Unsupervised learning of brain state dynamics during emotion imagination using high-density EEG.

This study applies adaptive mixture independent component analysis (AMICA) to learn a set of ICA models, each optimized by fitting a distributional model for each identified component process while maximizing component process independence within some subsets of time points of a multi-channel EEG dataset. Here, we applied 20-model AMICA decomposition to long-duration (1-2 h), high-density (128-channel) EEG data recorded while participants used guided imagination to imagine situations stimulating the experience of 15 specified emotions. These decompositions tended to return models identifying spatiotemporal EEG patterns or states within single emotion imagination periods. Model probability transitions reflected time-courses of EEG dynamics during emotion imagination, which varied across emotions. Transitions between models accounting for imagined "grief" and "happiness" were more abrupt and better aligned with participant reports, while transitions for imagined "contentment" extended into adjoining "relaxation" periods. The spatial distributions of brain-localizable independent component processes (ICs) were more similar within participants (across emotions) than emotions (across participants). Across participants, brain regions with differences in IC spatial distributions (i.e., dipole density) between emotion imagination versus relaxation were identified in or near the left rostrolateral prefrontal, posterior cingulate cortex, right insula, bilateral sensorimotor, premotor, and associative visual cortex. No difference in dipole density was found between positive versus negative emotions. AMICA models of changes in high-density EEG dynamics may allow data-driven insights into brain dynamics during emotional experience, possibly enabling the improved performance of EEG-based emotion decoding and advancing our understanding of emotion.

Modeling brain dynamic state changes with adaptive mixture independent component analysis.

There is a growing interest in neuroscience in assessing the continuous, endogenous, and nonstationary dynamics of brain network activity supporting the fluidity of human cognition and behavior. This non-stationarity may involve ever-changing formation and dissolution of active cortical sources and brain networks. However, unsupervised approaches to identify and model these changes in brain dynamics as continuous transitions between quasi-stable brain states using unlabeled, noninvasive recordings of brain activity have been limited. This study explores the use of adaptive mixture independent component analysis (AMICA) to model multichannel electroencephalographic (EEG) data with a set of ICA models, each of which decomposes an adaptively learned portion of the data into statistically independent sources. We first show that AMICA can segment simulated quasi-stationary EEG data and accurately identify ground-truth sources and source model transitions. Next, we demonstrate that AMICA decomposition, applied to 6-13 channel scalp recordings from the CAP Sleep Database, can characterize sleep stage dynamics, allowing 75% accuracy in identifying transitions between six sleep stages without use of EEG power spectra. Finally, applied to 30-channel data from subjects in a driving simulator, AMICA identifies models that account for EEG during faster and slower response to driving challenges, respectively. We show changes in relative probabilities of these models allow effective prediction of subject response speed and moment-by-moment characterization of state changes within single trials. AMICA thus provides a generic unsupervised approach to identifying and modeling changes in EEG dynamics. Applied to continuous, unlabeled multichannel data, AMICA may likely be used to detect and study any changes in cognitive states.

Using a standalone ear-EEG device for focal-onset seizure detection.

BACKGROUND: Seizure detection is challenging outside the clinical environment due to the lack of comfortable, reliable, and practical long-term neurophysiological monitoring devices. We developed a novel, discreet, unobstructive in-ear sensing system that enables long-term electroencephalography (EEG) recording. This is the first study we are aware of that systematically compares the seizure detection utility of in-ear EEG with that of simultaneously recorded intracranial EEG. In addition, we present a similar comparison between simultaneously recorded in-ear EEG and scalp EEG. METHODS: In this foundational research, we conducted a clinical feasibility study and validated the ability of the ear-EEG system to capture focal-onset seizures against 1255 hrs of simultaneous ear-EEG data along with scalp or intracranial EEG in 20 patients with refractory focal epilepsy (11 with scalp EEG, 8 with intracranial EEG, and 1 with both). RESULTS: In a blinded, independent review of the ear-EEG signals, two epileptologists were able to detect 86.4% of the seizures that were subsequently identified using the clinical gold standard EEG modalities, with a false detection rate of 0.1 per day, well below what has been reported for ambulatory monitoring. The few seizures not detected on the ear-EEG signals emanated from deep within the mesial temporal lobe or extra-temporally and remained very focal, without significant propagation. Following multiple sessions of recording for a median continuous wear time of 13 hrs, patients reported a high degree of tolerance for the device, with only minor adverse events reported by the scalp EEG cohort. CONCLUSIONS: These preliminary results demonstrate the potential of using ear-EEG to enable routine collection of complementary, prolonged, and remote neurophysiological evidence, which may permit real-time detection of paroxysmal events such as seizures and epileptiform discharges. This study suggests that the ear-EEG device may assist clinicians in making an epilepsy diagnosis, assessing treatment efficacy, and optimizing medication titration.

