EEG source localization: comparative study of classical and neural network methods.

We address the problem of estimating biopotential sources within the brain, based on EEG signals observed on the scalp. This problem, known as the inverse problem of electrophysiology, has no closed-form solution, and requires iterative techniques such as the Levenberg-Marquardt (LM) algorithm. Considering the nonlinear nature of the inverse problem, and the low signal to noise ratio inherent in EEG signals, a backpropagation neural network (BPN) has been recently proposed as a solution. The technique has not been properly compared with classical techniques such as the LM method, or with more recent neural network techniques such as the Radial Basis Function (RBF) network. In this paper, we provide improved strategies based on BPN and consider RBF networks in solving the inverse problem. We compare the performances of BPN, RBF and a hybrid technique with that of the classical LM method.     

Simultaneous Scalp Electroencephalography (EEG), Electromyography (EMG), and Whole-body Segmental Inertial Recording for Multi-modal Neural Decoding

Recent studies support the involvement of supraspinal networks in control of bipedal human walking. Part of this evidence encompasses studies, including our previous work, demonstrating that gait kinematics and limb coordination during treadmill walking can be inferred from the scalp electroencephalogram (EEG) with reasonably high decoding accuracies. These results provide impetus for development of non-invasive brain-machine-interface (BMI) systems for use in restoration and/or augmentation of gait- a primary goal of rehabilitation research. To date, studies examining EEG decoding of activity during gait have been limited to treadmill walking in a controlled environment. However, to be practically viable a BMI system must be applicable for use in everyday locomotor tasks such as over ground walking and turning. Here, we present a novel protocol for non-invasive collection of brain activity (EEG), muscle activity (electromyography (EMG)), and whole-body kinematic data (head, torso, and limb trajectories) during both treadmill and over ground walking tasks. By collecting these data in the uncontrolled environment insight can be gained regarding the feasibility of decoding unconstrained gait and surface EMG from scalp EEG.     

Topographic EEG mapping of the relaxation response

The purpose of this study was to assess the central nervous system effects of the relaxation response (RR) in novice subjects using a controlled, within-subjects design and topographic EEG mapping as the dependent measure. Twenty subjects listened to a RR and control audiotape presented in a counterbalanced order while EEG was recorded from 14 scalp locations. TheRR condition produced greater (p<.0164) reductions in frontal EEG beta activity relative to the control condition. No significant differences were observed for any other frequency band or scalp region. These findings suggest that elicitation of the RR produces significant reductions in cortical activation in anterior brain regions in novice subjects.     

From Oscillatory Transcranial Current Stimulation to Scalp EEG Changes: A Biophysical and Physiological Modeling Study

Both biophysical and neurophysiological aspects need to be considered to assess the impact of electric fields induced by transcranial current stimulation (tCS) on the cerebral cortex and the subsequent effects occurring on scalp EEG. The objective of this work was to elaborate a global model allowing for the simulation of scalp EEG signals under tCS. In our integrated modeling approach, realistic meshes of the head tissues and of the stimulation electrodes were first built to map the generated electric field distribution on the cortical surface. Secondly, source activities at various cortical macro-regions were generated by means of a computational model of neuronal populations. The model parameters were adjusted so that populations generated an oscillating activity around 10 Hz resembling typical EEG alpha activity. In order to account for tCS effects and following current biophysical models, the calculated component of the electric field normal to the cortex was used to locally influence the activity of neuronal populations. Lastly, EEG under both spontaneous and tACS-stimulated (transcranial sinunoidal tCS from 4 to 16 Hz) brain activity was simulated at the level of scalp electrodes by solving the forward problem in the aforementioned realistic head model. Under the 10 Hz-tACS condition, a significant increase in alpha power occurred in simulated scalp EEG signals as compared to the no-stimulation condition. This increase involved most channels bilaterally, was more pronounced on posterior electrodes and was only significant for tACS frequencies from 8 to 12 Hz. The immediate effects of tACS in the model agreed with the post-tACS results previously reported in real subjects. Moreover, additional information was also brought by the model at other electrode positions or stimulation frequency. This suggests that our modeling approach can be used to compare, interpret and predict changes occurring on EEG with respect to parameters used in specific stimulation configurations.     

