Computational Models in Electroencephalography

Decoding Imagined Speech from EEG Data: A Hybrid Deep Learning Approach to Capturing Spatial and Temporal Features

Neuroimaging is revolutionizing our ability to investigate the brain's structural and functional properties, enabling us to visualize brain activity during diverse mental processes and actions. One of the most widely used neuroimaging techniques is electroencephalography (EEG), which records electrical activity from the brain using electrodes positioned on the scalp. EEG signals capture both spatial (brain region) and temporal (time-based) data. While a high temporal resolution is achievable with EEG, spatial resolution is comparatively limited. Consequently, capturing both spatial and temporal information from EEG data to recognize mental activities remains challenging. In this paper, we represent spatial and temporal information obtained from EEG signals by transforming EEG data into sequential topographic brain maps. We then apply hybrid deep learning models to capture the spatiotemporal features of the EEG topographic images and classify imagined English words. The hybrid framework utilizes a sequential combination of three-dimensional convolutional neural networks (3DCNNs) and recurrent neural networks (RNNs). The experimental results reveal the effectiveness of the proposed approach, achieving an average accuracy of 77.8% in identifying imagined English speech.

Filtered Point Processes Tractably Capture Rhythmic And Broadband Power Spectral Structure in Neural Electrophysiological Recordings

Neural electrophysiological recordings arise from interacting rhythmic (oscillatory) and broadband (aperiodic) biological subprocesses. Both rhythmic and broadband processes contribute to the neural power spectrum, which decomposes the variance of a neural recording across frequencies. Although an extensive body of literature has successfully studied rhythms in various diseases and brain states, researchers only recently have systematically studied the characteristics of broadband effects in the power spectrum. Broadband effects can generally be categorized as 1) shifts in power across all frequencies, which correlate with changes in local firing rates and 2) changes in the overall shape of the power spectrum, such as the spectral slope or power law exponent. Shape changes are evident in various conditions and brain states, influenced by factors such as excitation to inhibition balance, age, and various diseases. It is increasingly recognized that broadband and rhythmic effects can interact on a sub-second timescale. For example, broadband power is time-locked to the phase of <1 Hz rhythms in propofol induced unconsciousness. Modeling tools that explicitly deal with both rhythmic and broadband contributors to the power spectrum and that capture their interactions are essential to help improve the interpretability of power spectral effects. Here, we introduce a tractable stochastic forward modeling framework designed to capture both narrowband and broadband spectral effects when prior knowledge or theory about the primary biophysical processes involved is available. Population-level neural recordings are modeled as the sum of filtered point processes (FPPs), each representing the contribution of a different biophysical process such as action potentials or postsynaptic potentials of different types. Our approach builds on prior neuroscience FPP work by allowing multiple interacting processes and time-varying firing rates and by deriving theoretical power spectra and cross-spectra. We demonstrate several properties of the models, including that they divide the power spectrum into frequency ranges dominated by rhythmic and broadband effects, and that they can capture spectral effects across multiple timescales, including sub-second cross-frequency coupling. The framework can be used to interpret empirically observed power spectra and cross-frequency coupling effects in biophysical terms, which bridges the gap between theoretical models and experimental results.

DELirium treatment with Transcranial Electrical Stimulation (DELTES): study protocol for a multicentre, randomised, double-blind, sham-controlled trial

INTRODUCTION

Delirium, a clinical manifestation of acute encephalopathy, is associated with extended hospitalisation, long-term cognitive dysfunction, increased mortality and high healthcare costs. Despite intensive research, there is still no targeted treatment. Delirium is characterised by electroencephalography (EEG) slowing, increased relative delta power and decreased functional connectivity. Recent studies suggest that transcranial alternating current stimulation (tACS) can entrain EEG activity, strengthen connectivity and improve cognitive functioning. Hence, tACS offers a potential treatment for augmenting EEG activity and reducing the duration of delirium. This study aims to evaluate the feasibility and assess the efficacy of tACS in reducing relative delta power.

