The EEG microstate topography is predominantly determined by intracortical sources in the alpha band.

Human brain electric activity can be measured at high temporal and fairly good spatial resolution via electroencephalography (EEG). The EEG microstate analysis is an increasingly popular method used to investigate this activity at a millisecond resolution by segmenting it into quasi-stable states of approximately 100 ms duration. These so-called EEG microstates were postulated to represent atoms of thoughts and emotions and can be classified into four classes of topographies A through D, which explain up to 90% of the variance of continuous EEG. The present study investigated whether these topographies are primarily driven by alpha activity originating from the posterior cingulate cortex (all topographies), left and right posterior cortices, and the anterior cingulate cortex (topographies A, B, and C, respectively). We analyzed two 64-channel resting state EEG datasets (N = 61 and N = 78) of healthy participants. Sources of head-surface signals were determined via exact low resolution electromagnetic tomography (eLORETA). The Hilbert transformation was applied to identify instantaneous source strength of four EEG frequency bands (delta through beta). These source strength values were averaged for each participant across time periods belonging to a particular microstate. For each dataset, these averages of the different microstate classes were compared for each voxel. Consistent differences across datasets were identified via a conjunction analysis. The intracortical strength and spatial distribution of alpha band activity mainly determined whether a head-surface topography of EEG microstate class A, B, C, or D was induced. EEG microstate class C was characterized by stronger alpha activity compared to all other classes in large portions of the cortex. Class A was associated with stronger left posterior alpha activity than classes B and D, and class B was associated with stronger right posterior alpha activity than A and D. Previous results indicated that EEG microstate dynamics reflect a fundamental mechanism of the human brain that is altered in different mental states in health and disease. They are characterized by systematic transitions between four head-surface topographies, the EEG microstate classes. Our results show that intra-cortical alpha oscillations, which likely reflect decreased cortical excitability, primarily account for the emergence of these classes. We suggest that microstate class dynamics reflect transitions between four global attractor states that are characterized by selective inhibition of specific intra-cortical regions.

The functional significance of EEG microstates--Associations with modalities of thinking.

The momentary, global functional state of the brain is reflected by its electric field configuration. Cluster analytical approaches consistently extracted four head-surface brain electric field configurations that optimally explain the variance of their changes across time in spontaneous EEG recordings. These four configurations are referred to as EEG microstate classes A, B, C, and D and have been associated with verbal/phonological, visual, subjective interoceptive-autonomic processing, and attention reorientation, respectively. The present study tested these associations via an intra-individual and inter-individual analysis approach. The intra-individual approach tested the effect of task-induced increased modality-specific processing on EEG microstate parameters. The inter-individual approach tested the effect of personal modality-specific parameters on EEG microstate parameters. We obtained multichannel EEG from 61 healthy, right-handed, male students during four eyes-closed conditions: object-visualization, spatial-visualization, verbalization (6 runs each), and resting (7 runs). After each run, we assessed participants' degrees of object-visual, spatial-visual, and verbal thinking using subjective reports. Before and after the recording, we assessed modality-specific cognitive abilities and styles using nine cognitive tests and two questionnaires. The EEG of all participants, conditions, and runs was clustered into four classes of EEG microstates (A, B, C, and D). RMANOVAs, ANOVAs and post-hoc paired t-tests compared microstate parameters between conditions. TANOVAs compared microstate class topographies between conditions. Differences were localized using eLORETA. Pearson correlations assessed interrelationships between personal modality-specific parameters and EEG microstate parameters during no-task resting. As hypothesized, verbal as opposed to visual conditions consistently affected the duration, occurrence, and coverage of microstate classes A and B. Contrary to associations suggested by previous reports, parameters were increased for class A during visualization, and class B during verbalization. In line with previous reports, microstate D parameters were increased during no-task resting compared to the three internal, goal-directed tasks. Topographic differences between conditions included particular sub-regions of components of the metabolic default mode network. Modality-specific personal parameters did not consistently correlate with microstate parameters except verbal cognitive style which correlated negatively with microstate class A duration and positively with class C occurrence. This is the first study that aimed to induce EEG microstate class parameter changes based on their hypothesized functional significance. Beyond the associations of microstate classes A and B with visual and verbal processing, respectively, our results suggest that a finely-tuned interplay between all four EEG microstate classes is necessary for the continuous formation of visual and verbal thoughts. Our results point to the possibility that the EEG microstate classes may represent the head-surface measured activity of intra-cortical sources primarily exhibiting inhibitory functions. However, additional studies are needed to verify and elaborate on this hypothesis.