不同状态下脑电图复杂性探索

Ｌｅｍｐｅｌ－Ｚｉｖ所定义的有限序列的复杂性反映了给定序列随其长度的增长出现新模式的速率,事实上它反映了序列接近随机的程度。将该复杂性度量运用于脑电分析,旨在克服分数维方法的缺陷。文中计算了八种实验条件下脑电图的复杂度,涉及看、听、休息和心算等基本的大脑功能状态,１３个被试的１６导数据被用于计算分析．结果显示了复杂度在不同电极位置及实验条件下都有变化,睁眼状态的复杂度高于闭眼,而施加任务时有额部大脑活动区域复杂度降低的现象。同时复杂度也提供了一些研究大脑高级认知活动的新思路。     

High Resolution Topography of Age-Related Changes in Non-Rapid Eye Movement Sleep Electroencephalography

Sleeping brain activity reflects brain anatomy and physiology. The aim of this study was to use high density (256 channel) electroencephalography (EEG) during sleep to characterize topographic changes in sleep EEG power across normal aging, with high spatial resolution. Sleep was evaluated in 92 healthy adults aged 18–65 years old using full polysomnography and high density EEG. After artifact removal, spectral power density was calculated for standard frequency bands for all channels, averaged across the NREM periods of the first 3 sleep cycles. To quantify topographic changes with age, maps were generated of the Pearson’s coefficient of the correlation between power and age at each electrode. Significant correlations were determined by statistical non-parametric mapping. Absolute slow wave power declined significantly with increasing age across the entire scalp, whereas declines in theta and sigma power were significant only in frontal regions. Power in fast spindle frequencies declined significantly with increasing age frontally, whereas absolute power of slow spindle frequencies showed no significant change with age. When EEG power was normalized across the scalp, a left centro-parietal region showed significantly less age-related decline in power than the rest of the scalp. This partial preservation was particularly significant in the slow wave and sigma bands. The effect of age on sleep EEG varies substantially by region and frequency band. This non-uniformity should inform the design of future investigations of aging and sleep. This study provides normative data on the effect of age on sleep EEG topography, and provides a basis from which to explore the mechanisms of normal aging as well as neurodegenerative disorders for which age is a risk factor.     

Electric Field Encephalography as a Tool for Functional Brain Research: A Modeling Study

We introduce the notion of Electric Field Encephalography (EFEG) based on measuring electric fields of the brain and demonstrate, using computer modeling, that given the appropriate electric field sensors this technique may have significant advantages over the current EEG technique. Unlike EEG, EFEG can be used to measure brain activity in a contactless and reference-free manner at significant distances from the head surface. Principal component analysis using simulated cortical sources demonstrated that electric field sensors positioned 3 cm away from the scalp and characterized by the same signal-to-noise ratio as EEG sensors provided the same number of uncorrelated signals as scalp EEG. When positioned on the scalp, EFEG sensors provided 2–3 times more uncorrelated signals. This significant increase in the number of uncorrelated signals can be used for more accurate assessment of brain states for non-invasive brain-computer interfaces and neurofeedback applications. It also may lead to major improvements in source localization precision. Source localization simulations for the spherical and Boundary Element Method (BEM) head models demonstrated that the localization errors are reduced two-fold when using electric fields instead of electric potentials. We have identified several techniques that could be adapted for the measurement of the electric field vector required for EFEG and anticipate that this study will stimulate new experimental approaches to utilize this new tool for functional brain research.     

