Analysis of the electromagnetic signals of the human brain: milestones, obstacles, and goals

Quantitative EEG measures in profoundly deaf and normal hearing individuals while performing a vibrotactile temporal discrimination task

Challenges in early oral language acquisition in profoundly deaf individuals have an impact on cognitive neurodevelopment. This has led to the exploration of alternative sound perception methods involving training of vibrotactile discrimination of sounds within the language spectrum. In particular, stimulus duration plays an important role in linguistic categorical perception. We comparatively evaluated vibrotactile temporal discrimination of sound and how specific training can modify the underlying electrical brain activity. Fifteen profoundly deaf (PD) and 15 normal-hearing (NH) subjects performed a vibrotactile oddball task with simultaneous EEG recording, before and after a short training period (5 one-hour sessions; in 2.5-3 weeks). The stimuli consisted of 700 Hz pure-tones with different duration (target: long 500 ms; non-target: short 250 ms). The sound-wave stimuli were delivered by a small device worn on the right index finger. A similar behavioral training effect was observed in both groups showing significant improvement in sound-duration discrimination. However, quantitative EEG measurements reveal distinct neurophysiological patterns characterized by higher and more diffuse delta band magnitudes in the PD group, together with a generalized decrement in absolute power in both groups that might reflect a facilitating process associated to learning. Furthermore, training-related changes were found in the beta-band in NH. Findings suggest PD have different cognitive adaptive mechanisms which are not a mere amplification effect due to greater cortical excitability.

A Novel, Fast, Reliable, and Data-Driven Method for Simultaneous Single-Trial Mining and Amplitude-Latency Estimation Based on Proximity Graphs and Network Analysis

Both amplitude and latency of single-trial EEG/MEG recordings provide valuable information regarding functionality of the human brain. In this article, we provided a data-driven graph and network-based framework for mining information from multi-trial event-related brain recordings. In the first part, we provide the general outline of the proposed methodological approach. In the second part, we provide a more detailed illustration, and present the obtained results on every step of the algorithmic procedure. To justify the proposed framework instead of presenting the analytic data mining and graph-based steps, we address the problem of response variability, a prerequisite to reliable estimates for both the amplitude and latency on specific N/P components linked to the nature of the stimuli. The major question addressed in this study is the selection of representative single-trials with the aim of uncovering a less noisey averaged waveform elicited from the stimuli. This graph and network-based algorithmic procedure increases the signal-to-noise (SNR) of the brain response, a key pre-processing step to reveal significant and reliable amplitude and latency at a specific time after the onset of the stimulus and with the right polarity (N or P). We demonstrated the whole approach using electroencephalography (EEG) auditory mismatch negativity (MMN) recordings from 42 young healthy controls. The method is novel, fast and data-driven succeeding first to reveal the true waveform elicited by MMN on different conditions (frequency, intensity, duration, etc.). The proposed graph-oriented algorithmic pipeline increased the SNR of the characteristic waveforms and the reliability of amplitude and latency within the adopted cohort. We also demonstrated how different EEG reference schemes (REST vs. average) can influence amplitude-latency estimation. Simulation results revealed robust amplitude-latency estimations under different SNR and amplitude-latency variations with the proposed algorithm.

Single-Trial Evoked Potential Estimating Based on Sparse Coding under Impulsive Noise Environment

Estimating single-trial evoked potentials (EPs) corrupted by the spontaneous electroencephalogram (EEG) can be regarded as signal denoising problem. Sparse coding has significant success in signal denoising and EPs have been proven to have strong sparsity over an appropriate dictionary. In sparse coding, the noise generally is considered to be a Gaussian random process. However, some studies have shown that the background noise in EPs may present an impulsive characteristic which is far from Gaussian but suitable to be modeled by the α-stable distribution (1 < α ≤ 2). Consequently, the performances of general sparse coding will degrade or even fail. In view of this, we present a new sparse coding algorithm using p-norm optimization in single-trial EPs estimating. The algorithm can track the underlying EPs corrupted by α-stable distribution noise, trial-by-trial, without the need to estimate the α value. Simulations and experiments on human visual evoked potentials and event-related potentials are carried out to examine the performance of the proposed approach. Experimental results show that the proposed method is effective in estimating single-trial EPs under impulsive noise environment.