Stimulus Design for Visual Evoked Potential Based Brain-Computer Interfaces.

Visual stimuli design plays an important role in brain-computer interfaces (BCIs) based on visual evoked potentials (VEPs). Variations in stimulus parameters have been shown to affect both decoding accuracy and subjective perception experience, implying the need for a trade-off in design. In this study, we comprehensively and systematically compared various combinations of amplitude contrast and spectral content parameters in the stimulus design to quantify their impact on decoding performance and subject comfort. Specifically, three parameters were investigated: 1) contrast level, 2) temporal pattern (periodic steady-state or pseudo-random code-modulated), and 3) frequency range. We collected electroencephalogram (EEG) data and subjective perception ratings from ten subjects and evaluated the decoding accuracy and subject comfort rating for different combinations of the stimulus parameters. Our results indicate that while high-frequency steady-state VEP (SSVEP) stimuli were rated the most comfortable, they also had the lowest decoding accuracy. Conversely, low-frequency SSVEP stimuli were rated the least comfortable but had the highest decoding accuracy. Standard and high-frequency M-sequence code-modulated VEPs (c-VEPs) produced intermediates between the two. We observed a consistent trade-off relationship between decoding accuracy and subjective comfort level across all parameters. Based on our findings, we offer c-VEP as a preferable stimulus for achieving reliable decoding accuracy while maintaining a reasonable level of comfortability.

Combining brain-computer interface and virtual reality for rehabilitation in neurological diseases: A narrative review.

BACKGROUND: The traditional rehabilitation for neurological diseases lacks the active participation of patients, its process is monotonous and tedious, and the effects need to be improved. Therefore, a new type of rehabilitation technology with more active participation combining brain-computer interface (BCI) with virtual reality (VR) has developed rapidly in recent years and has been used in rehabilitation in neurological diseases. OBJECTIVES: This narrative review analyzed and characterized the development and application of the new training system (BCI-VR) in rehabilitation of neurological diseases from the perspective of the BCI paradigm, to provide a pathway for future research in this field. METHODS: The review involved a search of the Web of Science-Science Citation Index/Social Sciences Citation Index and the China National Knowledge Infrastructure databases; 39 papers were selected. Advantages and challenges of BCI-VR - based neurological rehabilitation were analyzed in detail. RESULTS: Most BCI-VR studies included could be classified by 3 major BCI paradigms: motor imagery, P300, and steady-state visual-evoked potential. Integrating VR scenes into BCI systems could effectively promote the recovery process from nervous system injuries as compared with traditional methods. CONCLUSION: As compared with rehabilitation based on traditional BCI, rehabilitation based on BCI-VR can provide better feedback information for patients and promote the recovery of brain function. By solving the challenges and continual development, the BCI-VR system can be broadly applied to the clinical treatment of various neurological diseases.

Abnormal sleep physiology in children with 15q11.2-13.1 duplication (Dup15q) syndrome.

