The aim of the paper is to estimate cognitive workload level in accordance with subject-independent approach, applying classical machine learning methods combined with feature selection techniques. The procedure of data acquisition was based on registering the EEG signal of the person performing arithmetical tasks divided into six intervals of advancement. The analysis included the stages of preprocessing, feature extraction, and selection, while the final step covered multiclass classification performed with several models. The results discussed show high maximal accuracies achieved: ~91% for both the validation dataset and for the cross-validation approach for kNN model.

Multifractal and entropy analysis of resting-state electroencephalography reveals spatial organization in local dynamic functional connectivity

Functional connectivity of the brain fluctuates even in resting-state condition. It has been reported recently that fluctuations of global functional network topology and those of individual connections between brain regions expressed multifractal scaling. To expand on these findings, in this study we investigated if multifractality was indeed an inherent property of dynamic functional connectivity (DFC) on the regional level as well. Furthermore, we explored if local DFC showed region-specific differences in its multifractal and entropy-related features. DFC analyses were performed on 62-channel, resting-state electroencephalography recordings of twelve young, healthy subjects. Surrogate data testing verified the true multifractal nature of regional DFC that could be attributed to the presumed nonlinear nature of the underlying processes. Moreover, we found a characteristic spatial distribution of local connectivity dynamics, in that frontal and occipital regions showed stronger long-range correlation and higher degree of multifractality, whereas the highest values of entropy were found over the central and temporal regions. The revealed topology reflected well the underlying resting-state network organization of the brain. The presented results and the proposed analysis framework could improve our understanding on how resting-state brain activity is spatio-temporally organized and may provide potential biomarkers for future clinical research.

Functional Linear and Nonlinear Brain–Heart Interplay during Emotional Video Elicitation: A Maximum Information Coefficient Study

Brain and heart continuously interact through anatomical and biochemical connections. Although several brain regions are known to be involved in the autonomic control, the functional brain–heart interplay (BHI) during emotional processing is not fully characterized yet. To this aim, we investigate BHI during emotional elicitation in healthy subjects. The functional linear and nonlinear couplings are quantified using the maximum information coefficient calculated between time-varying electroencephalography (EEG) power spectra within the canonical bands (δ,θ,α,β and γ), and time-varying low-frequency and high-frequency powers from heartbeat dynamics. Experimental data were gathered from 30 healthy volunteers whose emotions were elicited through pleasant and unpleasant high-arousing videos. Results demonstrate that functional BHI increases during videos with respect to a resting state through EEG oscillations not including the γ band (>30 Hz). Functional linear coupling seems associated with a high-arousing positive elicitation, with preferred EEG oscillations in the θ band ([4,8) Hz) especially over the left-temporal and parietal cortices. Differential functional nonlinear coupling between emotional valence seems to mainly occur through EEG oscillations in the δ,θ,α bands and sympathovagal dynamics, as well as through δ,α,β oscillations and parasympathetic activity mainly over the right hemisphere. Functional BHI through δ and α oscillations over the prefrontal region seems primarily nonlinear. This study provides novel insights on synchronous heartbeat and cortical dynamics during emotional video elicitation, also suggesting that a nonlinear analysis is needed to fully characterize functional BHI.

Longitudinal EEG power in the first postnatal year differentiates autism outcomes

An aim of autism spectrum disorder (ASD) research is to identify early biomarkers that inform ASD pathophysiology and expedite detection. Brain oscillations captured in electroencephalography (EEG) are thought to be disrupted as core ASD pathophysiology. We leverage longitudinal EEG power measurements from 3 to 36 months of age in infants at low- and high-risk for ASD to test how and when power distinguishes ASD risk and diagnosis by age 3-years. Power trajectories across the first year, second year, or first three years postnatally were submitted to data-driven modeling to differentiate ASD outcomes. Power dynamics during the first postnatal year best differentiate ASD diagnoses. Delta and gamma frequency power trajectories consistently distinguish infants with ASD diagnoses from others. There is also a developmental shift across timescales towards including higher-frequency power to differentiate outcomes. These findings reveal the importance of developmental timing and trajectory in understanding pathophysiology and classifying ASD outcomes.