A MISO-ARX-Based Method for Single-Trial Evoked Potential Extraction

In this paper, we propose a novel method for solving the single-trial evoked potential (EP) estimation problem. In this method, the single-trial EP is considered as a complex containing many components, which may originate from different functional brain sites; these components can be distinguished according to their respective latencies and amplitudes and are extracted simultaneously by multiple-input single-output autoregressive modeling with exogenous input (MISO-ARX). The extraction process is performed in three stages: first, we use a reference EP as a template and decompose it into a set of components, which serve as subtemplates for the remaining steps. Then, a dictionary is constructed with these subtemplates, and EPs are preliminarily extracted by sparse coding in order to roughly estimate the latency of each component. Finally, the single-trial measurement is parametrically modeled by MISO-ARX while characterizing spontaneous electroencephalographic activity as an autoregression model driven by white noise and with each component of the EP modeled by autoregressive-moving-average filtering of the subtemplates. Once optimized, all components of the EP can be extracted. Compared with ARX, our method has greater tracking capabilities of specific components of the EP complex as each component is modeled individually in MISO-ARX. We provide exhaustive experimental results to show the effectiveness and feasibility of our method.

A multi-resolution approach to localize neural sources of P300 event-related brain potential

BACKGROUND AND OBJECTIVE

P300 is probably the most well-known component of event-related brain potentials (ERPs). Using an oddball paradigm, a P300 component can be identified, that is, elicited by the target stimuli recognition. Since P300 is associated with attention and memory operations of the brain, investigation of this component can improve our understanding of these mechanisms. The present study is aimed at identifying the P300 generators in 30 healthy subjects aged 18-30 years using time-reduction region-suppression linearly constrained minimum variance (TR-LCMV) beamformer.

METHODS

In our study, TR-LCMV beamformer with multi-resolution approach is proposed, coarse-resolution space to find the approximated coherent source locations, fine-resolution space to estimate covariance matrix for dimension reduction of determined regions, and normal-resolution space to localize the P300 generators in the brain.

RESULTS

Our results over simulated and real data showed that this approach is a suitable tool to the analysis of ERP fields with localizing superior and inferior frontal lobe, middle temporal gyrus, parietal lobe, and cingulate gyrus as the most prominent sources of P300. The result of P300 localization was finally compared with the other localization methods and it is demonstrated that enhanced performance is achieved.

CONCLUSIONS

Our results showed that the P300 originates from a widespread neuronal network in the brain and not from a specific region. Our finding over simulated and real data demonstrated the ability of the TR-LCMV algorithm for P300 source localization.

EEG oscillatory states: universality, uniqueness and specificity across healthy-normal, altered and pathological brain conditions

For the first time the dynamic repertoires and oscillatory types of local EEG states in 13 diverse conditions (examined over 9 studies) that covered healthy-normal, altered and pathological brain states were quantified within the same methodological and conceptual framework. EEG oscillatory states were assessed by the probability-classification analysis of short-term EEG spectral patterns. The results demonstrated that brain activity consists of a limited repertoire of local EEG states in any of the examined conditions. The size of the state repertoires was associated with changes in cognition and vigilance or neuropsychopathologic conditions. Additionally universal, optional and unique EEG states across 13 diverse conditions were observed. It was demonstrated also that EEG oscillations which constituted EEG states were characteristic for different groups of conditions in accordance to oscillations' functional significance. The results suggested that (a) there is a limit in the number of local states available to the cortex and many ways in which these local states can rearrange themselves and still produce the same global state and (b) EEG individuality is determined by varying proportions of universal, optional and unique oscillatory states. The results enriched our understanding about dynamic microstructure of EEG-signal.

