Detection and Separation of Event-related Potentials from Multi-Artifacts Contaminated EEG by Means of Independent Component Analysis

Event-related potentials (ERP) is an important type of brain dynamics in human cognition research. However, ERP is often submerged by the spontaneous brain activity EEG, for its relatively tiny scale. Further more, the brain activities collected from scalp electrodes are often inevitably contaminated by several kinds of artifacts, such as blinks, eye movements, muscle noise and power line interference. A new approach to correct these disturbances is presented using independent component analysis (ICA). This technique can effectively detect and extract ERP components from the measured electrodes recordings even if they are heavily contaminated. The results compare favorably to those obtained by parametric modeling. Besides, auto--adaptive projection of decomposed results to ERP components was also given. Through experiments, ICA proves to be highly capable of ERP extraction and S/N ratio improving.     

Automatic correction of artifact from single-trial event-related potentials by blind source separation using second order statistics only

Event-related potentials (ERP) are in general masked by various kinds of artifacts. To attenuate the effects of artifacts, various schemes have been introduced, such as epoch rejection, electro-oculogram (EOG) regression and independent component analysis (ICA). However, none of the existing techniques can automatically remove various kinds of artifacts from a single ERP epoch. EOG regression cannot handle artifacts other than ocular ones. ICA incorporating higher order statistics (HOS) normally requires data with large number of time samples in order that the solution is robust. In this paper we blindly separate the multi-channel ERP into source components by estimating the correlation matrices of the data. Since only second order statistics (SOS) is involved, the process performs well at the single epoch level. Automatic artifact identification is performed in the source domain by introducing objective criteria for various artifacts. Criteria are based on time domain signal amplitude for blink and spurious peak artifact, scalp distribution of signal power for eye movement artifact and power distribution of frequency components for muscle artifact. The correction procedure can be completed by removing the identified artifactual sources from the raw multi-channel ERP.     

A technique to consider mismatches between fMRI and EEG/MEG sources for fMRI-constrained EEG/MEG source imaging: a preliminary simulation study

fMRI-constrained EEG/MEG source imaging can be a powerful tool in studying human brain functions with enhanced spatial and temporal resolutions. Recent studies on the combination of fMRI and EEG/MEG have suggested that fMRI prior information could be readily implemented by simply imposing different weighting factors to cortical sources overlapping with the fMRI activations. It has been also reported, however, that such a hard constraint may cause severe distortions or elimination of meaningful EEG/MEG sources when there are distinct mismatches between the fMRI activations and the EEG/MEG sources. If one wants to obtain the actual EEG/MEG source locations and uses the fMRI prior information as just an auxiliary tool to enhance focality of the distributed EEG/MEG sources, it is reasonable to weaken the strength of fMRI constraint when severe mismatches between fMRI and EEG/MEG sources are observed. The present study suggests an efficient technique to automatically adjust the strength of fMRI constraint according to the mismatch level. The use of the proposed technique rarely affects the results of conventional fMRI-constrained EEG/MEG source imaging if no major mismatch between the two modalities is detected; while the new results become similar to those of typical EEG/MEG source imaging without fMRI constraint if the mismatch level is significant. A preliminary simulation study using realistic EEG signals demonstrated that the proposed technique can be a promising tool to selectively apply fMRI prior information to EEG/MEG source imaging.     

Accounting for microsaccadic artifacts in the EEG using independent component analysis and beamforming

Neuronal activity in the gamma‐band range was long considered a marker of object representation. However, scalp‐recorded EEG activity in this range is contaminated by a miniature saccade‐related muscle artifact. Independent component analysis (ICA) has been proposed as a method of removal of such artifacts. Alternatively, beamforming, a source analysis method in which potential sources of activity across the whole brain are scanned independently through the use of adaptive spatial filters, offers a promising method of accounting for the artifact without relying on its explicit removal. We present here the application of ICA‐based correction to a previously published dataset. Then, using beamforming, we examine the effect of ICA correction on the scalp‐recorded EEG signal and the extent to which genuine activity is recoverable before and after ICA correction. We find that beamforming attributes much of the scalp‐recorded gamma‐band signal before correction to deep frontal sources, likely the eye muscles, which generate the artifact related to each miniature saccade. Beamforming confirms that what is removed by ICA is predominantly this artifactual signal, and that what remains after correction plausibly originates in the visual cortex. Thus, beamforming allows researchers to confirm whether their removal procedures successfully removed the artifact. Our results demonstrate that ICA‐based correction brings about general improvements in signal‐to‐noise ratio suggesting it should be used along with, rather than be replaced by, beamforming.     