BACKGROUND: Sleep disturbances in autism spectrum disorder (ASD) represent a common and vexing comorbidity. Clinical heterogeneity amongst these warrants studies of the mechanisms associated with specific genetic etiologies. Duplications of 15q11.2-13.1 (Dup15q syndrome) are highly penetrant for neurodevelopmental disorders (NDDs) such as intellectual disability and ASD, as well as sleep disturbances. Genes in the 15q region, particularly UBE3A and a cluster of GABAA receptor genes, are critical for neural development, synaptic protein synthesis and degradation, and inhibitory neurotransmission. During awake electroencephalography (EEG), children with Dup15q syndrome demonstrate increased beta band oscillations (12-30 Hz) that likely reflect aberrant GABAergic neurotransmission. Healthy sleep rhythms, necessary for robust cognitive development, are also highly dependent on GABAergic neurotransmission. We therefore hypothesized that sleep physiology would be abnormal in children with Dup15q syndrome. METHODS: To test the hypothesis that elevated beta oscillations persist in sleep in Dup15q syndrome and that NREM sleep rhythms would be disrupted, we computed: (1) beta power, (2) spindle density, and (3) percentage of slow-wave sleep (SWS) in overnight sleep EEG recordings from a cohort of children with Dup15q syndrome (n = 15) and compared them to age-matched neurotypical children (n = 12). RESULTS: Children with Dup15q syndrome showed abnormal sleep physiology with elevated beta power, reduced spindle density, and reduced or absent SWS compared to age-matched neurotypical controls. LIMITATIONS: This study relied on clinical EEG where sleep staging was not available. However, considering that clinical polysomnograms are challenging to collect in this population, the ability to quantify these biomarkers on clinical EEG-routinely ordered for epilepsy monitoring-opens the door for larger-scale studies. While comparable to other human studies in rare genetic disorders, a larger sample would allow for examination of the role of seizure severity, medications, and developmental age that may impact sleep physiology. CONCLUSIONS: We have identified three quantitative EEG biomarkers of sleep disruption in Dup15q syndrome, a genetic condition highly penetrant for ASD. Insights from this study not only promote a greater mechanistic understanding of the pathophysiology defining Dup15q syndrome, but also lay the foundation for studies that investigate the association between sleep and cognition. Abnormal sleep physiology may undermine healthy cognitive development and may serve as a quantifiable and modifiable target for behavioral and pharmacological interventions.

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.

Modulation of frontal gamma oscillations improves working memory in schizophrenia.

Schizophrenia is a debilitating mental disorder that is associated with cognitive deficits. Impairments in cognition occur early in the course of illness and are associated with poor functional outcome, but have been difficult to treat with conventional treatments. Recent studies have implicated abnormal neural network dynamics and impaired connectivity in frontal brain regions as possible causes of cognitive deficits. For example, high-frequency, dorsal-lateral prefrontal oscillatory activity in the gamma range (30-50 Hz) is associated with impaired working memory in individuals with schizophrenia.In light of these findings, it may be possible to use EEG neurofeedback (EEG-NFB) to train individuals with schizophrenia to enhance frontal gamma activity to improve working memory and cognition. In a single-group, proof-of-concept study, 31 individuals with schizophrenia received 12 weeks of twice weekly EEG-NFB to enhance frontal gamma band response. EEG-NFB was well-tolerated, associated with increased gamma training threshold, and significant increases in frontal gamma power during an n-back working memory task. Additionally, EEG-NFB was associated with significant improvements in n-back performance and working memory, speed of processing, and reasoning and problem solving on neuropsychological tests. Change in gamma power was associated with change in cognition. Significant improvements in psychiatric symptoms were also found. These encouraging findings suggest EEG-NFB targeting frontal gamma activity may provide a novel effective approach to cognitive remediation in schizophrenia, although placebo-controlled trials are needed to assess the effects of non-treatment related factors.

Estimating coupling strength between multivariate neural series with multivariate permutation conditional mutual information.

Recently, coupling between groups of neurons or different brain regions has been widely studied to provide insights into underlying mechanisms of brain functions. To comprehensively understand the effect of such coupling, it is necessary to accurately extract the coupling strength information among multivariate neural signals from the whole brain. This study proposed a new method named multivariate permutation conditional mutual information (MPCMI) to quantitatively estimate the coupling strength of multivariate neural signals (MNS). The performance of the MPCMI method was validated on the simulated MNS generated by multi-channel neural mass model (MNMM). The coupling strength feature of simulated MNS extracted by MPCMI showed better performance compared with standard methods, such as permutation conditional mutual information (PCMI), multivariate Granger causality (MVGC), and Granger causality analysis (GCA). Furthermore, the MPCMI was applied to estimate the coupling strengths of two-channel resting-state electroencephalographic (rsEEG) signals from different brain regions of 19 patients with amnestic mild cognitive impairment (aMCI) with type 2 diabetes mellitus (T2DM) and 20 normal control (NC) with T2DM in Alpha1 and Alpha2 frequency bands. Empirical results showed that the MPCMI could effectively extract the coupling strength features that were significantly different between the aMCI and the NC. Hence, the proposed MPCMI method could be an effective estimate of coupling strengths of MNS, and might be a viable biomarker for clinical applications.

Evaluation of Artifact Subspace Reconstruction for Automatic EEG Artifact Removal.