EEG oscillatory states: universality, uniqueness and specificity across healthy-normal, altered and pathological brain conditions

For the first time the dynamic repertoires and oscillatory types of local EEG states in 13 diverse conditions (examined over 9 studies) that covered healthy-normal, altered and pathological brain states were quantified within the same methodological and conceptual framework. EEG oscillatory states were assessed by the probability-classification analysis of short-term EEG spectral patterns. The results demonstrated that brain activity consists of a limited repertoire of local EEG states in any of the examined conditions. The size of the state repertoires was associated with changes in cognition and vigilance or neuropsychopathologic conditions. Additionally universal, optional and unique EEG states across 13 diverse conditions were observed. It was demonstrated also that EEG oscillations which constituted EEG states were characteristic for different groups of conditions in accordance to oscillations' functional significance. The results suggested that (a) there is a limit in the number of local states available to the cortex and many ways in which these local states can rearrange themselves and still produce the same global state and (b) EEG individuality is determined by varying proportions of universal, optional and unique oscillatory states. The results enriched our understanding about dynamic microstructure of EEG-signal.

Toward high performance, weakly invasive brain computer interfaces using selective visual attention

Brain-computer interfaces have been proposed as a solution for paralyzed persons to communicate and interact with their environment. However, the neural signals used for controlling such prostheses are often noisy and unreliable, resulting in a low performance of real-world applications. Here we propose neural signatures of selective visual attention in epidural recordings as a fast, reliable, and high-performance control signal for brain prostheses. We recorded epidural field potentials with chronically implanted electrode arrays from two macaque monkeys engaged in a shape-tracking task. For single trials, we classified the direction of attention to one of two visual stimuli based on spectral amplitude, coherence, and phase difference in time windows fixed relative to stimulus onset. Classification performances reached up to 99.9%, and the information about attentional states could be transferred at rates exceeding 580 bits/min. Good classification can already be achieved in time windows as short as 200 ms. The classification performance changed dynamically over the trial and modulated with the task's varying demands for attention. For all three signal features, the information about the direction of attention was contained in the γ-band. The most informative feature was spectral amplitude. Together, these findings establish a novel paradigm for constructing brain prostheses as, for example, virtual spelling boards, promising a major gain in performance and robustness for human brain-computer interfaces.

Neurophysiological pharmacodynamic measures of groups and individuals extended from simple cognitive tasks to more "lifelike" activities

OBJECTIVE

Extend a method to track neurophysiological pharmacodynamics during repetitive cognitive testing to a more complex "lifelike" task.

METHODS

Alcohol was used as an exemplar psychoactive substance. An equation, derived in an exploratory analysis to detect alcohol's EEGs effects during repetitive cognitive testing, was validated in a Confirmatory Study on a new group whose EEGs after alcohol and placebo were recorded during working memory testing and while operating an automobile driving simulator.

RESULTS

The equation recognized alcohol by combining five times beta plus theta power. It worked well (p < .0001) when applied to both tasks in the confirmatory group. The maximum EEG effect occurred 2-2.5 h after drinking (>1 h after peak BAC) and remained at 90% at 3.5-4 h (BAC < 50% of peak). Individuals varied in the magnitude and timing of the EEG effect.

CONCLUSION

The equation tracked the EEG response to alcohol in the Confirmatory Study during both repetitive cognitive testing and a more complex "lifelike" task. The EEG metric was more sensitive to alcohol than several autonomic physiological measures, task performance measures or self-reports.

SIGNIFICANCE

Using EEG as a biomarker to track neurophysiological pharmacodynamics during complex "lifelike" activities may prove useful for assessing how drugs affect integrated brain functioning.