Integrated MEG/fMRI Model Validated Using Real Auditory Data

The main objective of this paper is to present methods and results for the estimation of parameters of our proposed integrated magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI) model. We use real auditory MEG and fMRI datasets from 7 normal subjects to estimate the parameters of the model. The MEG and fMRI data were acquired at different times, but the stimulus profile was the same for both techniques. We use independent component analysis (ICA) to extract activation-related signal from the MEG data. The stimulus-correlated ICA component is used to estimate MEG parameters of the model. The temporal and spatial information of the fMRI datasets are used to estimate fMRI parameters of the model. The estimated parameters have reasonable means and standard deviations for all subjects. Goodness of fit of the real data to our model shows the possibility of using the proposed model to simulate realistic datasets for evaluation of integrated MEG/fMRI analysis methods.     

一种独立成分分析的鲁棒算法及其在脑磁图数据分析中的应用

独立成分分析是一种新的信号处理统计方法。被广泛用于各个领域。在信号分析中面临的难题是：源信号的不同特性（既包括超高斯信号又包括亚高斯信号）；未知的独立源数目；传感器信号受到较大的加性噪声污染。针对以上难题，本文提出了一种独立成分分析的鲁棒算法。该方法先对观测数据作预处理，将包含噪声的高维传感器观测信号降维分解到信号子空间和噪声子空间。利用交叉验证法估计出独立源的数目（解决了独立成分分析本身不能确定源数目的缺陷）；然后利用快速稳定的FastICA算法分离独立成分。通过人工合成的数据和实际的脑磁图数据分析。验证了这种方法的功效。     

基于独立分量分析的脑电噪声消除

作为一种新的多元统计处理方法，独立分量分析(ICA)是解决盲源分离(BSS)问题的一个有效手段。在简要分析ICA理论及其算法的基础上，提出将其应用到脑电中的眼电伪迹的去除任务。实际采集的生理信号大多由相互独立的成分线性迭加而成，符合ICA要求源信号统计独立的基本假设。与传统方法相比，ICA这种空间滤波器不受信号频谱混迭的限制，消噪的同时能对有用信号的细节成分做到很好的保留，很大程度上弥补了时频域方法的不足。此外解混矩阵的逆可以用来反映独立源的空间分布模式，具有重要的生理意义。     

新的独立成分分析算法结合主成分分析实现fMRI信号的盲分离

用ICA算法来实现fMRI信号的盲源分离,可以提取出产生fMRI信号的多种源信号。但是在处理过程中存在两个困难:(1)fMRI数据的规模比较大,计算耗时;(2)计算量太大难免产生误差,给结果的分析带来不便。所以我们考虑对数据进行降维,但是如何确定源信号的个数也是一个难题。我们利用信息论的方法来估计源信号的个数,再使用主成分分析对数据进行降维。通过这样的处理,有效地确定了源信号的个数,减少了计算量。然后将一种新的ICA算法(New fixed-point,NewFP)用于处理降维后的数据。最后通过对实际的fMRI信号进行处理,结果表明新算法可以快速有效的分离fMRI信号,且准确性优于FastICA算法。     

ERP differences with vs. without concurrent fMRI.

The acquisition of ERPs concurrently with fMRI in cognitive paradigms is appealing, but technically challenging. Little is known about the effects of the fMRI environment on the time-course and topography of previously documented ERP effects. We examined the replicability of ERP differences in the scanner at the level of individual subjects, using two cognitive paradigms and two statistical procedures. ERP P3 differences found outside the scanner in both paradigms were also robustly detected in the ERPs acquired during fMRI scanning. These P3 effects had equivalent time-courses and scalp topographies inside and outside the scanner. This replication at the level of individual data-sets has implications for the clinical applicability of ERP-fMRI and, more generally, for the quality of scanner recorded ERPs.     

The neural sources of N170: Understanding timing of activation in face‐selective areas

The N170 ERP component has been widely identified as a face‐sensitive neural marker. Despite extensive investigations conducted to examine the neural sources of N170, there are two issues in prior literature: (a) few studies used individualized anatomy as head model for the cortical source analysis of the N170, and (b) the relationship between the N170 and face‐selective regions from fMRI studies is unclear. Here, we addressed these questions by presenting pictures of faces and houses to the same group of healthy adults and recording structural MRI, fMRI, and high‐density ERPs in separate sessions. Source analysis based on the participant's anatomy showed that the middle and posterior fusiform gyri were the primary neural sources for the face‐sensitive aspects of the N170. Source analysis based on regions of interest from the fMRI revealed that the fMRI‐defined fusiform face area was the major contributor to the N170. The current study suggests that the fusiform gyrus is a major neural contributor to the N170 ERP component and provides further insights about the spatiotemporal characteristics of face processing.     