METHODS AND ANALYSIS

A randomised, double-blind, sham-controlled trial will be conducted across three medical centres in the Netherlands. The study comprises two phases: a pilot phase (n=30) and a main study phase (n=129). Participants are patients aged 50 years and older who are diagnosed with delirium using the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision criteria (DSM-5-TR), that persists despite treatment of underlying causes. During the pilot phase, participants will be randomised (1:1) to receive either standardised (10 Hz) tACS or sham tACS. In the main study phase, participants will be randomised to standardised tACS, sham tACS or personalised tACS, in which tACS settings are tailored to the participant. All participants will undergo daily 30 min of (sham) stimulation for up to 14 days or until delirium resolution or hospital discharge. Sixty-four-channel resting-state EEG will be recorded pre- and post the first tACS session, and following the final tACS session. Daily delirium assessments will be acquired using the Intensive Care Delirium Screening Checklist and Delirium Observation Screening Scale. The pilot phase will assess the percentage of completed tACS sessions and increased care requirements post-tACS. The primary outcome variable is change in relative delta EEG power. Secondary outcomes include (1) delirium duration and severity, (2) quantitative EEG measurements, (3) length of hospital stay, (4) cognitive functioning at 3 months post-tACS and (5) tACS treatment burden. Study recruitment started in April 2024 and is ongoing.

ETHICS AND DISSEMINATION

The study has been approved by the Medical Ethics Committee of the Utrecht University Medical Center and the Institutional Review Boards of all participating centres. Trial results will be disseminated via peer-reviewed publications and conference presentations.

TRIAL REGISTRATION NUMBER

NCT06285721.

A tutorial on fitting joint models of M/EEG and behavior to understand cognition

We present motivation and practical steps necessary to find parameter estimates of joint models of behavior and neural electrophysiological data. This tutorial is written for researchers wishing to build joint models of human behavior and scalp and intracranial electroencephalographic (EEG) or magnetoencephalographic (MEG) data, and more specifically those researchers who seek to understand human cognition. Although these techniques could easily be applied to animal models, the focus of this tutorial is on human participants. Joint modeling of M/EEG and behavior requires some knowledge of existing computational and cognitive theories, M/EEG artifact correction, M/EEG analysis techniques, cognitive modeling, and programming for statistical modeling implementation. This paper seeks to give an introduction to these techniques as they apply to estimating parameters from neurocognitive models of M/EEG and human behavior, and to evaluate model results and compare models. Due to our research and knowledge on the subject matter, our examples in this paper will focus on testing specific hypotheses in human decision-making theory. However, most of the motivation and discussion of this paper applies across many modeling procedures and applications. We provide Python (and linked R) code examples in the tutorial and appendix. Readers are encouraged to try the exercises at the end of the document.

Investigation of the acute impact of rosemary consumption on brain activity in healthy volunteers

OBJECTIVES

Rosmarinus officinalis L. (rosemary) is a fragrant plant of the mint family, broadly known as a nourishment flavoring agent; it is additionally utilized in conventional people cures for its anti-inflammatory, diuretic, and antibacterial properties. Intense cognitive impacts from devouring plant-based flavonoids can be measured with electroencephalography (EEG), which records unconstrained brain movement. Brain activity can be evaluated amid independent states or whereas performing attentional assignments. This study aimed to determine the impact of rosemary consumption on cognitive consequences.

METHODS

Twenty volunteers took part in the study. EEG was taken for each volunteer twice, before drinking rosemary extract and around one hour after drinking it. EEG information was recorded with a Micromed recording framework inspecting rate of 512 Hz. EEG signals were prepared to be utilized in EEGLAB, an open-source toolbox within the MATLAB environment. The information obtained after the EEG recording was compared with the preliminary EEG information.

RESULTS

The signal's power spectral density in theta, delta, and beta frequency bands modestly increased in males and females. Even though there was a significant increase in power at the alpha frequency band in both sexes, this increment was not specific channel-wise.

DISCUSSION

The obtained data are consistent with the expected results and similar studies conducted, suggesting that the consumption of rosemary is beneficial for cognitive function in the short term. It is anticipated that forthcoming long-term studies will support the existing data.