Real-time change detection of steady-state evoked potentials

Steady-state evoked potentials (SSEP) are the electrical activity recorded from the scalp in response to high-rate sensory stimulation. SSEP consist of a constituent frequency component matching the stimulation rate, whose amplitude and phase remain constant with time and are sensitive to functional changes in the stimulated sensory system. Monitoring SSEP during neurosurgical procedures allows identification of an emerging impairment early enough before the damage becomes permanent. In routine practice, SSEP are extracted by averaging of the EEG recordings, allowing detection of neurological changes within approximately a minute. As an alternative to the relatively slow-responding empirical averaging, we present an algorithm that detects changes in the SSEP within seconds. Our system alerts when changes in the SSEP are detected by applying a two-step Generalized Likelihood Ratio Test (GLRT) on the unaveraged EEG recordings. This approach outperforms conventional detection and provides the monitor with a statistical measure of the likelihood that a change occurred, thus enhancing its sensitivity and reliability. The system's performance is analyzed using Monte Carlo simulations and tested on real EEG data recorded under coma.

Towards measuring brain function on groups of people in the real world

In three studies, EEGs from three groups of participants were recorded during progressively more real world situations after drinking alcoholic beverages that brought breath alcohol contents near the limit for driving in California 30 minutes after drinking. A simple equation that measured neurophysiological effects of alcohol in the first group of 15 participants performing repetitive cognitive tasks was applied to a second group of 15 operating an automobile driving simulator, and to a third group of 10 ambulatory people recorded simultaneously during a cocktail party. The equation derived from the first group quantified alcohol’s effect by combining measures of higher frequency (beta) and lower frequency (theta) power into a single score. It produced an Area Under the Receiver Operator Characteristic Curve of .73 (p<.05; 67% sensitivity in recognizing alcohol and 87% specificity in recognizing placebo). Applying the same equation to the second group operating the driving simulator, AUC was .95, (p<.0001; 93% sensitivity and 73% specificity), while for the cocktail party group AUC was .87 (p<.01; 80% sensitivity and 80% specificity). EEG scores were significantly related to breath alcohol content in all studies. Some individuals differed markedly from the overall response evident in their respective groups. The feasibility of measuring the neurophysiological effect of a psychoactive substance from an entire group of ambulatory people at a cocktail party suggests that future studies may be able to fruitfully apply brain function measures derived under rigorously controlled laboratory conditions to assess drug effects on groups of people interacting in real world situations.

A fast and reliable method for simultaneous waveform, amplitude and latency estimation of single-trial EEG/MEG data

The amplitude and latency of single-trial EEG/MEG signals may provide valuable information concerning human brain functioning. In this article we propose a new method to reliably estimate single-trial amplitude and latency of EEG/MEG signals. The advantages of the method are fourfold. First, no a-priori specified template function is required. Second, the method allows for multiple signals that may vary independently in amplitude and/or latency. Third, the method is less sensitive to noise as it models data with a parsimonious set of basis functions. Finally, the method is very fast since it is based on an iterative linear least squares algorithm. A simulation study shows that the method yields reliable estimates under different levels of latency variation and signal-to-noise ratioÕs. Furthermore, it shows that the existence of multiple signals can be correctly determined. An application to empirical data from a choice reaction time study indicates that the method describes these data accurately.

Editorial: EEG Phenomenology and Multiple Faces of Short-term EEG Spectral Pattern

An electroencephalogram (EEG) signal is extremely nonstationary, highly composite and very complex, all of which reflects the underlying integral neurodynamics. Understanding the EEG "grammar", its internal structural organization would place a "Rozetta stone" in researchers' hands, allowing them to more adequately describe the information processes of the brain in terms of EEG phenomenology. This Special Issue presents a framework where short-term EEG spectral pattern (SP) of a particular type is viewed as an information-rich event in EEG phenomenology. It is suggested that transition from one type of SP to another is accompanied by a "switch" between brain microstates in specific neuronal networks, or in cortex areas; and these microstates are reflected in EEG as piecewise stationary segments. In this context multiple faces of a short-term EEG SP reflect the poly-operational structure of brain activity.

