Ballistocardiogram artifact correction taking into account physiological signal preservation in simultaneous EEG-fMRI

Denoising of the gradient artifact present in simultaneous studies of muscle blood oxygen level dependent (BOLD) signal and electromyography (EMG)

The MR-induced gradient artifact affects EMG recordings during simultaneous muscle BOLD/EMG acquisitions. However, no dedicated hardware can remove the gradient artifact easily, and alternative methods are expensive and time-consuming. This study aimed to develop three denoising methods requiring different processing levels and MR-compatible hardware. At two time points, surface EMG was recorded from the lower leg of 6 participants (50:50 sex ratio, age = 26.24.6 yrs., height = 173.59.2 cm, weight = 71.511.4 kg) using a plantar flexion-based block design consisting of 30s of rest followed by 30s of flexion for 5 min, under three conditions: inside the MRI bore, with and without a BOLD sequence (3 T, BOLD sequence, GRE EPI, 10 slices, 64×64 matrix, 2 mm thickness, and TE/TR/flip = 35/3000 ms/70), and outside the MRI environment. Simultaneous BOLD/EMG recordings were denoised using average artifact subtraction with three methods of artifact template creation, each having varying timing and hardware requirements. Method M1 builds the artifact template by recording the scanner triggers coming from the MRI; M2 creates the artifact template with a constant artifact period computed as TR/[number of slices]; M3 estimates the artifact template by looking at the periodicity of the gradient artifact located in the EMG recordings. Following postprocessing, SNR analysis was performed, comparing rest-to-flexion periods, to assess the efficacy of denoising methods and to compare differences between conditions. Linear mixed-effects models showed no significant differences in the mean SNR between denoising methods (p = 0.656). Furthermore, EMG SNR measurements were significantly affected by the magnetic environment (p < 0.05) but not by muscle fatigue over time (p = 0.975). EMG recordings contaminated with gradient artifacts during simultaneous BOLD/EMG can be efficiently denoised using all proposed methods, with two methods requiring no extra hardware. With minimal post-processing, EMG can easily be performed during muscle BOLD MRI studies.

NeuXus open-source tool for real-time artifact reduction in simultaneous EEG-fMRI

The simultaneous acquisition of electroencephalography and functional magnetic resonance imaging (EEG-fMRI) allows the complementary study of the brain's electrophysiology and hemodynamics with high temporal and spatial resolution. One application with great potential is neurofeedback training of targeted brain activity, based on the real-time analysis of the EEG and/or fMRI signals. This depends on the ability to reduce in real time the severe artifacts affecting the EEG signal acquired with fMRI, mainly the gradient and pulse artifacts. A few methods have been proposed for this purpose, but they are either slow, hardware-dependent, publicly unavailable, or proprietary software. Here, we present a fully open-source and publicly available tool for real-time EEG artifact reduction in simultaneous EEG-fMRI recordings that is fast and applicable to any hardware. Our tool is integrated in the Python toolbox NeuXus for real-time EEG processing and adapts to a real-time scenario well-established artifact average subtraction methods combined with a long short-term memory network for R peak detection. We benchmarked NeuXus on three different datasets, in terms of artifact power reduction and background signal preservation in resting state, alpha-band power reactivity to eyes closure, and event-related desynchronization during motor imagery. We showed that NeuXus performed at least as well as the only available real-time tool for conventional hardware setups (BrainVision's RecView) and a well-established offline tool (EEGLAB's FMRIB plugin). We also demonstrated NeuXus' real-time ability by reporting execution times under 250 ms. In conclusion, we present and validate the first fully open-source and hardware-independent solution for real-time artifact reduction in simultaneous EEG-fMRI studies.

Comparison of Resting-State EEG Network Analyses With and Without Parallel MRI in Genetic Generalized Epilepsy

Genetic generalized epilepsy (GGE) is conceptualized as a brain disorder involving distributed bilateral networks. To study these networks, simultaneous EEG-fMRI measurements can be used. However, inside-MRI EEG suffers from strong MR-related artifacts; it is not established whether EEG-based metrics in EEG-fMRI resting-state measurements are suitable for the analysis of group differences at source-level. We evaluated the impact of the inside-MR measurement condition on statistical group comparisons of EEG on source-level power and functional connectivity in patients with GGE versus healthy controls. We studied the cross-modal spatial relation of statistical group differences in seed-based FC derived from EEG and parallel fMRI. We found a significant increase in power and a frequency-specific change in functional connectivity for the inside MR-scanner compared to the outside MR-scanner condition. For power, we found reduced group difference between GGE and controls both in terms of statistical significance as well as effect size. Group differences for ImCoh remained similar both in terms of statistical significance as well as effect size. We found increased seed-based FC for GGE patients from the thalamus to the precuneus cortex region in fMRI, and in the theta band of simultaneous EEG. Our findings suggest that the analysis of EEG functional connectivity based on ImCoh is suitable for MR-EEG, and that relative group difference in a comparison of patients with GGE against controls are preserved. Spatial correspondence of seed-based FC group differences between the two modalities was found for the thalamus.