Multimodal Integration of fMRI and EEG Data for High Spatial and Temporal Resolution Analysis of Brain Networks

Two major non-invasive brain mapping techniques, electroencephalography (EEG) and functional magnetic resonance imaging (fMRI), have complementary advantages with regard to their spatial and temporal resolution. We propose an approach based on the integration of EEG and fMRI, enabling the EEG temporal dynamics of information processing to be characterized within spatially well-defined fMRI large-scale networks. First, the fMRI data are decomposed into networks by means of spatial independent component analysis (sICA), and those associated with intrinsic activity and/or responding to task performance are selected using information from the related time-courses. Next, the EEG data over all sensors are averaged with respect to event timing, thus calculating event-related potentials (ERPs). The ERPs are subjected to temporal ICA (tICA), and the resulting components are localized with the weighted minimum norm (WMNLS) algorithm using the task-related fMRI networks as priors. Finally, the temporal contribution of each ERP component in the areas belonging to the fMRI large-scale networks is estimated. The proposed approach has been evaluated on visual target detection data. Our results confirm that two different components, commonly observed in EEG when presenting novel and salient stimuli, respectively, are related to the neuronal activation in large-scale networks, operating at different latencies and associated with different functional processes.     

Surface Laplacian of scalp electrical signals and independent component analysis resolve EMG contamination of electroencephalogram

The serious impact of electromyogram (EMG) contamination of electroencephalogram (EEG) is well recognised. The objective of this research is to demonstrate that combining independent component analysis with the surface Laplacian can eliminate EMG contamination of the EEG, and to validate that this processing does not degrade expected neurogenic signals. The method involves sequential application of ICA, using a manual procedure to identify and discard EMG components, followed by the surface Laplacian. The extent of decontamination is quantified by comparing processed EEG with EMG-free data that was recorded during pharmacologically induced neuromuscular paralysis. The combination of the ICA procedure and the surface Laplacian, with a flexible spherical spline, results in a strong suppression of EMG contamination at all scalp sites and frequencies. Furthermore, the ICA and surface Laplacian procedure does not impair the detection of well-known, cerebral responses; alpha activity with eyes-closed; ERP components (N1, P2) in response to an auditory oddball task; and steady state responses to photic and auditory stimulation. Finally, more flexible spherical splines increase the suppression of EMG by the surface Laplacian. We postulate this is due to ICA enabling the removal of local muscle sources of EMG contamination and the Laplacian transform being insensitive to distant (postural) muscle EMG contamination.     

Cortical Source Localization of Infant Cognition

Neuroimaging techniques such as positron emission topography (PET) and functional magnetic resonance imaging (fMRI) have been utilized with older children and adults to identify cortical sources of perceptual and cognitive processes. However, due to practical and ethical concerns, these techniques cannot be routinely applied to infant participants. An alternative to such neuroimaging techniques appropriate for use with infant participants is high-density electroencephalogram (EEG) recording and cortical source localization techniques. The current article provides an overview of a method developed for such analyses. The method consists of four steps: (1) recording high-density (e.g., 128-channel) EEG. (2) Analysis of individual participant raw segmented data with independent component analysis (ICA). (3) Estimation of equivalent current dipoles (ECDs) that represent cortical sources for the observed ICA component clusters. (4) Calculation of component activations in relation to experimental factors. We discuss an example of research applying this technique to investigate the development of visual attention and recognition memory. We also describe the application of “realistic head modeling” to address some of the current limitations of infant cortical source localization.     

Integrated Analysis of EEG and fMRI Using Sparsity of Spatial Maps

Integration of electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) is an open problem, which has motivated many researches. The most important challenge in EEG-fMRI integration is the unknown relationship between these two modalities. In this paper, we extract the same features (spatial map of neural activity) from both modality. Therefore, the proposed integration method does not need any assumption about the relationship of EEG and fMRI. We present a source localization method from scalp EEG signal using jointly fMRI analysis results as prior spatial information and source separation for providing temporal courses of sources of interest. The performance of the proposed method is evaluated quantitatively along with multiple sparse priors method and sparse Bayesian learning with the fMRI results as prior information. Localization bias and source distribution index are used to measure the performance of different localization approaches with or without a variety of fMRI-EEG mismatches on simulated realistic data. The method is also applied to experimental data of face perception of 16 subjects. Simulation results show that the proposed method is significantly stable against the noise with low localization bias. Although the existence of an extra region in the fMRI data enlarges localization bias, the proposed method outperforms the other methods. Conversely, a missed region in the fMRI data does not affect the localization bias of the common sources in the EEG-fMRI data. Results on experimental data are congruent with previous studies and produce clusters in the fusiform and occipital face areas (FFA and OFA, respectively). Moreover, it shows high stability in source localization against variations in different subjects.     