Brain areas and time course of emotional processing.

The aims of the present study were to identify brain regions involved in emotional processing as well as to follow the time sequence of these processes in the millisecond-range resolution using low resolution brain electromagnetic tomography (LORETA). Different emotional (happy, sad, angry, fearful, and disgust) and neutral faces were presented to 17 healthy, right-handed volunteers on a computer screen while 25-channel EEG data were recorded. Subjects were instructed to generate the same emotion as shown in the presented faces. Event-related potentials (ERPs) were computed for each emotion and neutral condition, and analyzed as sequences of potential distribution maps. Paired topographic analysis of variance tests of the ERP maps identified time segments of significant differences between responses to emotional and neutral faces. For these significant segments, statistical analyses of functional LORETA images were performed to identify active brain regions for the different emotions. Significant differences occurred in different time segments within the first 500 ms after stimulus onset. Each emotional condition showed specific activation patterns in different brain regions, changing over time. In the majority of significant time segments, activation was highest in the right frontal areas. Strongest activation was found in the happy, sad, and disgust conditions in extended fronto-temporal areas. Happy, sad, and disgust conditions also produced earlier and more widely distributed differences than anger and fear. Our findings are in good agreement with other brain-imaging studies (PET/fMRI). But unlike other imaging techniques, LORETA allows to follow the time sequence in the millisecond-range resolution.

Non-invasive localization of P300 sources in normal aging and age-associated memory impairment.

Cognitive event-related potentials were recorded from 17 EEG leads in an auditory two-tone paradigm in 43 patients aged 51-79 years with the diagnosis of age-associated memory impairment (AAMI), in age-and sex-matched normal controls and in a control group 10 years older than the AAMI patients. In addition to P300 latencies, amplitudes and topographies, three-dimensional current density distribution utilizing low-resolution brain electromagnetic tomography (LORETA) was computed. P300 latency was delayed and P300 amplitude was reduced in both AAMI and older subjects. Topographically this amplitude reduction was most pronounced frontally. LORETA revealed medial (frontal and parietal) and lateral (dorso- and ventrolateral prefrontal, middle/superior temporal, posterior superior temporal/inferior parietal) sources. Significant reductions in LORETA source strength in normal aging and in AAMI were found mainly medially frontally, right dorsolaterally prefrontally and right inferiorly parietally. Since these anatomically highly interconnected brain regions in the right hemisphere are part of a network associated with sustained attention, the results speak for a decline in attentional resource capacity in AAMI patients and elderly subjects.

Quantitative self-organizing maps for clustering electron tomograms.

Tomography emerges as a powerful methodology for determining the complex architectures of biological specimens that are better regarded from the structural point of view as singular entities. However, once the structure of a sufficiently large number of singular specimens is solved, quite possibly structural patterns start to emerge. This latter situation is addressed here, where the clustering of a set of 3D reconstructions using a novel quantitative approach is presented. In general terms, we propose a new variant of a self-organizing neural network for the unsupervised classification of 3D reconstructions. The novelty of the algorithm lies in its rigorous mathematical formulation that, starting from a large set of noisy input data, finds a set of "representative" items, organized onto an ordered output map, such that the probability density of this set of representative items resembles at its possible best the probability density of the input data. In this study, we evaluate the feasibility of application of the proposed neural approach to the problem of identifying similar 3D motifs within tomograms of insect flight muscle. Our experimental results prove that this technique is suitable for this type of problem, providing the electron microscopy community with a new tool for exploring large sets of tomogram data to find complex patterns.

Standardized low-resolution brain electromagnetic tomography (sLORETA): technical details.