Brain oscillations may be disrupted in children with autism spectrum disorder. The authors performed a longitudinal study of electroencephalography recordings and found that EEG recordings from the first year after birth can distinguish healthy children from children with autism spectrum disorder.

Reduced frontal gamma power at 24 months is associated with better expressive language in toddlers at risk for autism

Frontal gamma power has been associated with early language development in typically developing toddlers, and gamma band abnormalities have been observed in individuals with autism spectrum disorder (ASD), as well as high-risk infant siblings (those having an older sibling with ASD), as early as 6 months of age. The current study investigated differences in baseline frontal gamma power and its association with language development in toddlers at high versus low familial risk for autism. Electroencephalography recordings as well as cognitive and behavioral assessments were acquired at 24 months as part of prospective, longitudinal study of infant siblings of children with and without autism. Diagnosis of autism was determined at 24-36 months, and data were analyzed across three outcome groups-low-risk without ASD (n = 43), high-risk without ASD (n = 42), and high-risk with ASD (n = 16). High-risk toddlers without ASD had reduced baseline frontal gamma power (30-50 Hz) compared to low-risk toddlers. Among high-risk toddlers increased frontal gamma was only marginally associated with ASD diagnosis (P = 0.06), but significantly associated with reduced expressive language ability (P = 0.007). No association between gamma power and language was present in the low-risk group. These findings suggest that differences in gamma oscillations in high-risk toddlers may represent compensatory mechanisms associated with improved developmental outcomes. Autism Res 2019, 12: 1211-1224. © 2019 International Society for Autism Research, Wiley Periodicals, Inc. LAY SUMMARY: This study looked at differences in neural activity in the gamma range and its association with language in toddlers with and without increased risk for ASD. At 2 years of age, gamma power was lower in high-risk toddlers without ASD compared to a low-risk comparison group. Among high-risk toddlers both with and without later ASD, reduced gamma power was also associated with better language outcomes, suggesting that gamma power may be a marker of language development in high-risk children.

Evaluation of Artifact Subspace Reconstruction for Automatic Artifact Components Removal in Multi-Channel EEG Recordings

OBJECTIVE

Artifact subspace reconstruction (ASR) is an automatic, online-capable, component-based method that can effectively remove transient or large-amplitude artifacts contaminating electroencephalographic (EEG) data. However, the effectiveness of ASR and the optimal choice of its parameter have not been systematically evaluated and reported, especially on actual EEG data.

METHODS

This paper systematically evaluates ASR on 20 EEG recordings taken during simulated driving experiments. Independent component analysis (ICA) and an independent component classifier are applied to separate artifacts from brain signals to quantitatively assess the effectiveness of the ASR.

RESULTS

ASR removes more eye and muscle components than brain components. Even though some eye and muscle components retain after ASR cleaning, the power of their temporal activities is reduced. Study results also showed that ASR cleaning improved the quality of a subsequent ICA decomposition.

CONCLUSIONS

Empirical results show that the optimal ASR parameter is between 20 and 30, balancing between removing non-brain signals and retaining brain activities.

SIGNIFICANCE

With an appropriate choice of parameter, ASR can be a powerful and automatic artifact removal approach for offline data analysis or online real-time EEG applications such as clinical monitoring and brain-computer interfaces.

Wavelet p-Leader Non-Gaussian Multiscale Expansions for EEG series: an Exploratory Study on Cold-Pressor Test

Brain dynamics recorded through electroencephalography (EEG) have been proven to be the output of a nonstationary and nonlinear system. Thus, multifractality of EEG series has been exploited as a useful tool for a neurophysiological characterization in health and disease. However, the role of EEG multifractality under peripheral stress is unknown. In this study, we propose to make use of a novel tool, the recently defined non-Gaussian multiscale analysis, to investigate brain dynamics in the range of 4-8Hz following a cold-pressor test versus a resting state. The method builds on the wavelet p-leader multifractal spectrum to quantify different types of departure from Gaussian and linear properties, and is compared here to standard linear descriptive indices. Results suggest that the proposed non-Gaussian multiscale indices were able to detect expected changes over the somatosensory and premotor cortices, over regions different from those detected by linear analyses. They further indicate that preferred responses for the contralateral somatosensory cortex occur at scales 2.5s and 5s. These findings contribute to the characterization of the so-called central autonomic network, linking dynamical changes at a peripheral and a central nervous system levels.