EEG-LLAMAS: A low-latency neurofeedback platform for artifact reduction in EEG-fMRI

Simultaneous EEG-fMRI is a powerful multimodal technique for imaging the brain, but its use in neurofeedback experiments has been limited by EEG noise caused by the MRI environment. Neurofeedback studies typically require analysis of EEG in real time, but EEG acquired inside the scanner is heavily contaminated with ballistocardiogram (BCG) artifact, a high-amplitude artifact locked to the cardiac cycle. Although techniques for removing BCG artifacts do exist, they are either not suited to real-time, low-latency applications, such as neurofeedback, or have limited efficacy. We propose and validate a new open-source artifact removal software called EEG-LLAMAS (Low Latency Artifact Mitigation Acquisition Software), which adapts and advances existing artifact removal techniques for low-latency experiments. We first used simulations to validate LLAMAS in data with known ground truth. We found that LLAMAS performed better than the best publicly-available real-time BCG removal technique, optimal basis sets (OBS), in terms of its ability to recover EEG waveforms, power spectra, and slow wave phase. To determine whether LLAMAS would be effective in practice, we then used it to conduct real-time EEG-fMRI recordings in healthy adults, using a steady state visual evoked potential (SSVEP) task. We found that LLAMAS was able to recover the SSVEP in real time, and recovered the power spectra collected outside the scanner better than OBS. We also measured the latency of LLAMAS during live recordings, and found that it introduced a lag of less than 50 ms on average. The low latency of LLAMAS, coupled with its improved artifact reduction, can thus be effectively used for EEG-fMRI neurofeedback. A limitation of the method is its use of a reference layer, a piece of EEG equipment which is not commercially available, but can be assembled in-house. This platform enables closed-loop experiments which previously would have been prohibitively difficult, such as those that target short-duration EEG events, and is shared openly with the neuroscience community.

The MotoNet: A 3 Tesla MRI-Conditional EEG Net with Embedded Motion Sensors

We introduce a new electroencephalogram (EEG) net, which will allow clinicians to monitor EEG while tracking head motion. Motion during MRI limits patient scans, especially of children with epilepsy. EEG is also severely affected by motion-induced noise, predominantly ballistocardiogram (BCG) noise due to the heartbeat.

METHODS

The MotoNet was built using polymer thick film (PTF) EEG leads and motion sensors on opposite sides in the same flex circuit. EEG/motion measurements were made with a standard commercial EEG acquisition system in a 3 Tesla (T) MRI. A Kalman filtering-based BCG correction tool was used to clean the EEG in healthy volunteers.

RESULTS

MRI safety studies in 3 T confirmed the maximum heating below 1 °C. Using an MRI sequence with spatial localization gradients only, the position of the head was linearly correlated with the average motion sensor output. Kalman filtering was shown to reduce the BCG noise and recover artifact-clean EEG.

CONCLUSIONS

The MotoNet is an innovative EEG net design that co-locates 32 EEG electrodes with 32 motion sensors to improve both EEG and MRI signal quality. In combination with custom gradients, the position of the net can, in principle, be determined. In addition, the motion sensors can help reduce BCG noise.

Preservation of EEG spectral power features during simultaneous EEG-fMRI

Introduction

Electroencephalographic (EEG) data quality is severely compromised when recorded inside the magnetic resonance (MR) environment. Here we characterized the impact of the ballistocardiographic (BCG) artifact on resting-state EEG spectral properties and compared the effectiveness of seven common BCG correction methods to preserve EEG spectral features. We also assessed if these methods retained posterior alpha power reactivity to an eyes closure-opening (EC-EO) task and compared the results from EEG-informed fMRI analysis using different BCG correction approaches.

Method

Electroencephalographic data from 20 healthy young adults were recorded outside the MR environment and during simultaneous fMRI acquisition. The gradient artifact was effectively removed from EEG-fMRI acquisitions using Average Artifact Subtraction (AAS). The BCG artifact was corrected with seven methods: AAS, Optimal Basis Set (OBS), Independent Component Analysis (ICA), OBS followed by ICA, AAS followed by ICA, PROJIC-AAS and PROJIC-OBS. EEG signal preservation was assessed by comparing the spectral power of traditional frequency bands from the corrected rs-EEG-fMRI data with the data recorded outside the scanner. We then assessed the preservation of posterior alpha functional reactivity by computing the ratio between the EC and EO conditions during the EC-EO task. EEG-informed fMRI analysis of the EC-EO task was performed using alpha power-derived BOLD signal predictors obtained from the EEG signals corrected with different methods.

Results

The BCG artifact caused significant distortions (increased absolute power, altered relative power) across all frequency bands. Artifact residuals/signal losses were present after applying all correction methods. The EEG reactivity to the EC-EO task was better preserved with ICA-based correction approaches, particularly when using ICA feature extraction to isolate alpha power fluctuations, which allowed to accurately predict hemodynamic signal fluctuations during the EEG-informed fMRI analysis.

Discussion

Current software solutions for the BCG artifact problem offer limited efficiency to preserve the EEG spectral power properties using this particular EEG setup. The state-of-the-art approaches tested here can be further refined and should be combined with hardware implementations to better preserve EEG signal properties during simultaneous EEG-fMRI. Existing and novel BCG artifact correction methods should be validated by evaluating signal preservation of both ERPs and spontaneous EEG spectral power.

Simultaneous EEG-fMRI: What Have We Learned and What Does the Future Hold?

Simultaneous EEG-fMRI has developed into a mature measurement technique in the past 25 years. During this time considerable technical and analytical advances have been made, enabling valuable scientific contributions to a range of research fields. This review will begin with an introduction to the measurement principles involved in EEG and fMRI and the advantages of combining these methods. The challenges faced when combining the two techniques will then be considered. An overview of the leading application fields where EEG-fMRI has made a significant contribution to the scientific literature and emerging applications in EEG-fMRI research trends is then presented.