Simultaneous oscillatory encoding of "hot" and "cold" information during social interactions in the monkey medial prefrontal cortex

Social interactions in primates require social cognition abilities such as anticipating the partner's future choices as well as pure cognitive skills involving processing task-relevant information. The medial prefrontal cortex (mPFC) has been implicated in these cognitive processes. Here, we investigated the neural oscillations underlying the complex social behaviors involving the interplay of social roles (Actor vs. Observer) and interaction types (whether working with a "Good" or "Bad" partner). We found opposite power modulations of the beta and gamma bands by social roles, indicating dedicated processing for task-related information. Concurrently, the interaction type was conveyed by lower frequencies, which are commonly associated with neural circuits linked to performance and reward monitoring. Thus, the mPFC exhibits parallel coding of both "cold" processes (purely cognitive) and "hot" processes (reward and social-related). This allocation of neural resources gives the mPFC a key neural node, flexibly integrating multiple sources of information during social interactions.

Bi-sinusoidal light stimulation reveals an enhanced response power and reduced phase coherence at the visual cortex in migraine

Introduction

Migraine is associated with enhanced visual sensitivity during and outside attacks. Processing of visual information is a highly non-linear process involving complex interactions across (sub)cortical networks. In this exploratory study, we combined electroencephalography with bi-sinusoidal light stimulation to assess non-linear features of visual processing in participants with migraine.

Methods

Twenty participants with migraine (10 with aura, 10 without aura) and ten non-headache controls were measured (outside attacks). Participants received bi-sinusoidal 13 + 23 Hz red light visual stimulation. Electroencephalography spectral power and multi-spectral phase coherence were compared between groups at the driving stimulation frequencies together with multiples and combinations of these frequencies (harmonic and intermodulation frequencies) caused by non-linearities.

Results

Only at the driving frequency of 13 Hz higher spectral power was found in migraine with aura participants compared with those with migraine without aura and controls. Differences in phase coherence were present for 2nd, 4th, and 5th-order non-linearities in those with migraine (migraine with and without aura) compared with controls. Bi-sinusoidal light stimulation revealed evident non-linearities in the brain's electroencephalography response up to the 5th order with reduced phase coherence for higher order interactions in interictal participants with migraine.

Discussion

Insight into interictal non-linear visual processing may help understand brain dynamics underlying migraine attack susceptibility. Future research is needed to determine the clinical value of the results.

Increased noise relates to abnormal excitation-inhibition balance in schizophrenia: a combined empirical and computational study

Electroencephalography studies link sensory processing issues in schizophrenia to increased noise level-noise here is background spontaneous activity-as measured by the signal-to-noise ratio. The mechanism, however, of such increased noise is unknown. We investigate if this relates to changes in cortical excitation-inhibition balance, which has been observed to be atypical in schizophrenia, by combining electroencephalography and computational modeling. Our electroencephalography task results, for which the local field potentials can be used as a proxy, show lower signal-to-noise ratio due to higher noise in schizophrenia. Both electroencephalography rest and task states exhibit higher levels of excitation in the functional excitation-inhibition (as a proxy of excitation-inhibition balance). This suggests a relationship between increased noise and atypical excitation in schizophrenia, which was addressed by using computational modeling. A Leaky Integrate-and-Fire model was used to simulate the effects of varying degrees of noise on excitation-inhibition balance, local field potential, NMDA current, and . Results show a noise-related increase in the local field potential, excitation in excitation-inhibition balance, pyramidal NMDA current, and spike rate. Mutual information and mediation analysis were used to explore a cross-level relationship, showing that the cortical local field potential plays a key role in transferring the effect of noise to the cellular population level of NMDA.