Association of Perceived Maternal Stress During the Perinatal Period With Electroencephalography Patterns in 2-Month-Old Infants

IMPORTANCE

Variation in child responses to adversity creates a clinical challenge to identify children most resilient or susceptible to later risk for disturbances in cognition and health. Advances in establishing scalable biomarkers can lead to early identification and mechanistic understanding of the association of early adversity with neurodevelopment.

OBJECTIVES

To examine whether maternal reports of stress are associated with patterns in resting electroencephalography at 2 months of age and whether unique electroencephalographic profiles associated with risk and resiliency factors can be identified.

DESIGN, SETTING, AND PARTICIPANTS

For this cohort study, a population-based sample of 113 mother-infant dyads was recruited from January 1, 2016, to March 1, 2018, during regularly scheduled pediatric visits before infants were 2 months of age from 2 primary care clinics in Boston, Massachusetts, and Los Angeles, California, that predominantly serve families from low-income backgrounds. Data are reported from a single time point, when infants were aged 2 months, of an ongoing cohort study longitudinally following the mother-infant dyads.

EXPOSURES

Maternal reported exposure to stressful life events and perceived stress.

MAIN OUTCOMES AND MEASURES

Spectral power (absolute and relative) in different frequency bands (Δ, θ, low and high α, β, and γ) from infant resting electroencephalography (EEG) and EEG profiles across frequency bands determined by latent profile analysis.

RESULTS

Of 113 enrolled infants, 70 (mean [SD] age, 2.42 [0.37] months; 35 girls [50%]) provided usable EEG data. In multivariable hierarchical linear regressions, maternal perceived stress was significantly and negatively associated with absolute β (β = -0.007; 95% CI, -0.01 to -0.001; semipartial r = -0.25) and γ power (β = -0.008; 95% CI, -0.01 to -0.002; semipartial r = -0.28). Maternal educational level was significantly and positively associated with power in high α, β, and γ bands after adjusting for covariates (high school: γ: β = 0.108; 95% CI, 0.014-0.203; semipartial r = -0.236; associate's degree or higher: high α: β = 0.133; 95% CI, 0.018-0.248; semipartial r = 0.241; β: β = 0.167; 95% CI, 0.055-0.279; semipartial r = 0.309; and γ: β = 0.183; 95% CI, 0.066-0.299; semipartial r = 0.323). Latent profile analysis identified 2 unique profiles for absolute and relative power. Maternal perceived stress (β = 0.13; 95% CI, 0.01-0.25; adjusted odds ratio [AOR], 1.14; 95% CI, 1.01-1.28) and maternal educational level (high school: β = 3.00; 95% CI, 0.35-5.65; AOR, 20.09; 95% CI, 1.42-283.16; associate's degree or higher: β = 4.12; 95% CI, 1.45-6.79; AOR, 61.56; 95% CI, 4.28-885.01) were each associated with unique profile membership.

CONCLUSIONS AND RELEVANCE

These findings suggest that unique contributions of caregiver stress and maternal educational level on infant neurodevelopment are detectable at 2 months; EEG might be a promising tool to identify infants most susceptible to parental stress and to reveal mechanisms by which neurodevelopment is associated with adversity. Additional studies validating subgroups across larger cohorts with different stressors and at different ages are required before use at the individual level in clinical settings.

Early Parenting Intervention and Adverse Family Environments Affect Neural Function in Middle Childhood

BACKGROUND

Growing work points to the negative impact of early adverse experiences on the developing brain. An outstanding question concerns the extent to which early intervention can normalize trajectories of brain development in at-risk children. We tested this within the context of a randomized clinical trial of an early parenting program, the Attachment and Biobehavioral Catch-up (ABC), delivered to parents and infants monitored for maltreatment by Child Protective Services.

METHODS

Families participated in the randomized clinical trial when children were 2.5 years of age or younger. Parenting and home adversity was measured at baseline. Children were followed longitudinally, and resting brain activity was measured electrophysiologically (n = 106) when children reached 8 years of age. Spectral power was quantified and compared across children assigned to the experimental intervention (ABC), a control intervention, and a low-risk comparison group (n = 76) recruited at the follow-up assessment.