EEG-fMRI: Ballistocardiogram Artifact Reduction by Surrogate Method for Improved Source Localization

For the analysis of simultaneous EEG-fMRI recordings, it is vital to use effective artifact removal tools. This applies in particular to the ballistocardiogram (BCG) artifact which is difficult to remove without distorting signals of interest related to brain activity. Here, we documented the use of surrogate source models to separate the artifact-related signals from brain signals with minimal distortion of the brain activity of interest. The artifact topographies used for surrogate separation were created automatically using principal components analysis (PCA-S) or by manual selection of artifact components utilizing independent components analysis (ICA-S). Using real resting-state data from 55 subjects superimposed with simulated auditory evoked potentials (AEP), both approaches were compared with three established BCG artifact removal methods: Blind Source Separation (BSS), Optimal Basis Set (OBS), and a mixture of both (OBS-ICA). Each method was evaluated for its applicability for ERP and source analysis using the following criteria: the number of events surviving artifact threshold scans, signal-to-noise ratio (SNR), error of source localization, and signal variance explained by the dipolar model. Using these criteria, PCA-S and ICA-S fared best overall, with highly significant differences to the established methods, especially in source localization. The PCA-S approach was also applied to a single subject Berger experiment performed in the MRI scanner. Overall, the removal of BCG artifacts by the surrogate methods provides a substantial improvement for the analysis of simultaneous EEG-fMRI data compared to the established methods.

Optimizing EEG Source Reconstruction with Concurrent fMRI-Derived Spatial Priors

Reconstructing EEG sources involves a complex pipeline, with the inverse problem being the most challenging. Multiple inversion algorithms are being continuously developed, aiming to tackle the non-uniqueness of this problem, which has been shown to be partially circumvented by including prior information in the inverse models. Despite a few efforts, there are still current and persistent controversies regarding the inversion algorithm of choice and the optimal set of spatial priors to be included in the inversion models. The use of simultaneous EEG-fMRI data is one approach to tackle this problem. The spatial resolution of fMRI makes fMRI derived spatial priors very convenient for EEG reconstruction, however, only task activation maps and resting-state networks (RSNs) have been explored so far, overlooking the recent, but already accepted, notion that brain networks exhibit dynamic functional connectivity fluctuations. The lack of a systematic comparison between different source reconstruction algorithms, considering potentially more brain-informative priors such as fMRI, motivates the search for better reconstruction models. Using simultaneous EEG-fMRI data, here we compared four different inversion algorithms (minimum norm, MN; low resolution electromagnetic tomography, LORETA; empirical Bayes beamformer, EBB; and multiple sparse priors, MSP) under a Bayesian framework (as implemented in SPM), each with three different sets of priors consisting of: (1) those specific to the algorithm; (2) those specific to the algorithm plus fMRI task activation maps and RSNs; and (3) those specific to the algorithm plus fMRI task activation maps and RSNs and network modules of task-related dFC states estimated from the dFC fluctuations. The quality of the reconstructed EEG sources was quantified in terms of model-based metrics, namely the expectation of the posterior probability P(model|data) and variance explained of the inversion models, and the overlap/proportion of brain regions known to be involved in the visual perception tasks that the participants were submitted to, and RSN templates, with/within EEG source components. Model-based metrics suggested that model parsimony is preferred, with the combination MSP and priors specific to this algorithm exhibiting the best performance. However, optimal overlap/proportion values were found using EBB and priors specific to this algorithm and fMRI task activation maps and RSNs or MSP and considering all the priors (algorithm priors, fMRI task activation maps and RSNs and dFC state modules), respectively, indicating that fMRI spatial priors, including dFC state modules, might contain useful information to recover EEG source components reflecting neuronal activity of interest. Our main results show that providing fMRI spatial derived priors that reflect the dynamics of the brain might be useful to map neuronal activity more accurately from EEG-fMRI. Furthermore, this work paves the way towards a more informative selection of the optimal EEG source reconstruction approach, which may be critical in future studies.

Automated pipeline for EEG artifact reduction (APPEAR) recorded during fMRI

Objective.Simultaneous electroencephalography-functional magnetic resonance imaging (EEG-fMRI) recordings offer a high spatiotemporal resolution approach to study human brain and understand the underlying mechanisms mediating cognitive and behavioral processes. However, the high susceptibility of EEG to MRI-induced artifacts hinders a broad adaptation of this approach. More specifically, EEG data collected during fMRI acquisition are contaminated with MRI gradients and ballistocardiogram artifacts, in addition to artifacts of physiological origin. There have been several attempts for reducing these artifacts with manual and time-consuming pre-processing, which may result in biasing EEG data due to variations in selecting steps order, parameters, and classification of artifactual independent components. Thus, there is a strong urge to develop a fully automatic and comprehensive pipeline for reducing all major EEG artifacts. In this work, we introduced an open-access toolbox with a fully automatic pipeline for reducing artifacts from EEG data collected simultaneously with fMRI (refer to APPEAR).Approach.The pipeline integrates average template subtraction and independent component analysis to suppress both MRI-related and physiological artifacts. To validate our results, we tested APPEAR on EEG data recorded from healthy control subjects during resting-state (n= 48) and task-based (i.e. event-related-potentials (ERPs);n= 8) paradigms. The chosen gold standard is an expert manual review of the EEG database.Main results.We compared manually and automated corrected EEG data during resting-state using frequency analysis and continuous wavelet transformation and found no significant differences between the two corrections. A comparison between ERP data recorded during a so-called stop-signal task (e.g. amplitude measures and signal-to-noise ratio) also showed no differences between the manually and fully automatic fMRI-EEG-corrected data.Significance.APPEAR offers the first comprehensive open-source toolbox that can speed up advancement of EEG analysis and enhance replication by avoiding experimenters' preferences while allowing for processing large EEG-fMRI cohorts composed of hundreds of subjects with manageable researcher time and effort.