Investigation of frequency components embedded in EEG recordings underlying neuronal mechanism of cognitive control and attentional functions

Attentional cognitive control regulates the perception to enhance human behaviour. The current study examines the atltentional mechanisms in terms of time and frequency of EEG signals. The cognitive load is higher for processing local attentional stimulus, thereby demanding higher response time (RT) with low response accuracy (RA). On the other hand, the global attentional mechanisms broadly promote the perception while demanding a low cognitive load with faster RT and high RA. Attentional mechanisms refer to perceptual systems that afford and allocate the adaptive behaviours for prioritizing the processing of relevant stimuli based on the local and global features. The early sensory component of C1, which was associated with the local attentional mechanism, showed higher amplitudes than the global attentional mechanisms in parieto-occipital regions. Further, the local attentional mechanisms were also sustained in N2 and P3 components increasing higher amplitude in the left and right hemispheric sides of temporal regions (T7 and T8). Theta band frequency had shown higher power spectrum density (PSD) values while processing local attentional mechanisms. However, the significance of other frequency bands was noticeably minute. Hence, integrating the attentional mechanisms in terms of ERP and frequency signatures, a hybrid custom weight allocation model (CWAM) was built to assess and predict the contribution of insignificant channels to significant ones. The CWAM model was formulated based on the computational linear regression derivatives. All the derivatives are computationally derived the significant score while channelizing the hierarchical performance of each channel with respect to the frequent and deviant occurrences of global-local stimulus. This model enables us to configure the neural dynamicity of cognitive allocation of resources within the different locations of the human brain while processing the attentional stimulus. CWAM is reported to be the first model to evaluate the performance of the non-significant channels for enhancing the response of significant channels. The findings of the CWAM model suggest that the brain's performance may be determined by the underlying contribution of the non-significant channels.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11571-022-09888-x.

Multi-modal and multi-model interrogation of large-scale functional brain networks

Existing whole-brain models are generally tailored to the modelling of a particular data modality (e.g., fMRI or MEG/EEG). We propose that despite the differing aspects of neural activity each modality captures, they originate from shared network dynamics. Building on the universal principles of self-organising delay-coupled nonlinear systems, we aim to link distinct features of brain activity - captured across modalities - to the dynamics unfolding on a macroscopic structural connectome. To jointly predict connectivity, spatiotemporal and transient features of distinct signal modalities, we consider two large-scale models - the Stuart Landau and Wilson and Cowan models - which generate short-lived 40 Hz oscillations with varying levels of realism. To this end, we measure features of functional connectivity and metastable oscillatory modes (MOMs) in fMRI and MEG signals - and compare them against simulated data. We show that both models can represent MEG functional connectivity (FC), functional connectivity dynamics (FCD) and generate MOMs to a comparable degree. This is achieved by adjusting the global coupling and mean conduction time delay and, in the WC model, through the inclusion of balance between excitation and inhibition. For both models, the omission of delays dramatically decreased the performance. For fMRI, the SL model performed worse for FCD and MOMs, highlighting the importance of balanced dynamics for the emergence of spatiotemporal and transient patterns of ultra-slow dynamics. Notably, optimal working points varied across modalities and no model was able to achieve a correlation with empirical FC higher than 0.4 across modalities for the same set of parameters. Nonetheless, both displayed the emergence of FC patterns that extended beyond the constraints of the anatomical structure. Finally, we show that both models can generate MOMs with empirical-like properties such as size (number of brain regions engaging in a mode) and duration (continuous time interval during which a mode appears). Our results demonstrate the emergence of static and dynamic properties of neural activity at different timescales from networks of delay-coupled oscillators at 40 Hz. Given the higher dependence of simulated FC on the underlying structural connectivity, we suggest that mesoscale heterogeneities in neural circuitry may be critical for the emergence of parallel cross-modal functional networks and should be accounted for in future modelling endeavours.

Excitation-Inhibition Imbalance in Migraine: From Neurotransmitters to Brain Oscillations

Migraine is among the most common and debilitating neurological disorders typically affecting people of working age. It is characterised by a unilateral, pulsating headache often associated with severe pain. Despite the intensive research, there is still little understanding of the pathophysiology of migraine. At the electrophysiological level, altered oscillatory parameters have been reported within the alpha and gamma bands. At the molecular level, altered glutamate and GABA concentrations have been reported. However, there has been little cross-talk between these lines of research. Thus, the relationship between oscillatory activity and neurotransmitter concentrations remains to be empirically traced. Importantly, how these indices link back to altered sensory processing has to be clearly established as yet. Accordingly, pharmacologic treatments have been mostly symptom-based, and yet sometimes proving ineffective in resolving pain or related issues. This review provides an integrative theoretical framework of excitation-inhibition imbalance for the understanding of current evidence and to address outstanding questions concerning the pathophysiology of migraine. We propose the use of computational modelling for the rigorous formulation of testable hypotheses on mechanisms of homeostatic imbalance and for the development of mechanism-based pharmacological treatments and neurostimulation interventions.