Effects of dipole position,orientation and noise on the accuracy of EEG source localization

Background  

The electroencephalogram (EEG) reflects the electrical activity in the brain on the surface of scalp. A major challenge in this field is the localization of sources in the brain responsible for eliciting the EEG signal measured at the scalp. In order to estimate the location of these sources, one must correctly model the sources, i.e., dipoles, as well as the volume conductor in which the resulting currents flow. In this study, we investigate the effects of dipole depth and orientation on source localization with varying sets of simulated random noise in 4 realistic head models.     

Effects of cavities on EEG dipole localization and their relations with surface electrode positions

Effects of cavities in the human head on EEG dipole localization have been investigated by computer simulation. The human head is represented by a homogeneous spherical conductor including an eccentric spherical cavity which approximates effects of actual cavities inside the head. The homogeneous sphere model is used for assessing the effects caused by neglecting the cavity in the volume conductor model in the inverse dipole fitting procedure. Four electrode configurations have been examined to investigate their relation to the EEG inverse dipole solution. After examination of 2520 dipoles in the brain, the effects of cavities in the human head are found to be negligible when the dipole is located in the cortex or in the subcortex. When the dipole is located in the brain stem, the EEG inverse dipole solution is strongly affected by the cavity and is sensitive to the electrode configuration on the scalp. The EEG inverse dipole solution in the deep brain is sensitive to inhomogeneity in the lower part of the head when a single positive or negative potential pole is observed by the electrodes on the scalp, and at the same time is sensitive to the extent of the scalp covered by the electrodes. In conclusion, the electrodes should cover as much of the upper scalp as possible for deep source localization.     

Investigating Social Cognition in Infants and Adults Using Dense Array Electroencephalography (dEEG)

Dense array electroencephalography (_dEEG), which provides a non-invasive window for measuring brain activity and a temporal resolution unsurpassed by any other current brain imaging technology^1,2, is being used increasingly in the study of social cognitive functioning in infants and adults. While _dEEG is enabling researchers to examine brain activity patterns with unprecedented levels of sensitivity, conventional EEG recording systems continue to face certain limitations, including 1) poor spatial resolution and source localization^3,4,2) the physical discomfort for test subjects of enduring the individual application of numerous electrodes to the surface of the scalp, and 3) the complexity for researchers of learning to use multiple software packages to collect and process data. Here we present an overview of an established methodology that represents a significant improvement on conventional methodologies for studying EEG in infants and adults. Although several analytical software techniques can be used to establish indirect indices of source localization to improve the spatial resolution of _dEEG, the HydroCel Geodesic Sensor Net (HCGSN) by Electrical Geodesics, Inc. (EGI), a dense sensory array that maintains equal distances among adjacent recording electrodes on all surfaces of the scalp, further enhances spatial resolution^4,5,6 compared to standard _dEEG systems. The sponge-based HCGSN can be applied rapidly and without scalp abrasion, making it ideal for use with adults^7,8, children^9,10,11, and infants¹², in both research and clinical^{4,5,6,13,14,15} settings. This feature allows for considerable cost and time savings by decreasing the average net application time compared to other _dEEG systems. Moreover, the HCGSN includes unified, seamless software applications for all phases of data, greatly simplifying the collection, processing, and analysis of _dEEG data.The HCGSN features a low-profile electrode pedestal, which, when filled with electrolyte solution, creates a sealed microenvironment and an electrode-scalp interface. In all Geodesic _dEEG systems, EEG sensors detect changes in voltage originating from the participant''s scalp, along with a small amount of electrical noise originating from the room environment. Electrical signals from all sensors of the Geodesic sensor net are received simultaneously by the amplifier, where they are automatically processed, packaged, and sent to the data-acquisition computer (DAC). Once received by the DAC, scalp electrical activity can be isolated from artifacts for analysis using the filtering and artifact detection tools included in the EGI software. Typically, the HCGSN can be used continuously for only up to two hours because the electrolyte solution dries out over time, gradually decreasing the quality of the scalp-electrode interface.In the Parent-Infant Research Lab at the University of Toronto, we are using _dEEG to study social cognitive processes including memory, emotion, goals, intentionality, anticipation, and executive functioning in both adult and infant participants.     