One of the greatest challenges that hinder the decoding and application of electroencephalography (EEG) is that EEG recordings almost always contain artifacts - non-brain signals. Among existing automatic artifact-removal methods, artifact subspace reconstruction (ASR) is an online and realtime capable, component-based method that can effectively remove transient or large-amplitude artifacts. However, the effectiveness of ASR and the optimal choice of its parameter have not been evaluated and reported, especially on real EEG data. This study systematically validates ASR on ten EEG recordings in a simulated driving experiment. Independent component analysis (ICA) is applied to separate artifacts from brain signals to allow a quantitative assessment of ASR's effectiveness in removing various types of artifacts and preserving brain activities. Empirical results show that the optimal ASR parameter is between 10 and 100, which is small enough to remove activities from artifacts and eye-related components and large enough to retain signals from brain-related components. With the appropriate choice of the 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.

Online Automatic Artifact Rejection using the Real-time EEG Source-mapping Toolbox (REST).

Non-brain contributions to electroencephalographic (EEG) signals, often referred to as artifacts, can hamper the analysis of scalp EEG recordings. This is especially true when artifacts have large amplitudes (e.g., movement artifacts), or occur continuously (like eye-movement artifacts). Offline automated pipelines can detect and reduce artifact in EEG data, but no good solution exists for online processing of EEG data in near real time. Here, we propose the combined use of online artifact subspace reconstruction (ASR) to remove large amplitude transients, and online recursive independent component analysis (ORICA) combined with an independent component (IC) classifier to compute, classify, and remove artifact ICs. We demonstrate the efficacy of the proposed pipeline using 2 EEG recordings containing series of (1) movement and muscle artifacts, and (2) cued blinks and saccades. This pipeline is freely available in the Real-time EEG Sourcemapping Toolbox (REST) for MATLAB (The Mathworks, Inc.).

Monitoring alert and drowsy states by modeling EEG source nonstationarity.

OBJECTIVE: As a human brain performs various cognitive functions within ever-changing environments, states of the brain characterized by recorded brain activities such as electroencephalogram (EEG) are inevitably nonstationary. The challenges of analyzing the nonstationary EEG signals include finding neurocognitive sources that underlie different brain states and using EEG data to quantitatively assess the state changes. APPROACH: This study hypothesizes that brain activities under different states, e.g. levels of alertness, can be modeled as distinct compositions of statistically independent sources using independent component analysis (ICA). This study presents a framework to quantitatively assess the EEG source nonstationarity and estimate levels of alertness. The framework was tested against EEG data collected from 10 subjects performing a sustained-attention task in a driving simulator. MAIN RESULTS: Empirical results illustrate that EEG signals under alert versus drowsy states, indexed by reaction speeds to driving challenges, can be characterized by distinct ICA models. By quantifying the goodness-of-fit of each ICA model to the EEG data using the model deviation index (MDI), we found that MDIs were significantly correlated with the reaction speeds (r = -0.390 with alertness models and r = 0.449 with drowsiness models) and the opposite correlations indicated that the two models accounted for sources in the alert and drowsy states, respectively. Based on the observed source nonstationarity, this study also proposes an online framework using a subject-specific ICA model trained with an initial (alert) state to track the level of alertness. For classification of alert against drowsy states, the proposed online framework achieved an averaged area-under-curve of 0.745 and compared favorably with a classic power-based approach. SIGNIFICANCE: This ICA-based framework provides a new way to study changes of brain states and can be applied to monitoring cognitive or mental states of human operators in attention-critical settings or in passive brain-computer interfaces.

Real-Time Adaptive EEG Source Separation Using Online Recursive Independent Component Analysis.

Independent component analysis (ICA) has been widely applied to electroencephalographic (EEG) biosignal processing and brain-computer interfaces. The practical use of ICA, however, is limited by its computational complexity, data requirements for convergence, and assumption of data stationarity, especially for high-density data. Here we study and validate an optimized online recursive ICA algorithm (ORICA) with online recursive least squares (RLS) whitening for blind source separation of high-density EEG data, which offers instantaneous incremental convergence upon presentation of new data. Empirical results of this study demonstrate the algorithm's: 1) suitability for accurate and efficient source identification in high-density (64-channel) realistically-simulated EEG data; 2) capability to detect and adapt to nonstationarity in 64-ch simulated EEG data; and 3) utility for rapidly extracting principal brain and artifact sources in real 61-channel EEG data recorded by a dry and wearable EEG system in a cognitive experiment. ORICA was implemented as functions in BCILAB and EEGLAB and was integrated in an open-source Real-time EEG Source-mapping Toolbox (REST), supporting applications in ICA-based online artifact rejection, feature extraction for real-time biosignal monitoring in clinical environments, and adaptable classifications in brain-computer interfaces.