Data-driven beamforming technique to attenuate ballistocardiogram artefacts in electroencephalography-functional magnetic resonance imaging without detecting cardiac pulses in electrocardiography recordings

Simultaneous recording of electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) is a very promising non‐invasive neuroimaging technique. However, EEG data obtained from the simultaneous EEG–fMRI are strongly influenced by MRI‐related artefacts, namely gradient artefacts (GA) and ballistocardiogram (BCG) artefacts. When compared to the GA correction, the BCG correction is more challenging to remove due to its inherent variabilities and dynamic changes over time. The standard BCG correction (i.e., average artefact subtraction [AAS]), require detecting cardiac pulses from simultaneous electrocardiography (ECG) recording. However, ECG signals are also distorted and will become problematic for detecting reliable cardiac peaks. In this study, we focused on a beamforming spatial filtering technique to attenuate all unwanted source activities outside of the brain. Specifically, we applied the beamforming technique to attenuate the BCG artefact in EEG–fMRI, and also to recover meaningful task‐based neural signals during an attentional network task (ANT) which required participants to identify visual cues and respond accurately. We analysed EEG–fMRI data in 20 healthy participants during the ANT, and compared four different BCG corrections (non‐BCG corrected, AAS BCG corrected, beamforming + AAS BCG corrected, beamforming BCG corrected). We demonstrated that the beamforming approach did not only significantly reduce the BCG artefacts, but also significantly recovered the expected task‐based brain activity when compared to the standard AAS correction. This data‐driven beamforming technique appears promising especially for longer data acquisition of sleep and resting EEG–fMRI. Our findings extend previous work regarding the recovery of meaningful EEG signals by an optimized suppression of MRI‐related artefacts.

The relationship between EEG and fMRI connectomes is reproducible across simultaneous EEG-fMRI studies from 1.5T to 7T

Both electroencephalography (EEG) and functional Magnetic Resonance Imaging (fMRI) are non-invasive methods that show complementary aspects of human brain activity. Despite measuring different proxies of brain activity, both the measured blood-oxygenation (fMRI) and neurophysiological recordings (EEG) are indirectly coupled. The electrophysiological and BOLD signal can map the underlying functional connectivity structure at the whole brain scale at different timescales. Previous work demonstrated a moderate but significant correlation between resting-state functional connectivity of both modalities, however there is a wide range of technical setups to measure simultaneous EEG-fMRI and the reliability of those measures between different setups remains unknown. This is true notably with respect to different magnetic field strengths (low and high field) and different spatial sampling of EEG (medium to high-density electrode coverage). Here, we investigated the reproducibility of the bimodal EEG-fMRI functional connectome in the most comprehensive resting-state simultaneous EEG-fMRI dataset compiled to date including a total of 72 subjects from four different imaging centers. Data was acquired from 1.5T, 3T and 7T scanners with simultaneously recorded EEG using 64 or 256 electrodes. We demonstrate that the whole-brain monomodal connectivity reproducibly correlates across different datasets and that a moderate crossmodal correlation between EEG and fMRI connectivity of r ≈ 0.3 can be reproducibly extracted in low- and high-field scanners. The crossmodal correlation was strongest in the EEG-β frequency band but exists across all frequency bands. Both homotopic and within intrinsic connectivity network (ICN) connections contributed the most to the crossmodal relationship. This study confirms, using a considerably diverse range of recording setups, that simultaneous EEG-fMRI offers a consistent estimate of multimodal functional connectomes in healthy subjects that are dominantly linked through a functional core of ICNs across spanning across the different timescales measured by EEG and fMRI. This opens new avenues for estimating the dynamics of brain function and provides a better understanding of interactions between EEG and fMRI measures. This observed level of reproducibility also defines a baseline for the study of alterations of this coupling in pathological conditions and their role as potential clinical markers.

Excitatory-inhibitory balance within EEG microstates and resting-state fMRI networks: assessed via simultaneous trimodal PET-MR-EEG imaging

The symbiosis of neuronal activities and glucose energy metabolism is reflected in the generation of functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) signals. However, their association with the balance between neuronal excitation and inhibition (E/I-B), which is closely related to the activities of glutamate and γ-aminobutyric acid (GABA) and the receptor availability (RA) of GABAA and mGluR5, remains unexplored. This research investigates these associations during the resting state (RS) condition using simultaneously recorded PET/MR/EEG (trimodal) data. The trimodal data were acquired from three studies using different radio-tracers such as, [11C]ABP688 (ABP) (N = 9), [11C]Flumazenil (FMZ) (N = 10) and 2-[18F]fluoro-2-deoxy-D-glucose (FDG) (N = 10) targeted to study the mGluR5, GABAA receptors and glucose metabolism respectively. Glucose metabolism and neuroreceptor binding availability (non-displaceable binding potential (BPND)) of GABAA and mGluR5 were found to be significantly higher and closely linked within core resting-state networks (RSNs). The neuronal generators of EEG microstates and the fMRI measures were most tightly associated with the BPND of GABAA relative to mGluR5 BPND and the glucose metabolism, emphasising a predominance of inhibitory processes within in the core RSNs at rest. Changes in the neuroreceptors leading to an altered coupling with glucose metabolism may render the RSNs vulnerable to psychiatric conditions. The paradigm employed here will likely help identify the precise neurobiological mechanisms behind these alterations in fMRI functional connectivity and EEG oscillations, potentially benefitting individualised healthcare treatment measures.