Evaluating the efficacy of fully automated approaches for the selection of eyeblink ICA components

Independent component analysis (ICA) offers a powerful approach for the isolation and removal of eyeblink artifacts from EEG signals. Manual identification of the eyeblink ICA component by inspection of scalp map projections, however, is prone to error, particularly when nonartifactual components exhibit topographic distributions similar to the blink. The aim of the present investigation was to determine the extent to which automated approaches for selecting eyeblink‐related ICA components could be utilized to replace manual selection. We evaluated popular blink selection methods relying on spatial features (EyeCatch), combined stereotypical spatial and temporal features (ADJUST), and a novel method relying on time series features alone (icablinkmetrics) using both simulated and real EEG data. The results of this investigation suggest that all three methods of automatic component selection are able to accurately identify eyeblink‐related ICA components at or above the level of trained human observers. However, icablinkmetrics, in particular, appears to provide an effective means of automating ICA artifact rejection while at the same time eliminating human errors inevitable during manual component selection and false positive component identifications common in other automated approaches. Based upon these findings, best practices for (a) identifying artifactual components via automated means, and (b) reducing the accidental removal of signal‐related ICA components are discussed.     

Source localization of two evoked magnetic field components using two alternative procedures

Summary The purpose of this study was to compare the relative efficacy of two methods in assessing the location of the sources of the N100 and P200 components of evoked magnetic fields (EMFs) to transient tone stimuli. EMFs to left ear stimulation, containing both components, were recorded over the right hemisphere of six normal subjects. The magnetic scalp distributions calculated at several adjacent time points, covering the duration of each component's peak, were used to estimate the source parameters of each component. Good estimates of the source of both components were obtained from all magnetic field distributions. The averaged spatial parameters derived from all distributions of each component as well as the parameters derived from the distribution that gave the best source estimate for each component were projected onto magnetic resonance images of subject's head. It was found that the source of each component is located on the superior surface of the temporal lobe and that the source of the P200 component is anterior to the N100 source in all subjects using both procedures.     

空间ICA在fMRI数据上的应用与分析

独立成分分析(ICA)技术试图将多维数据分解成若干个相互统计独立的分量。时间ICA和空间ICA都可以用于分析功能核磁共振成像(fMRI)数据。但由于fMRI数据空间维数远远大于时间维数，为计算方便，在分析fMRI数据时。则更多的使用空间ICA方法。本文在单任务激励实验中，利用ICA方法从fMRI数据中分离出若干个与任务相关的独立分量，其中包括与任务相关的恒定分量(CTR)和与任务相关的暂态分量(TTR)；通过将这些独立分量进行空间映射，得到了与任务相关的脑部激活区域。将此结果与SPM的分析比较，得到了一致的结果。在对结果的分析中，我们进一步指出了ICA方法的特点和局限性。     

Fetal source extraction from magnetocardiographic recordings by dependent component analysis

Fetal magnetocardiography (fMCG) has been extensively reported in the literature as a non-invasive, prenatal technique that can be used to monitor various functions of the fetal heart. However, fMCG signals often have low signal-to-noise ratio (SNR) and are contaminated by strong interference from the mother's magnetocardiogram signal. A promising, efficient tool for extracting signals, even under low SNR conditions, is blind source separation (BSS), or independent component analysis (ICA). Herein we propose an algorithm based on a variation of ICA, where the signal of interest is extracted using a time delay obtained from an autocorrelation analysis. We model the system using autoregression, and identify the signal component of interest from the poles of the autocorrelation function. We show that the method is effective in removing the maternal signal, and is computationally efficient. We also compare our results to more established ICA methods, such as FastICA.     

基于独立分量分析的拟牛顿方法在脑电多源定位问题中的应用

在时空源模型的基础上，应用基于独立分量分析的拟牛顿方法进行多源的分离及定位，源分离的过程使得多偶极子的定位问题转化成几个单偶极子的定位，此方法的另一个优点是可以获得独立源的数目。计算机仿真表明：基于独立分量分析的拟牛顿方法在定位精度、计算时间及抗噪性能等方面都要优于传统的非线性优化方法。