RESULTS

Higher early home adversity was associated with electrophysiological profiles indicative of cortical delays/immaturity in middle childhood, based on relatively greater power in lower frequency bands (theta, 4-6 Hz, and low alpha, 6-9 Hz) and lower power in a higher frequency band (high alpha, 9-12 Hz). Children assigned to ABC showed relatively greater high-frequency power (beta, 12-20 Hz) than children assigned to the control intervention. Beta power in the ABC did not differ from that of the low-risk comparison group.

CONCLUSIONS

Maltreatment risk and home adversity can affect indicators of middle childhood brain maturation. Early parenting programs can support more normative patterns of neural function during middle childhood.

Neural correlates of facial emotion processing in infancy

In the present study we examined the neural correlates of facial emotion processing in the first year of life using ERP measures and cortical source analysis. EEG data were collected cross-sectionally from 5- (N = 49), 7- (N = 50), and 12-month-old (N = 51) infants while they were viewing images of angry, fearful, and happy faces. The N290 component was found to be larger in amplitude in response to fearful and happy than angry faces in all posterior clusters and showed largest response to fear than the other two emotions only over the right occipital area. The P400 and Nc components were found to be larger in amplitude in response to angry than happy and fearful faces over central and frontal scalp. Cortical source analysis of the N290 component revealed greater cortical activation in the right fusiform face area in response to fearful faces. This effect started to emerge at 5 months and became well established at 7 months, but it disappeared at 12 months. The P400 and Nc components were primarily localized to the PCC/Precuneus where heightened responses to angry faces were observed. The current results suggest the detection of a fearful face in infants' brain can happen shortly (~200-290 ms) after the stimulus onset, and this process may rely on the face network and develop substantially between 5 to 7 months of age. The current findings also suggest the differential processing of angry faces occurred later in the P400/Nc time window, which recruits the PCC/Precuneus and is associated with the allocation of infants' attention.

BEAPP: The Batch Electroencephalography Automated Processing Platform

Electroencephalography (EEG) offers information about brain function relevant to a variety of neurologic and neuropsychiatric disorders. EEG contains complex, high-temporal-resolution information, and computational assessment maximizes our potential to glean insight from this information. Here we present the Batch EEG Automated Processing Platform (BEAPP), an automated, flexible EEG processing platform incorporating freely available software tools for batch processing of multiple EEG files across multiple processing steps. BEAPP does not prescribe a specified EEG processing pipeline; instead, it allows users to choose from a menu of options for EEG processing, including steps to manage EEG files collected across multiple acquisition setups (e.g., for multisite studies), minimize artifact, segment continuous and/or event-related EEG, and perform basic analyses. Overall, BEAPP aims to streamline batch EEG processing, improve accessibility to computational EEG assessment, and increase reproducibility of results.

Computational Testing for Automated Preprocessing 2: Practical Demonstration of a System for Scientific Data-Processing Workflow Management for High-Volume EEG

Existing tools for the preprocessing of EEG data provide a large choice of methods to suitably prepare and analyse a given dataset. Yet it remains a challenge for the average user to integrate methods for batch processing of the increasingly large datasets of modern research, and compare methods to choose an optimal approach across the many possible parameter configurations. Additionally, many tools still require a high degree of manual decision making for, e.g., the classification of artifacts in channels, epochs or segments. This introduces extra subjectivity, is slow, and is not reproducible. Batching and well-designed automation can help to regularize EEG preprocessing, and thus reduce human effort, subjectivity, and consequent error. The Computational Testing for Automated Preprocessing (CTAP) toolbox facilitates: (i) batch processing that is easy for experts and novices alike; (ii) testing and comparison of preprocessing methods. Here we demonstrate the application of CTAP to high-resolution EEG data in three modes of use. First, a linear processing pipeline with mostly default parameters illustrates ease-of-use for naive users. Second, a branching pipeline illustrates CTAP's support for comparison of competing methods. Third, a pipeline with built-in parameter-sweeping illustrates CTAP's capability to support data-driven method parameterization. CTAP extends the existing functions and data structure from the well-known EEGLAB toolbox, based on Matlab, and produces extensive quality control outputs. CTAP is available under MIT open-source licence from https://github.com/bwrc/ctap.