Electronic evaluation for video commercials by impression index

How to evaluate the effect of commercials is significantly important in neuromarketing. In this paper, we proposed an electronic way to evaluate the influence of video commercials on consumers by impression index. The impression index combines both the memorization and attention index during consumers observing video commercials by tracking the EEG activity. It extracts features from scalp EEG to evaluate the effectiveness of video commercials in terms of time–frequency-space domain. And, the general global field power was used as an impression index for evaluation of video commercial scenes as time series. Results of experiment demonstrate that the proposed approach is able to track variations of the cerebral activity related to cognitive task such as observing video commercials, and help to judge whether the scene in video commercials is impressive or not by EEG signals.     

EEG and implanted sources in the brain

Localisation procedures are based on models of the EEG that are relatively simple. The models are based on assumptions and choices of parameters that can be mistaken. Thus, it is crucial to validate the localisation procedures used in EEG. One of the options is to use the data obtained with electrodes that are implanted within the brain of an epileptic patient as part of the pre-surgical evaluation. When one of two neighbouring electrodes is used as a current source and the other as a current sink this can be regarded as a current dipole. The current injected has to be below the threshold for activation of cells. The position of this dipole can be deduced from magnetic resonance or X-ray images. The current dipole gives rise to a potential distribution at the scalp that can be measured by EEG. The measurements can be compared with the potential distribution that is calculated in a forward computation. Another method is to use the measured potential at the scalp to localize the source and to compare the result with the actual position of the dipole. In this paper the measured potential distributions at the scalp due to implanted dipoles were used to evaluate different volume conductor models. Since intracerebral and subdural electrodes were introduced through trephine holes over the fronto-central areas, and the diameter of the holes was rather large, approximately 23 mm, special effort was put into modelling the skull. Two important assumptions could be validated in this study: the electric currents within the head are Ohmic and a dipole can be used to model the induced electric activity of pairs of contacts on subdural electrodes or intra cerebral electrodes.     

Patient-Specific Epileptic Seizure Prediction in Long-Term Scalp EEG Signal Using Multivariate Statistical Process Control

An accurate epileptic seizure prediction algorithm can alleviate the problem and reduce risks in the life of a patient suffering from epilepsy. The main motive of this work is to propose a model which can predict seizures well in advance of its occurrence. Multivariate statistical process control (MSPC) has been used for seizure predictions in long-term scalp EEG signal. It has been observed that excessive neuronal activity in the preictal period of seizure changes the electrical characteristic from chaotic to rhythmic behavior. These changes have been utilized for prediction. Eight temporal based features are used for predicting the seizures by using multivariate statistical process control, which is widely known as an anomaly monitoring method. 90 seizures from the CHB-MIT EEG data of ten patients are analyzed.ResultThe results of the proposed method demonstrated that 80 seizures out of 90 in preictal period were correctly predicted prior to the seizure onset, thereby giving a sensitivity of 88.89%. The false positive rate is observed to 0.39 per hour.ConclusionThis study proposed a temporal based patient-specific epileptic seizure prediction method using MSPC in long-term scalp EEG signals. It also provides the possibility of realizing an EEG-based epileptic seizure prediction system which requires less computational power.SignificanceThe proposed method does not require preictal data for modeling. The extracted features are computationally easy. The tested result shows good accuracy on the CHB-MIT data base.     

Transient induced gamma-band response in EEG as a manifestation of miniature saccades

The induced gamma-band EEG response (iGBR) recorded on the scalp is widely assumed to reflect synchronous neural oscillation associated with object representation, attention, memory, and consciousness. The most commonly reported EEG iGBR is a broadband transient increase in power at the gamma range approximately 200-300 ms following stimulus onset. A conspicuous feature of this iGBR is the trial-to-trial poststimulus latency variability, which has been insufficiently addressed. Here, we show, using single-trial analysis of concomitant EEG and eye tracking, that this iGBR is tightly time locked to the onset of involuntary miniature eye movements and reflects a saccadic "spike potential." The time course of the iGBR is related to an increase in the rate of saccades following a period of poststimulus saccadic inhibition. Thus, whereas neuronal gamma-band oscillations were shown conclusively with other methods, the broadband transient iGBR recorded by scalp EEG reflects properties of miniature saccade dynamics rather than neuronal oscillations.     