Attention improves object representation in visual cortical field potentials

Selective attention improves perception and modulates neuronal responses, but how attention-dependent changes of cortical activity improve the processing of attended objects is an open question. Changes in total signal strength or enhancements in signal-to-noise ratio have been proposed as putative mechanisms. However, it is still not clear whether, and to what extent, these processes contribute to the large perceptual improvements. We studied the ability to discriminate states of activity in visual cortex evoked by differently shaped objects depending on selective attention in monkeys. We found that gamma-band activity from V4 and V1 contains a high amount of information about stimulus shape, which increases for V4 recordings considerably with attention in successful trials, but not in case of behavioral errors. This effect resulted from enhanced differences between the stimulus-specific distributions of power spectral amplitudes. It could be explained neither by enhancements of signal-to-noise ratios, nor by changes in total signal power. Instead our results indicate that attention causes underlying cortical network states to become more distinct for different stimuli, providing a new neurophysiological explanation for improvements of behavioral performance by attention. The absence of the enhancement in discriminability in trials with behavioral errors demonstrates the relevance of this novel neural mechanism for perception.

Coherent oscillatory activity in monkey area v4 predicts successful allocation of attention

Attention serves to select objects from often complex scenes for enhanced processing and perception. In particular, the perception of shape depends critically on attention for integrating the various parts of the selected object into a coherent representation of object shape. To study whether oscillatory neuronal synchrony may serve as a mechanism of attention in shape perception, we introduced a novel shape-tracking task requiring sustained attention to a morphing shape. Attention was found to strongly increase oscillatory currents underlying the recorded field potentials in the gamma-frequency range, thus indicating enhanced neuronal synchrony within the population of V4 neurons representing the attended stimulus. Errors indicating a misdirection of attention to the distracter instead of the target were preceded by a corresponding shift of oscillatory activity from the target's neuronal representation to that of the distracter. No such effect was observed for errors unrelated to attention. Modulations of the attention-dependent enhancement of oscillatory activity occurred in correspondence with changing attentional demands during the course of a trial. The specificity of the effect of attentional errors together with the close coupling between attentional demand and oscillatory activity support the hypothesis that oscillatory neuronal synchrony serves as a mechanism of attention.

Warp-averaging event-related potentials

OBJECTIVE

To align the repeated single trials of the event-related potential (ERP) in order to get an improved estimate of the ERP.

METHODS

A new implementation of the dynamic time warping is applied to compute a warp-average of the single trials. The trilinear modeling method is applied to filter the single trials prior to alignment. Alignment is based on normalized signals and their estimated derivatives. These features reduce the misalignment due to aligning the random alpha waves, explaining amplitude differences in latency differences, or the seemingly small amplitudes of some components.

RESULTS

Simulations and applications to visually evoked potentials show significant improvement over some commonly used methods.

CONCLUSIONS

The new implementation of the dynamic time warping can be used to align the major components (P1, N1, P2, N2, P3) of the repeated single trials. The average of the aligned single trials is an improved estimate of the ERP. This could lead to more accurate results in subsequent analysis.

Subspace averaging of steady-state visual evoked potentials

A new algorithm for doing signal averaging of steady-state visual evoked potentials (VEP's) is described. The subspace average is obtained by finding the orthogonal projection of the VEP measurement vector onto the signal subspace, which is based on a sinusoidal VEP signal model. The subspace average is seen to out-perform the conventional average using a new signal-to-noise-ratio-based performance measure on simulated and actual VEP data.

Overcoming selective ensemble averaging: unsupervised identification of event-related brain potentials

We present a novel approach to the problem of event-related potential (ERP) identification, based on a competitive artificial neural network (ANN) structure. Our method uses ensembled electroencephalogram (EEG) data just as used in conventional averaging, however without the need for a priori data subgrouping into distinct categories (e.g., stimulus- or event-related), and thus avoids conventional assumptions on response invariability. The competitive ANN, often described as a winner takes all neural structure, is based on dynamic competition among the net neurons where learning takes place only with the winning neuron. Using a simple single-layered structure, the proposed scheme results in convergence of the actual neural weights to the embedded ERP patterns. The method is applied to real event-related potential data recorded during a common odd-ball type paradigm. For the first time, within-session variable signal patterns are automatically identified, dismissing the strong and limiting requirement of a priori stimulus-related selective grouping of the recorded data. The results present new possibilities in ERP research.