Tracking non-stationary EEG sources using adaptive online recursive independent component analysis.

Electroencephalographic (EEG) source-level analyses such as independent component analysis (ICA) have uncovered features related to human cognitive functions or artifactual activities. Among these methods, Online Recursive ICA (ORICA) has been shown to achieve fast convergence in decomposing high-density EEG data for real-time applications. However, its adaptation performance has not been fully explored due to the difficulty in choosing an appropriate forgetting factor: the weight applied to new data in a recursive update which determines the trade-off between the adaptation capability and convergence quality. This study proposes an adaptive forgetting factor for ORICA (adaptive ORICA) to learn and adapt to non-stationarity in the EEG data. Using a realistically simulated non-stationary EEG dataset, we empirically show adaptive forgetting factors outperform other commonly-used non-adaptive rules when underlying source dynamics are changing. Standard offline ICA can only extract a subset of the changing sources while adaptive ORICA can recover all. Applied to actual EEG data recorded from a task-switching experiments, adaptive ORICA can learn and re-learn the task-related components as they change. With an adaptive forgetting factor, adaptive ORICA can track non-stationary EEG sources, opening many new online applications in brain-computer interfaces and in monitoring of brain dynamics.

Real-time EEG Source-mapping Toolbox (REST): Online ICA and source localization.

The Electroencephalogram (EEG) is a noninvasive functional brain activity recording method that shows promise for becoming a 3-D cortical imaging modality with high temporal resolution. Currently, most of the tools developed for EEG analysis focus mainly on offline processing. This study introduces and demonstrates the Real-time EEG Source-mapping Toolbox (REST), an extension to the widely distributed EEGLAB environment. REST allows blind source separation of EEG data in real-time using Online Recursive Independent Component Analysis (ORICA), plus near real-time localization of separated sources. Two source localization methods are available to fit equivalent current dipoles or estimate spatial source distributions of selected sources. Selected measures of raw EEG data or component activations (e.g. time series of the data, spectral changes over time, equivalent current dipoles, etc.) can be visualized in near real-time. Finally, this study demonstrates the accuracy and functionality of REST with data from two experiments and discusses some relevant applications.

Online recursive independent component analysis for real-time source separation of high-density EEG.

Online Independent Component Analysis (ICA) algorithms have recently seen increasing development and application across a range of fields, including communications, biosignal processing, and brain-computer interfaces. However, prior work in this domain has primarily focused on algorithmic proofs of convergence, with application limited to small `toy' examples or to relatively low channel density EEG datasets. Furthermore, there is limited availability of computationally efficient online ICA implementations, suitable for real-time application. This study describes an optimized online recursive ICA algorithm (ORICA), with online recursive least squares (RLS) whitening, for blind source separation of high-density EEG data. It is implemented as an online-capable plugin within the open-source BCILAB (EEGLAB) framework. We further derive and evaluate a block-update modification to the ORICA learning rule. We demonstrate the algorithm's suitability for accurate and efficient source identification in high density (64-channel) realistically-simulated EEG data, as well as real 61-channel EEG data recorded by a dry and wearable EEG system in a cognitive experiment.

High-resolution structural and functional assessments of cerebral microvasculature using 3D Gas ΔR2*-mMRA.

The ability to evaluate the cerebral microvascular structure and function is crucial for investigating pathological processes in brain disorders. Previous angiographic methods based on blood oxygen level-dependent (BOLD) contrast offer appropriate visualization of the cerebral vasculature, but these methods remain to be optimized in order to extract more comprehensive information. This study aimed to integrate the advantages of BOLD MRI in both structural and functional vascular assessments. The BOLD contrast was manipulated by a carbogen challenge, and signal changes in gradient-echo images were computed to generate ΔR2* maps. Simultaneously, a functional index representing the regional cerebral blood volume was derived by normalizing the ΔR2* values of a given region to those of vein-filled voxels of the sinus. This method is named 3D gas ΔR2*-mMRA (microscopic MRA). The advantages of using 3D gas ΔR2*-mMRA to observe the microvasculature include the ability to distinguish air-tissue interfaces, a high vessel-to-tissue contrast, and not being affected by damage to the blood-brain barrier. A stroke model was used to demonstrate the ability of 3D gas ΔR2*-mMRA to provide information about poststroke revascularization at 3 days after reperfusion. However, this technique has some limitations that cannot be overcome and hence should be considered when it is applied, such as magnifying vessel sizes and predominantly revealing venous vessels.