Advanced Modeling and Signal Processing Methods in Brain-Computer Interfaces Based on a Vector of Cyclic Rhythmically Connected Random Processes

In this study is substantiated the new mathematical model of vector of electroencephalographic signals, registered under the conditions of multiple repetitions of the mental control influences of brain-computer interface operator, in the form of a vector of cyclic rhythmically connected random processes, which, due to taking into account the stochasticity and cyclicity, the variability and commonality of the rhythm of the investigated signals have a number of advantages over the known models. This new model opens the way for the study of multidimensional distribution functions; initial, central, and mixed moment functions of higher order such as for each electroencephalographic signal separately; as well as for their respective compatible probabilistic characteristics, among which the most informative characteristics can be selected. This provides an increase in accuracy in the detection (classification) of mental control influences of the brain-computer interface operators. Based on the developed mathematical model, the statistical processing methods of vector of electroencephalographic signals are substantiated, which consist of statistical evaluation of its probabilistic characteristics and make it possible to conduct an effective joint statistical estimation of the probability characteristics of electroencephalographic signals. This provides the basis for coordinated integration of information from different sensors. The use of moment functions of higher order and their spectral images in the frequency domain, as informative characteristics in brain-computer interface systems, are substantiated. Their significant sensitivity to the mental controlling influence of the brain-computer interface operator is experimentally established. The application of Bessel's inequality to the problems of reducing the dimensions (from 500 to 20 numbers) of the vectors of informative features makes it possible to significantly reduce the computational complexity of the algorithms for the functioning of brain-computer interface systems. Namely, we experimentally established that only the first 20 values of the Fourier transform of the estimation of moment functions of higher-order electroencephalographic signals are sufficient to form the vector of informative features in brain-computer interface systems, because these spectral components make up at least 95% of the total energy of the corresponding statistical estimate of the moment functions of higher-order electroencephalographic signals.

Revisiting Riemannian geometry-based EEG decoding through approximate joint diagonalization

Objective.The wider adoption of Riemannian geometry in electroencephalography (EEG) processing is hindered by two factors: (a) it involves the manipulation of complex mathematical formulations and, (b) it leads to computationally demanding tasks. The main scope of this work is to simplify particular notions of Riemannian geometry and provide an efficient and comprehensible scheme for neuroscientific explorations.Approach.To overcome the aforementioned shortcomings, we exploit the concept of approximate joint diagonalization in order to reconstruct the spatial covariance matrices assuming the existence of (and identifying) a common eigenspace in which the application of Riemannian geometry is significantly simplified.Main results.The employed reconstruction process abides to physiologically plausible assumptions, reduces the computational complexity in Riemannian geometry schemes and bridges the gap between rigorous mathematical procedures and computational neuroscience. Our approach is both formally established and experimentally validated by employing real and synthetic EEG data.Significance.The implications of the introduced reconstruction process are highlighted by reformulating and re-introducing two signal processing methodologies, namely the 'Symmetric Positive Definite (SPD) Matrix Quantization' and the 'Coding over SPD Atoms'. The presented approach paves the way for robust and efficient neuroscientific explorations that exploit Riemannian geometry schemes.

A multiscale brain network model links Alzheimer’s disease-mediated neuronal hyperactivity to large-scale oscillatory slowing

Background

Neuronal hyperexcitability and inhibitory interneuron dysfunction are frequently observed in preclinical animal models of Alzheimer’s disease (AD). This study investigates whether these microscale abnormalities explain characteristic large-scale magnetoencephalography (MEG) activity in human early-stage AD patients.