Simultaneous fMRI-EEG-Based Characterisation of NREM Parasomnia Disease: Methods and Limitations

Functional magnetic resonance imaging (fMRI) techniques and electroencephalography (EEG) were used to investigate sleep with a focus on impaired arousal mechanisms in disorders of arousal (DOAs). With a prevalence of 2–4% in adults, DOAs are significant disorders that are currently gaining attention among physicians. The paper describes a simultaneous EEG and fMRI experiment conducted in adult individuals with DOAs (n=10). Both EEG and fMRI data were validated by reproducing well established EEG and fMRI associations. A method for identification of both brain functional areas and EEG rhythms associated with DOAs in shallow sleep was designed. Significant differences between patients and controls were found in delta, theta, and alpha bands during awakening epochs. General linear models of the blood-oxygen-level-dependent signal have shown the secondary visual cortex and dorsal posterior cingulate cortex to be associated with alpha spectral power fluctuations, and the precuneus with delta spectral power fluctuations, specifically in patients and not in controls. Future EEG–fMRI sleep studies should also consider subject comfort as an important aspect in the experimental design.

Pushing the Limits of EEG: Estimation of Large-Scale Functional Brain Networks and Their Dynamics Validated by Simultaneous fMRI

Functional magnetic resonance imaging (fMRI) is the technique of choice for detecting large-scale functional brain networks and to investigate their dynamics. Because fMRI measures brain activity indirectly, electroencephalography (EEG) has been recently considered a feasible tool for detecting such networks, particularly the resting-state networks (RSNs). However, a truly unbiased validation of such claims is still missing, which can only be accomplished by using simultaneously acquired EEG and fMRI data, due to the spontaneous nature of the activity underlying the RSNs. Additionally, EEG is still poorly explored for the purpose of mapping task-specific networks, and no studies so far have been focused on investigating networks' dynamic functional connectivity (dFC) with EEG. Here, we started by validating RSNs derived from the continuous reconstruction of EEG sources by directly comparing them with those derived from simultaneous fMRI data of 10 healthy participants, and obtaining an average overlap (quantified by the Dice coefficient) of 0.4. We also showed the ability of EEG to map the facial expressions processing network (FEPN), highlighting regions near the posterior superior temporal sulcus, where the FEPN is anchored. Then, we measured the dFC using EEG for the first time in this context, estimated dFC brain states using dictionary learning, and compared such states with those obtained from the fMRI. We found a statistically significant match between fMRI and EEG dFC states, and determined the existence of two matched dFC states which contribution over time was associated with the brain activity at the FEPN, showing that the dynamics of FEPN can be captured by both fMRI and EEG. Our results push the limits of EEG toward being used as a brain imaging tool, while supporting the growing literature on EEG correlates of (dynamic) functional connectivity measured with fMRI, and providing novel insights into the coupling mechanisms underlying the two imaging techniques.

Investigating the variability of cardiac pulse artifacts across heartbeats in simultaneous EEG-fMRI recordings: A 7T study

Electroencephalography (EEG) recordings performed in magnetic resonance imaging (MRI) scanners are affected by complex artifacts caused by heart function, often termed pulse artifacts (PAs). PAs can strongly compromise EEG data quality, and remain an open problem for EEG-fMRI. This study investigated the properties and mechanisms of PA variability across heartbeats, which has remained largely unaddressed to date, and evaluated its impact on PA correction approaches. Simultaneous EEG-fMRI was performed at 7T on healthy participants at rest or under visual stimulation, with concurrent recordings of breathing and cardiac activity. PA variability was found to contribute to EEG variance with more than 500 μV² at 7T, which extrapolates to 92 μV² at 3T. Clustering analyses revealed that PA variability not only is linked to variations in head position/orientation, as previously hypothesized, but also, and more importantly, to the respiratory cycle and to heart rate fluctuations. The latter mechanisms are associated to short-timescale variability (even across consecutive heartbeats), and their importance varied across EEG channels. In light of this PA variability, three PA correction techniques were compared: average artifact subtraction (AAS), optimal basis sets (OBS), and an approach based on K-means clustering. All methods allowed the recovery of visual evoked potentials from the EEG data; nonetheless, OBS and K-means tended to outperform AAS, likely due to the inability of the latter in modeling short-timescale variability. Altogether, these results offer novel insights into the dynamics and underlying mechanisms of the pulse artifact, with important consequences for its correction, relevant to most EEG-fMRI applications.

Identification of epileptic brain states by dynamic functional connectivity analysis of simultaneous EEG-fMRI: a dictionary learning approach

Most fMRI studies of the brain's intrinsic functional connectivity (FC) have assumed that this is static; however, it is now clear that it changes over time. This is particularly relevant in epilepsy, which is characterized by a continuous interchange between epileptic and normal brain states associated with the occurrence of epileptic activity. Interestingly, recurrent states of dynamic FC (dFC) have been found in fMRI data using unsupervised learning techniques, assuming either their sparse or non-sparse combination. Here, we propose an l₁-norm regularized dictionary learning (l₁-DL) approach for dFC state estimation, which allows an intermediate and flexible degree of sparsity in time, and demonstrate its application in the identification of epilepsy-related dFC states using simultaneous EEG-fMRI data. With this l₁-DL approach, we aim to accommodate a potentially varying degree of sparsity upon the interchange between epileptic and non-epileptic dFC states. The simultaneous recording of the EEG is used to extract time courses representative of epileptic activity, which are incorporated into the fMRI dFC state analysis to inform the selection of epilepsy-related dFC states. We found that the proposed l₁-DL method performed best at identifying epilepsy-related dFC states, when compared with two alternative methods of extreme sparsity (k-means clustering, maximum; and principal component analysis, minimum), as well as an l₀-norm regularization framework (l₀-DL), with a fixed amount of temporal sparsity. We further showed that epilepsy-related dFC states provide novel insights into the dynamics of epileptic networks, which go beyond the information provided by more conventional EEG-correlated fMRI analysis, and which were concordant with the clinical profile of each patient. In addition to its application in epilepsy, our study provides a new dFC state identification method of potential relevance for studying brain functional connectivity dynamics in general.