Detection of P300 waves in single trials by the wavelet transform (WT)

The P300 response is conventionally obtained by averaging the responses to the task-relevant (target) stimuli of the oddball paradigm. However, it is well known that cognitive ERP components show a high variability due to changes of cognitive state during an experimental session. With simple tasks such changes may not be demonstrable by the conventional method of averaging the sweeps chosen according to task-relevance. Therefore, the present work employed a response-based classification procedure to choose the trials containing the P300 component from the whole set of sweeps of an auditory oddball paradigm. For this purpose, the most significant response property reflecting the P300 wave was identified by using the wavelet transform (WT). The application of a 5 octave quadratic B-spline-WT on single sweeps yielded discrete coefficients in each octave with an appropriate time resolution for each frequency range. The main feature indicating a P300 response was the positivity of the 4th delta (0.5-4 Hz) coefficient (310-430 ms) after stimulus onset. The average of selected single sweeps from the whole set of data according to this criterion yielded more enhanced P300 waves compared with the average of the target responses, and the average of the remaining sweeps showed a significantly smaller positivity in the P300 latency range compared with the average of the non-target responses. The combination of sweeps classified according to the task-based and response-based criteria differed significantly. This suggests an influence of changes in cognitive state on the presence of the P300 wave which cannot be assessed by task performance alone.

Trends in EEG source localization

The concepts underlying the quantitative localization of the sources of the EEG inside the brain are reviewed along with the current and emerging approaches to the problem. The concepts mentioned include monopolar and dipolar source models and head models ranging from the spherical to the more realistic based on boundary and finite elements. The forward and inverse problems in electroencephalography are discussed, including the non-uniqueness of the inverse problem. The approaches to the solution of the inverse problem described include single and multiple time-slice localization, equivalent dipole localization and the weighted minimum norm. The multiple time-slice localization approach is highlighted as probably the best available at this time and is discussed in terms of the spatiotemporal model of the EEG. The effect of noise corruption, artifacts and the number of recording electrodes on the accuracy of source localization is also mentioned. It is suggested that the main appeal of the minimum norm is that it does not assume a model for the sources and provides an estimate of the current density everywhere in the three dimensional volume of the head.

Modeling and estimation of single evoked brain potential components

In this paper, we present a novel approach to solving the single-trial evoked-potential estimation problem. Recognizing that different components of an evoked potential complex may originate from different functional brain sites and can be distinguished according to their respective latencies and amplitudes, we propose an estimation approach based on identification of evoked potential components on a single-trial basis. The estimation process is performed in two stages: first, an average evoked potential is calculated and decomposed into a set of components, with each component serving as a subtemplate for the next stage; then, the single measurement is parametrically modeled by a superposition of an emulated ongoing electroencephalographic activity and a linear combination of latency and amplitude-corrected component templates. Once optimized, the model provides the two assumed signal contributions, namely the ongoing brain activity and the single evoked brain response. The estimator's performance is analyzed analytically and via simulation, verifying its capability to extract single components at low signal-to-noise ratios typical of evoked potential data. Finally, two applications are presented, demonstrating the improved analysis capabilities gained by using the proposed approach. The first application deals with movement related brain potentials, where a change of the single evoked response due to external loading is detected. The second application involves cognitive event-related brain potentials, where a dynamic change of two overlapping components throughout the experimental session is detected and tracked.

A fuzzy clustering approach to study the auditory P50 component in schizophrenia

We have recently provided evidence that selective evoked response averaging based on a fuzzy clustering approach is a useful way to increase the signal-to-noise ratio, particularly when recording low-amplitude components, such as the auditory P50. We have also reported that, when stimuli are delivered in pairs (S1 followed by S2) with a short interstimulus interval, the first stimulus (S1) results in synchronization of the EEG producing a large-amplitude evoked response, whereas the second stimulus (S2) causes phase opposition resulting in a lower amplitude average evoked response. In the current study we reanalyzed data previously obtained from 13 normal volunteers and 17 chronic schizophrenia patients. Our results show that the partial EPs corresponding to the S1 stimulus are highly synchronized in normal subjects but not in schizophrenia patients. However, such a synchronization is not present after delivery of the S2 stimulus, neither in normal controls nor in patients. These findings are in agreement with previous reports of decreased amplitude of the S1 response without a significant further decrease in the amplitude of the S2 response in schizophrenia patients.