Toward electrocorticographic control of a dexterous upper limb prosthesis: building brain-machine interfaces

One of the most exciting and compelling areas of research and development is building brain machine interfaces (BMIs) for controlling prosthetic limbs. Prosthetic limb technology is advancing rapidly, and the modular prosthetic limb (MPL) of the Johns Hopkins University/ Applied Physics Laboratory (JHU/APL) permits actuation with 17 degrees of freedom in 26 articulating joints. There are many signals from the brain that can be leveraged, including the spiking rates of neurons in the cortex, electrocorticographic (ECoG) signals from the surface of the cortex, and electroencephalographic (EEG) signals from the scalp. Unlike microelectrodes that record spikes, ECoG does not penetrate the cortex and has a higher spatial specificity, signal-to-noise ratio, and bandwidth than EEG signals. We have implemented an ECoG-based system for controlling the MPL in the Johns Hopkins Hospital Epilepsy Monitoring Unit, where patients are implanted with ECoG electrode grids for clinical seizure mapping and asked to perform various recorded finger or grasp movements. We have shown that low-frequency local motor potentials (LMPs) and ECoG power in the high gamma frequency (70,150 Hz) range correlate well with grasping parameters, and they stand out as good candidate features for closed-loop control of the MPL.     

Coherence and phase locking in the scalp EEG and between LORETA model sources, and microstates as putative mechanisms of brain temporo-spatial functional organization.

Brain electric mechanisms of temporary, functional binding between brain regions are studied using computation of scalp EEG coherence and phase locking, sensitive to time differences of few milliseconds. However, such results if computed from scalp data are ambiguous since electric sources are spatially oriented. Non-ambiguous results can be obtained using calculated time series of strength of intracerebral model sources. This is illustrated applying LORETA modeling to EEG during resting and meditation. During meditation, time series of LORETA model sources revealed a tendency to decreased left-right intracerebral coherence in the delta band, and to increased anterior-posterior intracerebral coherence in the theta band. An alternate conceptualization of functional binding is based on the observation that brain electric activity is discontinuous, i.e., that it occurs in chunks of up to about 100 ms duration that are detectable as quasi-stable scalp field configurations of brain electric activity, called microstates. Their functional significance is illustrated in spontaneous and event-related paradigms, where microstates associated with imagery- versus abstract-type mentation, or while reading positive versus negative emotion words showed clearly different regions of cortical activation in LORETA tomography. These data support the concept that complete brain functions of higher order such as a momentary thought might be incorporated in temporal chunks of processing in the range of tens to about 100 ms as quasi-stable brain states; during these time windows, subprocesses would be accepted as members of the ongoing chunk of processing.     

Only Low Frequency Event-Related EEG Activity Is Compromised in Multiple Sclerosis: Insights from an Independent Component Clustering Analysis