Exploring the relative efficacy of motion artefact correction techniques for EEG data acquired during simultaneous fMRI

Simultaneous EEG‐fMRI allows multiparametric characterisation of brain function, in principle enabling a more complete understanding of brain responses; unfortunately the hostile MRI environment severely reduces EEG data quality. Simply eliminating data segments containing gross motion artefacts [MAs] (generated by movement of the EEG system and head in the MRI scanner's static magnetic field) was previously believed sufficient. However recently the importance of removal of all MAs has been highlighted and new methods developed. A systematic comparison of the ability to remove MAs and retain underlying neuronal activity using different methods of MA detection and post‐processing algorithms is needed to guide the neuroscience community. Using a head phantom, we recorded MAs while simultaneously monitoring the motion using three different approaches: Reference Layer Artefact Subtraction (RLAS), Moiré Phase Tracker (MPT) markers and Wire Loop Motion Sensors (WLMS). These EEG recordings were combined with EEG responses to simple visual tasks acquired on a subject outside the MRI environment. MAs were then corrected using the motion information collected with each of the methods combined with different analysis pipelines. All tested methods retained the neuronal signal. However, often the MA was not removed sufficiently to allow accurate detection of the underlying neuronal signal. We show that the MA is best corrected using the RLAS combined with post‐processing using a multichannel, recursive least squares (M‐RLS) algorithm. This method needs to be developed further to enable practical utility; thus, WLMS combined with M‐RLS currently provides the best compromise between EEG data quality and practicalities of motion detection.

Adaptive optimal basis set for BCG artifact removal in simultaneous EEG-fMRI

Electroencephalography (EEG) signals recorded during simultaneous functional magnetic resonance imaging (fMRI) are contaminated by strong artifacts. Among these, the ballistocardiographic (BCG) artifact is the most challenging, due to its complex spatio-temporal dynamics associated with ongoing cardiac activity. The presence of BCG residuals in EEG data may hide true, or generate spurious correlations between EEG and fMRI time-courses. Here, we propose an adaptive Optimal Basis Set (aOBS) method for BCG artifact removal. Our method is adaptive, as it can estimate the delay between cardiac activity and BCG occurrence on a beat-to-beat basis. The effective creation of an optimal basis set by principal component analysis (PCA) is therefore ensured by a more accurate alignment of BCG occurrences. Furthermore, aOBS can automatically estimate which components produced by PCA are likely to be BCG artifact-related and therefore need to be removed. The aOBS performance was evaluated on high-density EEG data acquired with simultaneous fMRI in healthy subjects during visual stimulation. As aOBS enables effective reduction of BCG residuals while preserving brain signals, we suggest it may find wide application in simultaneous EEG-fMRI studies.

Clustering-Constrained ICA for Ballistocardiogram Artifacts Removal in Simultaneous EEG-fMRI

Combination of electroencephalogram (EEG) recording and functional magnetic resonance imaging (fMRI) plays a potential role in neuroimaging due to its high spatial and temporal resolution. However, EEG is easily influenced by ballistocardiogram (BCG) artifacts and may cause false identification of the related EEG features, such as epileptic spikes. There are many related methods to remove them, however, they do not consider the time-varying features of BCG artifacts. In this paper, a novel method using clustering algorithm to catch the BCG artifacts' features and together with the constrained ICA (ccICA) is proposed to remove the BCG artifacts. We first applied this method to the simulated data, which was constructed by adding the BCG artifacts to the EEG signal obtained from the conventional environment. Then, our method was tested to demonstrate the effectiveness during EEG and fMRI experiments on 10 healthy subjects. In simulated data analysis, the value of error in signal amplitude (Er) computed by ccICA method was lower than those from other methods including AAS, OBS, and cICA (p < 0.005). In vivo data analysis, the Improvement of Normalized Power Spectrum (INPS) calculated by ccICA method in all electrodes was much higher than AAS, OBS, and cICA methods (p < 0.005). We also used other evaluation index (e.g., power analysis) to compare our method with other traditional methods. In conclusion, our novel method successfully and effectively removed BCG artifacts in both simulated and vivo EEG data tests, showing the potentials of removing artifacts in EEG-fMRI applications.

Heart-Brain Interactions in the MR Environment: Characterization of the Ballistocardiogram in EEG Signals Collected During Simultaneous fMRI

The ballistocardiographic (BCG) artifact is linked to cardiac activity and occurs in electroencephalographic (EEG) recordings acquired inside the magnetic resonance (MR) environment. Its variability in terms of amplitude, waveform shape and spatial distribution over subject's scalp makes its attenuation a challenging task. In this study, we aimed to provide a detailed characterization of the BCG properties, including its temporal dependency on cardiac events and its spatio-temporal dynamics. To this end, we used high-density EEG data acquired during simultaneous functional MR imaging in six healthy volunteers. First, we investigated the relationship between cardiac activity and BCG occurrences in the EEG recordings. We observed large variability in the delay between ECG and subsequent BCG events (ECG-BCG delay) across subjects and non-negligible epoch-by-epoch variations at the single subject level. The inspection of spatial-temporal variations revealed a prominent non-stationarity of the BCG signal. We identified five main BCG waves, which were common across subjects. Principal component analysis revealed two spatially distinct patterns to explain most of the variance (85% in total). These components are possibly related to head rotation and pulse-driven scalp expansion, respectively. Our results may inspire the development of novel, more effective methods for the removal of the BCG, capable of isolating and attenuating artifact occurrences while preserving true neuronal activity.