Methods

To simulate spontaneous electrophysiological activity, we used a whole-brain computational network model comprised of 78 neural masses coupled according to human structural brain topology. We modified relevant model parameters to simulate six literature-based cellular scenarios of AD and compare them to one healthy and six contrast (non-AD-like) scenarios. The parameters include excitability, postsynaptic potentials, and coupling strength of excitatory and inhibitory neuronal populations. Whole-brain spike density and spectral power analyses of the simulated data reveal mechanisms of neuronal hyperactivity that lead to oscillatory changes similar to those observed in MEG data of 18 human prodromal AD patients compared to 18 age-matched subjects with subjective cognitive decline.

Results

All but one of the AD-like scenarios showed higher spike density levels, and all but one of these scenarios had a lower peak frequency, higher spectral power in slower (theta, 4–8Hz) frequencies, and greater total power. Non-AD-like scenarios showed opposite patterns mainly, including reduced spike density and faster oscillatory activity. Human AD patients showed oscillatory slowing (i.e., higher relative power in the theta band mainly), a trend for lower peak frequency and higher total power compared to controls. Combining model and human data, the findings indicate that neuronal hyperactivity can lead to oscillatory slowing, likely due to hyperexcitation (by hyperexcitability of pyramidal neurons or greater long-range excitatory coupling) and/or disinhibition (by reduced excitability of inhibitory interneurons or weaker local inhibitory coupling strength) in early AD.

Conclusions

Using a computational brain network model, we link findings from different scales and models and support the hypothesis of early-stage neuronal hyperactivity underlying E/I imbalance and whole-brain network dysfunction in prodromal AD.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13195-022-01041-4.

EEG Markers of Treatment Resistance in Idiopathic Generalized Epilepsy: From Standard EEG Findings to Advanced Signal Analysis

Idiopathic generalized epilepsy (IGE) represents a common form of epilepsy in both adult and pediatric epilepsy units. Although IGE has been long considered a relatively benign epilepsy syndrome, a remarkable proportion of patients could be refractory to treatment. While some clinical prognostic factors have been largely validated among IGE patients, the impact of routine electroencephalography (EEG) findings in predicting drug resistance is still controversial and a growing number of authors highlighted the potential importance of capturing the sleep state in this setting. In addition, the development of advanced computational techniques to analyze EEG data has opened new opportunities in the identification of reliable and reproducible biomarkers of drug resistance in IGE patients. In this manuscript, we summarize the EEG findings associated with treatment resistance in IGE by reviewing the results of studies considering standard EEGs, 24-h EEG recordings, and resting-state protocols. We discuss the role of 24-h EEG recordings in assessing seizure recurrence in light of the potential prognostic relevance of generalized fast discharges occurring during sleep. In addition, we highlight new and promising biomarkers as identified by advanced EEG analysis, including hypothesis-driven functional connectivity measures of background activity and data-driven quantitative findings revealed by machine learning approaches. Finally, we thoroughly discuss the methodological limitations observed in existing studies and briefly outline future directions to identify reliable and replicable EEG biomarkers in IGE patients.

Structure-function models of temporal, spatial, and spectral characteristics of non-invasive whole brain functional imaging

We review recent advances in using mathematical models of the relationship between the brain structure and function that capture features of brain dynamics. We argue the need for models that can jointly capture temporal, spatial, and spectral features of brain functional activity. We present recent work on spectral graph theory based models that can accurately capture spectral as well as spatial patterns across multiple frequencies in MEG reconstructions.

Kuramoto model for populations of quadratic integrate-and-fire neurons with chemical and electrical coupling

We derive the Kuramoto model (KM) corresponding to a population of weakly coupled, nearly identical quadratic integrate-and-fire (QIF) neurons with both electrical and chemical coupling. The ratio of chemical to electrical coupling determines the phase lag of the characteristic sine coupling function of the KM and critically determines the synchronization properties of the network. We apply our results to uncover the presence of chimera states in two coupled populations of identical QIF neurons. We find that the presence of both electrical and chemical coupling is a necessary condition for chimera states to exist. Finally, we numerically demonstrate that chimera states gradually disappear as coupling strengths cease to be weak.