Scalp electric potentials (electroencephalograms) and extracranial magnetic fields (magnetoencephalograms) are due to the primary (impressed) current density distribution that arises from neuronal postsynaptic processes. A solution to the inverse problem--the computation of images of electric neuronal activity based on extracranial measurements--would provide important information on the time-course and localization of brain function. In general, there is no unique solution to this problem. In particular, an instantaneous, distributed, discrete, linear solution capable of exact localization of point sources is of great interest, since the principles of linearity and superposition would guarantee its trustworthiness as a functional imaging method, given that brain activity occurs in the form of a finite number of distributed hot spots. Despite all previous efforts, linear solutions, at best, produced images with systematic nonzero localization errors. A solution reported here yields images of standardized current density with zero localization error. The purpose of this paper is to present the technical details of the method, allowing researchers to test, check, reproduce and validate the new method.

Structural and energetic processes related to P300: LORETA findings in depression and effects of antidepressant drugs.

Noninvasive electrophysiological neuroimaging applied to cognitive components of event-related potentials (ERPs) may differentiate between structural and energetic processes related to information processing. The structural level, revealed by the location of the local maxima of the current source density distribution, describes the time-dependent network of activated brain areas. The magnitude of the source strength, a measure of the energetic component, describes the allocation of processing resources. ERPs were recorded in an odd-ball paradigm and low-resolution brain electromagnetic tomography (LORETA) was applied for standard and target ERP components. In a group of 60 menopausal depressed patients of 45-60 years of age, reduced P300 source strength was observed bilaterally, temporally and medially prefrontally reaching to rostal parts of the anterior cingulate, compared with 29 age-matched controls. In a double-blind, placebo-controlled study, 2 mg of the antidepressant citalopram induced a significant increase of P300 source strength in the (left) prefrontal cortex and precuneus compared with placebo, reaching to the posterior cingulate. Similar increases were observed after 800 mg S-adenosyl-L-methionine (SAMe) administered intravenously in ten young healthy subjects aged 22-33, and they were even more pronounced in ten elderly healthy subjects aged 56-71. Thus, ERP-tomography identified changes in energetic sources in brain areas predominantly involved in depression and in antidepressant action.

EEG topography and tomography in diagnosis and treatment of mental disorders: evidence for a key-lock principle.

Clinically well-defined diagnostic subgroups of mental disorders, such as schizophrenia with predominantly plus and minus symptomatology, major depression, generalized anxiety disorder, agoraphobia, obsessive-compulsive disorder, multiinfarct dementia, senile dementia of the Alzheimer type and alcohol dependence, show electroencephalogram (EEG) maps that differ statistically both from each other and from normal controls. Representative drugs of the main psychopharmacological classes, such as sedative and nonsedative neuroleptics and antidepressants, tranquilizers, hypnotics, psychostimulants and cognition-enhancing drugs, induce significant and typical changes to normal human brain function compared with placebo, in which many variables are opposite to the above-mentioned differences between psychiatric patients and normal controls. Thus, by considering these differences between psychotropic drugs and placebo in normal subjects, as well as between mental disorder patients and normal controls, it may be possible to choose the optimum drug for a specific patient according to a key-lock principle, since the drug should normalize the deviant brain function. This is supported by low-resolution brain electromagnetic tomography (LORETA), which identifies brain regions affected by psychiatric disorders and psychotropic drugs.

Classification and evaluation of the pharmacodynamics of psychotropic drugs by single-lead pharmaco-EEG, EEG mapping and tomography (LORETA).

Utilizing computer-assisted quantitative analyses of human scalp-recorded electroencephalogram (EEG) in combination with certain statistical procedures (quantitative pharmaco-EEG) and mapping techniques (pharmaco-EEG mapping), it is possible to classify psychotropic substances and objectively evaluate their bioavailability at the target organ: the human brain. Specifically, one may determine at an early stage of drug development whether a drug is effective on the central nervous system (CNS) compared with placebo, what its clinical efficacy will be like, at which dosage it acts, when it acts and the equipotent dosages of different galenic formulations. Pharmaco-EEG profiles and maps of neuroleptics, antidepressants, tranquilizers, hypnotics, psychostimulants and nootropics/cognition-enhancing drugs will be described in this paper. Methodological problems, as well as the relationships between acute and chronic drug effects, alterations in normal subjects and patients, CNS effects, therapeutic efficacy and pharmacokinetic and pharmacodynamic data will be discussed. In recent times, imaging of drug effects on the regional brain electrical activity of healthy subjects by means of EEG tomography such as low-resolution electromagnetic tomography (LORETA) has been used for identifying brain areas predominantly involved in psychopharmacological action. This will be demonstrated for the representative drugs of the four main psychopharmacological classes, such as 3 mg haloperidol for neuroleptics, 20 mg citalopram for antidepressants, 2 mg lorazepam for tranquilizers and 20 mg methylphenidate for psychostimulants. LORETA demonstrates that these psychopharmacological classes affect brain structures differently.