EEG-Informed fMRI: A Review of Data Analysis Methods

The simultaneous acquisition of electroencephalography (EEG) with functional magnetic resonance imaging (fMRI) is a very promising non-invasive technique for the study of human brain function. Despite continuous improvements, it remains a challenging technique, and a standard methodology for data analysis is yet to be established. Here we review the methodologies that are currently available to address the challenges at each step of the data analysis pipeline. We start by surveying methods for pre-processing both EEG and fMRI data. On the EEG side, we focus on the correction for several MR-induced artifacts, particularly the gradient and pulse artifacts, as well as other sources of EEG artifacts. On the fMRI side, we consider image artifacts induced by the presence of EEG hardware inside the MR scanner, and the contamination of the fMRI signal by physiological noise of non-neuronal origin, including a review of several approaches to model and remove it. We then provide an overview of the approaches specifically employed for the integration of EEG and fMRI when using EEG to predict the blood oxygenation level dependent (BOLD) fMRI signal, the so-called EEG-informed fMRI integration strategy, the most commonly used strategy in EEG-fMRI research. Finally, we systematically review methods used for the extraction of EEG features reflecting neuronal phenomena of interest.

EEG synchronization measures predict epilepsy-related BOLD-fMRI fluctuations better than commonly used univariate metrics

OBJECTIVE

We hypothesize that the hypersynchronization associated with epileptic activity is best described by EEG synchronization measures, and propose to use these as predictors of epilepsy-related BOLD fluctuations.

METHODS

We computed the phase synchronization index (PSI) and global field synchronization (GFS), within two frequency bands, a broadband (1-45 Hz) and a narrower band focused on the presence of epileptic activity (3-10 Hz). The associated epileptic networks were compared with those obtained using conventional unitary regressors and two power-weighted metrics (total power and root mean square frequency), on nine simultaneous EEG-fMRI datasets from four epilepsy patients, exhibiting inter-ictal epileptiform discharges (IEDs).

RESULTS

The average PSI within 3-10 Hz achieved the best performance across several measures reflecting reliability in all datasets. The results were cross-validated through electrical source imaging of the IEDs. The applicability of PSI when no IEDs are recorded on the EEG was evaluated on three additional patients, yielding partially plausible networks in all cases.

CONCLUSIONS

Epileptic networks can be mapped based on the EEG PSI metric within an IED-specific frequency band, performing better than commonly used EEG metrics.

SIGNIFICANCE

This is the first study to investigate EEG synchronization measures as potential predictors of epilepsy-related BOLD fluctuations.

Online Reduction of Artifacts in EEG of Simultaneous EEG-fMRI Using Reference Layer Adaptive Filtering (RLAF)

Simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) allow us to study the active human brain from two perspectives concurrently. Signal processing based artifact reduction techniques are mandatory for this, however, to obtain reasonable EEG quality in simultaneous EEG-fMRI. Current artifact reduction techniques like average artifact subtraction (AAS), typically become less effective when artifact reduction has to be performed on-the-fly. We thus present and evaluate a new technique to improve EEG quality online. This technique adds up with online AAS and combines a prototype EEG-cap for reference recordings of artifacts, with online adaptive filtering and is named reference layer adaptive filtering (RLAF). We found online AAS + RLAF to be highly effective in improving EEG quality. Online AAS + RLAF outperformed online AAS and did so in particular online in terms of the chosen performance metrics, these being specifically alpha rhythm amplitude ratio between closed and opened eyes (3-45% improvement), signal-to-noise-ratio of visual evoked potentials (VEP) (25-63% improvement), and VEPs variability (16-44% improvement). Further, we found that EEG quality after online AAS + RLAF is occasionally even comparable with the offline variant of AAS at a 3T MRI scanner. In conclusion RLAF is a very effective add-on tool to enable high quality EEG in simultaneous EEG-fMRI experiments, even when online artifact reduction is necessary.

Removal of BCG artefact from concurrent fMRI-EEG recordings based on EMD and PCA

BACKGROUND

Simultaneous electroencephalography (EEG) and functional magnetic resonance image (fMRI) acquisitions provide better insight into brain dynamics. Some artefacts due to simultaneous acquisition pose a threat to the quality of the data. One such problematic artefact is the ballistocardiogram (BCG) artefact.

METHODS

We developed a hybrid algorithm that combines features of empirical mode decomposition (EMD) with principal component analysis (PCA) to reduce the BCG artefact. The algorithm does not require extra electrocardiogram (ECG) or electrooculogram (EOG) recordings to extract the BCG artefact.

RESULTS

The method was tested with both simulated and real EEG data of 11 participants. From the simulated data, the similarity index between the extracted BCG and the simulated BCG showed the effectiveness of the proposed method in BCG removal. On the other hand, real data were recorded with two conditions, i.e. resting state (eyes closed dataset) and task influenced (event-related potentials (ERPs) dataset). Using qualitative (visual inspection) and quantitative (similarity index, improved normalized power spectrum (INPS) ratio, power spectrum, sample entropy (SE)) evaluation parameters, the assessment results showed that the proposed method can efficiently reduce the BCG artefact while preserving the neuronal signals.

COMPARISON WITH EXISTING METHODS

Compared with conventional methods, namely, average artefact subtraction (AAS), optimal basis set (OBS) and combined independent component analysis and principal component analysis (ICA-PCA), the statistical analyses of the results showed that the proposed method has better performance, and the differences were significant for all quantitative parameters except for the power and sample entropy.