Cognitive impairment (CI), often examined with neuropsychological tests such as the Paced Auditory Serial Addition Test (PASAT), affects approximately 65% of multiple sclerosis (MS) patients. The P3b event-related potential (ERP), evoked when an infrequent target stimulus is presented, indexes cognitive function and is typically compared across subjects'' scalp electroencephalography (EEG) data. However, the clustering of independent components (ICs) is superior to scalp-based EEG methods because it can accommodate the spatiotemporal overlap inherent in scalp EEG data. Event-related spectral perturbations (ERSPs; event-related mean power spectral changes) and inter-trial coherence (ITCs; event-related consistency of spectral phase) reveal a more comprehensive overview of EEG activity. Ninety-five subjects (56 MS patients, 39 controls) completed visual and auditory two-stimulus P3b event-related potential tasks and the PASAT. MS patients were also divided into CI and non-CI groups (n = 18 in each) based on PASAT scores. Data were recorded from 128-scalp EEG channels and 4 IC clusters in the visual, and 5 IC clusters in the auditory, modality were identified. In general, MS patients had significantly reduced ERSP theta power versus controls, and a similar pattern was observed for CI vs. non-CI MS patients. The ITC measures were also significantly different in the theta band for some clusters. The finding that MS patients had reduced P3b task-related theta power in both modalities is a reflection of compromised connectivity, likely due to demyelination, that may have disrupted early processes essential to P3b generation, such as orientating and signal detection. However, for posterior sources, MS patients had a greater decrease in alpha power, normally associated with enhanced cognitive function, which may reflect a compensatory mechanism in response to the compromised early cognitive processing.     

Independent EEG sources are dipolar

Independent component analysis (ICA) and blind source separation (BSS) methods are increasingly used to separate individual brain and non-brain source signals mixed by volume conduction in electroencephalographic (EEG) and other electrophysiological recordings. We compared results of decomposing thirteen 71-channel human scalp EEG datasets by 22 ICA and BSS algorithms, assessing the pairwise mutual information (PMI) in scalp channel pairs, the remaining PMI in component pairs, the overall mutual information reduction (MIR) effected by each decomposition, and decomposition 'dipolarity' defined as the number of component scalp maps matching the projection of a single equivalent dipole with less than a given residual variance. The least well-performing algorithm was principal component analysis (PCA); best performing were AMICA and other likelihood/mutual information based ICA methods. Though these and other commonly-used decomposition methods returned many similar components, across 18 ICA/BSS algorithms mean dipolarity varied linearly with both MIR and with PMI remaining between the resulting component time courses, a result compatible with an interpretation of many maximally independent EEG components as being volume-conducted projections of partially-synchronous local cortical field activity within single compact cortical domains. To encourage further method comparisons, the data and software used to prepare the results have been made available (http://sccn.ucsd.edu/wiki/BSSComparison).     

Electroencephalographic imaging of higher brain function.

High temporal resolution is necessary to resolve the rapidly changing patterns of brain activity that underlie mental function. Electroencephalography (EEG) provides temporal resolution in the millisecond range. However, traditional EEG technology and practice provide insufficient spatial detail to identify relationships between brain electrical events and structures and functions visualized by magnetic resonance imaging or positron emission tomography. Recent advances help to overcome this problem by recording EEGs from more electrodes, by registering EEG data with anatomical images, and by correcting the distortion caused by volume conduction of EEG signals through the skull and scalp. In addition, statistical measurements of sub-second interdependences between EEG time-series recorded from different locations can help to generate hypotheses about the instantaneous functional networks that form between different cortical regions during perception, thought and action. Example applications are presented from studies of language, attention and working memory. Along with its unique ability to monitor brain function as people perform everyday activities in the real world, these advances make modern EEG an invaluable complement to other functional neuroimaging modalities.     

Des sources neuronales aux signaux EEG : modèle et interprétation des activités épileptiques

This work deals with the interpretation of electrophysiological patients recorded in epileptic patients candidate to surgery. This issue is addressed through a physiologically relevant model for the generation of scalp and intracerebral electroencephalographic (EEG) signals. The proposed model is based on a spatiotemporal representation of the sources of brain activity, which combines a distributed dipole source model and a model of coupled neuronal populations. Signals recorded by sensors (scalp and intracerebral) are then computed by solving the forward problem in the head volume conductor. In this paper, the EEG generation model is used to study the influence of some source-related parameters (spatial extent, position, synchronization) on simulated signals, during epileptic transient activity (interictal spikes). Results show that the model allows for studying, on the one hand, the relationship between the spatiotemporal organization of neuronal sources and the properties of the observed signals and, on the other hand, the relationship between surface and depth EEG signals.