Functional imaging with low-resolution brain electromagnetic tomography (LORETA): a review.

This paper reviews several recent publications that have successfully used the functional brain imaging method known as LORETA. Emphasis is placed on the electrophysiological and neuroanatomical basis of the method, on the localization properties of the method, and on the validation of the method in real experimental human data. Papers that criticize LORETA are briefly discussed. LORETA publications in the 1994-1997 period based localization inference on images of raw electric neuronal activity. In 1998, a series of papers appeared that based localization inference on the statistical parametric mapping methodology applied to high-time resolution LORETA images. Starting in 1999, quantitative neuroanatomy was added to the methodology, based on the digitized Talairach atlas provided by the Brain Imaging Centre, Montreal Neurological Institute. The combination of these methodological developments has placed LORETA at a level that compares favorably to the more classical functional imaging methods, such as PET and fMRI.

Perceptual and cognitive event-related potentials in neuropsychopharmacology: methodological aspects and clinical applications (pharmaco-ERP topography and tomography).

Middle latency and late components of event-related brain potential (ERPs) are closely related to perceptual and cognitive information processing, respectively. In a double-blind, placebo-controlled study, the acute effects of lorazepam (2 mg), haloperidol (3 mg), methylphenidate (20 mg) and citalopram (20 mg) on ERP latencies, amplitudes, topographies and tomographies were investigated in 20 healthy subjects of 23-34 years of age. After automatic artifact minimization and rejection, standard N1 and P2 and target N2 and P300 components were determined. The tranquilizer lorazepam prolonged P300 latency, which indicates an impairment of stimulus evaluation time. Low-resolution brain electromagnetic tomography (LORETA) revealed decreases in N1 and P300 source strength in those brain regions with relevant generators of these components, which reflects impairments of attentional and cognitive processing resources. The neuroleptic haloperidol decreased N1 and P300 source strength predominantly in those brain regions not involved in the generation of these components, suggesting a shift of resources. The psychostimulant methylphenidate increased P300 source strength in brain regions with major P3b generators, indicating increases in energetic resources associated with stimulus encoding. The antidepressant citalopram increased N1 and P3b source strength in multiple brain regions specifically in the left prefrontal cortex, a brain region in which reduced blood flow and metabolism was found in depressed patients.

EEG source localization and global dimensional complexity in high- and low- hypnotizable subjects: a pilot study.

Individuals differ in hypnotizability. Information on hypnotizability-related EEG characteristics is controversial and incomplete, particularly on intracerebral source localization and EEG dimensionality. 19-channel, eyes-closed resting EEGs from right-handed, healthy, 8 high- and 4 low-hynotizable subjects (age: 26.7 +/- 7.3 years) were analyzed. Hypnotizability was rated after the subjects' ability to attain a deep hypnotic stage (amnesia). FFT Dipole Approximation analysis in seven EEG frequency bands showed significant differences (p < 0.04) of source gravity center locations for theta (6.5-8 Hz, more posterior and more left for highs), beta-1 and beta-2 frequencies (12.5-18 and 18.5-21 Hz; both more posterior and more right for highs). Low Resolution Electromagnetic Tomography (LORETA) specified the cortical anteriorization of beta-1 and beta-2 in low hypnotizables. Power spectral analysis of Global Field Power time series (curves) showed no overall power differences in any band. Full-band Global Dimensional Complexity was higher in high-hypnotizable subjects (p < 0.02). Thus, before hypnosis, high and low hypnotizables were in different brain electric states, with more posterior brain activity gravity centers (excitatory right, routine or relaxation left) and higher dimensional complexity (higher arousal) in high than low hypnotizables.

Limbic activity in slow wave sleep in a healthy subject with alpha-delta sleep.