CONCLUSIONS

The proposed method does not require any reference signal, prior information or assumption to extract the BCG artefact. It will be very useful in circumstances where the reference signal is not available.

Improved 7 Tesla resting-state fMRI connectivity measurements by cluster-based modeling of respiratory volume and heart rate effects

Several strategies have been proposed to model and remove physiological noise from resting-state fMRI (rs-fMRI) data, particularly at ultrahigh fields (7 T), including contributions from respiratory volume (RV) and heart rate (HR) signal fluctuations. Recent studies suggest that these contributions are highly variable across subjects and that physiological noise correction may thus benefit from optimization at the subject or even voxel level. Here, we systematically investigated the impact of the degree of spatial specificity (group, subject, newly proposed cluster, and voxel levels) on the optimization of RV and HR models. For each degree of spatial specificity, we measured the fMRI signal variance explained (VE) by each model, as well as the functional connectivity underlying three well-known resting-state networks (RSNs) obtained from the fMRI data after removal of RV+HR contributions. Whole-brain, high-resolution rs-fMRI data were acquired from twelve healthy volunteers at 7 T, while simultaneously recording their cardiac and respiratory signals. Although VE increased with spatial specificity up to the voxel level, the accuracy of functional connectivity measurements improved only up to the cluster level, and subsequently decreased at the voxel level. This suggests that voxelwise modeling over-fits to local fluctuations with no physiological meaning. In conclusion, our results indicate that 7 T rs-fMRI connectivity measurements improve if a cluster-based physiological noise correction approach is employed in order to take into account the individual spatial variability in the HR and RV contributions.

Reference layer adaptive filtering (RLAF) for EEG artifact reduction in simultaneous EEG-fMRI

OBJECTIVE

Simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) combines advantages of both methods, namely high temporal resolution of EEG and high spatial resolution of fMRI. However, EEG quality is limited due to severe artifacts caused by fMRI scanners.

APPROACH

To improve EEG data quality substantially, we introduce methods that use a reusable reference layer EEG cap prototype in combination with adaptive filtering. The first method, reference layer adaptive filtering (RLAF), uses adaptive filtering with reference layer artifact data to optimize artifact subtraction from EEG. In the second method, multi band reference layer adaptive filtering (MBRLAF), adaptive filtering is performed on bandwidth limited sub-bands of the EEG and the reference channels.

MAIN RESULTS

The results suggests that RLAF outperforms the baseline method, average artifact subtraction, in all settings and also its direct predecessor, reference layer artifact subtraction (RLAS), in lower (<35 Hz) frequency ranges. MBRLAF is computationally more demanding than RLAF, but highly effective in all EEG frequency ranges. Effectivity is determined by visual inspection, as well as root-mean-square voltage reduction and power reduction of EEG provided that physiological EEG components such as occipital EEG alpha power and visual evoked potentials (VEP) are preserved. We demonstrate that both, RLAF and MBRLAF, improve VEP quality. For that, we calculate the mean-squared-distance of single trial VEP to the mean VEP and estimate single trial VEP classification accuracies. We found that the average mean-squared-distance is lowest and the average classification accuracy is highest after MBLAF. RLAF was second best.

SIGNIFICANCE

In conclusion, the results suggests that RLAF and MBRLAF are potentially very effective in improving EEG quality of simultaneous EEG-fMRI. Highlights We present a new and reusable reference layer cap prototype for simultaneous EEG-fMRI We introduce new algorithms for reducing EEG artifacts due to simultaneous fMRI The algorithms combine a reference layer and adaptive filtering Several evaluation criteria suggest superior effectivity in terms of artifact reduction We demonstrate that physiological EEG components are preserved.

Physiological noise correction using ECG-derived respiratory signals for enhanced mapping of spontaneous neuronal activity with simultaneous EEG-fMRI

The study of spontaneous brain activity based on BOLD-fMRI may be seriously compromised by the presence of signal fluctuations of non-neuronal origin, most prominently due to cardiac and respiratory mechanisms. Methods used for modeling and correction of the so-called physiological noise usually rely on the concurrent measurement of cardiac and respiratory signals. In simultaneous EEG-fMRI recordings, which are primarily aimed at the study of spontaneous brain activity, the electrocardiogram (ECG) is typically measured as part of the EEG setup but respiratory data are not generally available. Here, we propose to use the ECG-derived respiratory (EDR) signal estimated by Empirical Mode Decomposition (EMD) as a surrogate of the respiratory signal, for retrospective physiological noise correction of typical simultaneous EEG-fMRI data. A physiological noise model based on these physiological signals (P-PNM) complemented with fMRI-derived noise regressors was generated, and evaluated, for 17 simultaneous EEG-fMRI datasets acquired from a group of seven epilepsy patients imaged at 3T. The respiratory components of P-PNM were found to explain BOLD variance significantly in addition to the cardiac components, suggesting that the EDR signal was successfully extracted from the ECG, and P-PNM outperformed an image-based model (I-PNM) in terms of total BOLD variance explained. Further, the impact of the correction using P-PNM on fMRI mapping of patient-specific epileptic networks and the resting-state default mode network (DMN) was assessed in terms of sensitivity and specificity and, when compared with an ICA-based procedure and a standard pre-processing pipeline, P-PNM achieved the best performance. Overall, our results support the feasibility and utility of extracting physiological noise models of the BOLD signal resorting to ECG data exclusively, with substantial impact on the simultaneous EEG-fMRI mapping of resting-state networks, and, most importantly, epileptic networks where sensitivity and specificity are still limited.