All-night electroencephalographic (EEG) activity was recorded in a healthy subject with known alpha-delta sleep. Recordings were made from all 19 of the 10/20 system electrode sites, and low resolution electromagnetic tomography (LORETA) was used to estimate intracerebral current densities. Sleep stages were compared within classical frequency bands by statistical parametric mapping (SPM). With the onset of sleep, occipital alpha abated. With increasing depth of sleep, alpha power increased in a region comprising the left frontal lobe, the anterior and parietal cingulum, and the anterior and medial right front lobe. In slow wave sleep (SWS), frontal alpha power was much greater than in wakefulness. The maximum of frontal alpha power of SWS was localised symmetrically in the left and right anterior cingulum. The observed alpha activity was different from the occipital alpha characteristic of wakefulness; it was a distinct activity of separate origin. The anterior limbic lobes seemed to play an active part in SWS in this healthy volunteer with an alpha-delta sleep pattern.

Spatial structure of brain electric fields during intermittent photic stimulation.

EEG changes in 27 young healthy male right-handed volunteers on intermittent photic stimulation (IPS) were estimated using global field power (GFP), EEG microstate modeling and analysis (EMMA), and low-resolution electromagnetic brain tomography (LORETA). The GFP significantly increased at flashing frequency and high harmonics. Three model maps were extracted with the EMMA procedure, from which high alternation rates of each microstate were observed. Moreover, two of the three model maps contributed very highly, occurring most frequently. LORETA imaging of the three model maps obtained from the EMMA procedure showed that both visual dominant cortical areas were activated, especially in the left hemisphere. These results suggest that IPS does not cause peculiar spatial configurations of the brain electric field, but does cause acceleration and deviation of the microstate alternation. Also, a functional laterality between hemispheres might be enhanced by symmetric IPS.

A novel neural network technique for analysis and classification of EM single-particle images.

We propose a novel self-organizing neural network for the unsupervised classification of electron microscopy (EM) images of biological macromolecules. The radical novelty of the algorithm lies in its rigorous mathematical formulation that, starting from a large set of possibly very noisy input data, finds a set of "representative" data items, organized onto an ordered output map, such that the probability density of this set of representative items resembles at its possible best the probability density of the input data. In a way, it summarizes large amounts of information into a concise description that rigorously keeps the basic pattern of the input data distribution. In this application to the field of three-dimensional EM of single particles, two different data sets have been used; one comprised 2458 rotational power spectra of individual negative stain images of the G40P helicase of Bacillus subtilis bacteriophage SPP1, and the other contained 2822 cryoelectron images of SV40 large T-antigen. Our experimental results prove that this technique is indeed very successful, providing the user with the capability of exploring complex patterns in a succinct, informative, and objective manner. The above facts, together with the consideration that the integration of this new algorithm with commonly used software packages is immediate, prompt us to propose it as a valuable new tool in the analysis of large collections of noisy data.

Brain mapping of bilateral visual interactions in children.

Interhemispheric interactions were studied with functional brain mapping of visual processing. Children performed a reaction time task with uni- and bilateral targets and nontargets. The visual evoked potential (VEP) was segmented into P1a, P1b, and N1 microstates using map rather than channel features. Map latencies, amplitudes and sources were tested for bilateral interactions. Bilateral targets yielded shorter VEP map latencies but later response onsets than unilateral ones. Source analyses of the unilateral VEPs indicated a transition from contra- (P1a) to ipsilateral (P1b) visual cortex activation (interhemispheric transfer). Bilateral VEPs were smaller than the summed unilateral VEPs in all microstates. indicating that interhemispheric interactions both precede and follow interhemispheric transfer. Brain mapping of uni- and bilateral VEPs in children thus revealed several distinct forms of interhemispheric interactions in the same, early time range.

Source localization of EEG activity during hypnotically induced anxiety and relaxation.

The engagement of different brain regions which implement subjectively experienced emotional states in normals is not completely clarified. Emotional states can conveniently be induced by hypnosis-based suggestions. We studied brain electric activity during hypnotically induced anxiety and relaxation in 11 right-handed normals (5 males, 6 females, mean age 26.5+/-7.6 years). After induction of light hypnosis, anxiety and then relaxation was suggested using a standardized text (reverse sequence in half of the subjects). Nineteen-channel, eyes-closed EEG (20 artifact-free s/subject) was analyzed (source localization using FFT approximation and low resolution electromagnetic tomography, LORETA). Global tests revealed the strongest difference (P<0.005) between EEG source gravity center locations during the two emotional states in the excitatory beta-2 EEG frequency band (18.5-21 Hz). Post hoc tests showed that the sources were located more right during anxiety than during relaxation (P=0.01). LORETA specified that anxiety showed maximally stronger activity than relaxation in right Brodmann area 10, and relaxation showed maximally stronger activity than anxiety in left Brodmann area 22. Clearly, the two induced emotional states were associated with activity of different neural populations. Our results agree with reports on brain activity shifted to the right (especially fronto-temporal) during negative compared with positive emotions, and support the role of beta-2 EEG frequency in emotional states.

Low-resolution brain electromagnetic tomography revealed simultaneously active frontal and parietal sleep spindle sources in the human cortex.

Analyses of scalp-recorded sleep spindles have demonstrated topographically distinct slow and fast spindle waves. In the present paper, the electrical activity in the brain corresponding to different types of sleep spindles was estimated by means of low-resolution electromagnetic tomography. In its new implementation, this method is based on realistic head geometry and solution space is restricted to the cortical gray matter and hippocampus. In multichannel all-night electroencephalographic recordings, 10-20 artifact-free 1.25-s epochs with frontally, parietally and approximately equally distributed spindles were marked visually in 10 normal healthy subjects aged 20-35years. As a control condition, artifact-free non-spindle epochs 1-3s before or after the corresponding spindle episodes were marked. Low-resolution electromagnetic tomography demonstrated, independent of the scalp distribution, a distributed spindle source in the prefrontal cortex (Brodmann areas 9 and 10), oscillating with a frequency below 13Hz, and in the precuneus (Brodmann area 7), oscillating with a frequency above 13Hz. In extremely rare cases only the prefrontal or the parietal source was active. Brodmann areas 9 and 10 have principal connections to the dorsomedial thalamic nucleus; Brodmann area 7 is connected to the lateroposterior, laterodorsal and rostral intralaminar centrolateral thalamic nuclei. Thus, the localized cortical brain regions are directly connected with adjacent parts of the dorsal thalamus, where sleep spindles are generated. The results demonstrated simultaneously active cortical spindle sources which differed in frequency by approximately 2Hz and were located in brain regions known to be critically involved in the processing of sensory input, which is in line with the assumed functional role of sleep spindles.

Anterior cingulate activity as a predictor of degree of treatment response in major depression: evidence from brain electrical tomography analysis.

OBJECTIVE: The anterior cingulate cortex has been implicated in depression. Results are best interpreted by considering anatomic and cytoarchitectonic subdivisions. Evidence suggests depression is characterized by hypoactivity in the dorsal anterior cingulate, whereas hyperactivity in the rostral anterior cingulate is associated with good response to treatment. The authors tested the hypothesis that activity in the rostral anterior cingulate during the depressed state has prognostic value for the degree of eventual response to treatment. Whereas prior studies used hemodynamic imaging, this investigation used EEG. METHOD: The authors recorded 28-channel EEG data for 18 unmedicated patients with major depression and 18 matched comparison subjects. Clinical outcome was assessed after nortriptyline treatment. Of the 18 depressed patients, 16 were considered responders 4-6 months after initial assessment. A median split was used to classify response, and the pretreatment EEG data of patients showing better (N=9) and worse (N=9) responses were analyzed with low-resolution electromagnetic tomography, a new method to compute three-dimensional cortical current density for given EEG frequency bands according to a Talairach brain atlas. RESULTS: The patients with better responses showed hyperactivity (higher theta activity) in the rostral anterior cingulate (Brodmann's area 24/32). Follow-up analyses demonstrated the specificity of this finding, which was not confounded by age or pretreatment depression severity. CONCLUSIONS: These results, based on electrophysiological imaging, not only support hemodynamic findings implicating activation of the anterior cingulate as a predictor of response in depression, but they also suggest that differential activity in the rostral anterior cingulate is associated with gradations of response.