Low-level EMFs are transduced like other stimuli

The aims of this study were to test the theory that transduction of low-level electromagnetic fields (EMFs) is mediated like other stimuli, and to determine the false-negative rate of the method used to assess the occurrence of transduction (intra-subject comparison of stimulus and non-stimulus states (ICOS)). A light stimulus was chosen as a basis of comparison because light could be applied and removed at precise time points, similar to the manner in which EMFs were controlled. Subjects exposed to a weak light stimulus during 2-second epochs exhibited alterations in brain electrical activity that were similar to those previously observed in subjects exposed to EMFs. The false-negative rate of the ICOS method was 61%, since it registered an effect in only 39% of the subjects (11/28) whereas all subjects were actually aware of the light. In a second group of subjects that were exposed to 0.8 G (1.5 or 10 Hz), 58% (11/19) exhibited similar alterations in brain activity, as determined using ICOS. Previous measurements in the same subjects using a different method showed that the EMFs actually affected brain electrical activity in all subjects; consequently, the false-negative rate was 42% when an EMF was used as the stimulus. The results suggested that the post-transduction brain electrical processes in human subjects were similar in the cases of EMF and light stimuli, as hypothesized, and that the high negative rate of the ICOS method (here and in previous studies) was composed partly or entirely of false-negative results.

A robust parametric estimator for single-trial movement related brain potentials

Current estimators for single-trial evoked potentials (EP's) require a signal-to-noise ratio (SNR) of 0 dB or better to obtain high quality estimations, yet many types of EP's suffer from substantially lower SNR's. This paper presents a robust-evoked-potential-estimator (REPE) facilitating high quality estimations of single movement related EP's with a relatively low SNR. The estimator is based on a standard ARX model, enhanced to support estimation under poor SNR conditions. The REPE was tested successfully on a computer simulated data set giving reliable single-trial estimations for the low SNR range of around -20 dB. THe REPE was also applied to experimental data, producing clear single-trial estimations of movement related brain signals recorded in a classic scenario of self-paced finger tapping experiment.

Adaptive estimation of single response evoked potentials

A new adaptive filtering algorithm and structure is developed to estimate response-to-response variations in evoked responses. The evoked responses are modeled as the sum of three uncorrelated signal components: ensemble average, noise, and stochastic signal variation. A two stage time sequenced filter structure exhibiting improved convergence characteristics is developed along with a modified P-vector algorithm (mPa) which eliminates the need for a separate desired signal electrode. The mPa adaptive filter is tested with simulated and human EP data. The mPa filter is able to estimate signal variations from one response to the next.

Segmented matched filtering of single event related evoked potentials

A fast segmentation-based Matched Filtering (MF) technique of single trial Evoked Potentials (EP's) is presented. MF improves the Signal-to-Noise Ratio of single EP's, reducing the number of repetitions necessary to obtain high quality signals by an order of magnitude. A computer simulation and analysis of experimental data of Movement Related Potentials and cognitive Event Related Potentials demonstrate the superior capabilities of MF compared to traditional Ensemble Averaging.

Estimation of single sweep steady-state visual evoked potentials by adaptive line enhancement

An adaptive line enhancer (ALE) is used to obtain estimates of the single sweep steady-state visual evoked potential (SSVEP). The method is seen to enhance the estimated signal-to-noise ratio of the single sweep SSVEP by as much as 10 dB.

A multistage system to detect epileptiform activity in the EEG

A PC-based system has been developed to automatically detect epileptiform activity in sixteen-channel bipolar EEG's. The system consists of three stages: data collection, feature extraction, and event detection. The feature extractor employs a mimetic approach to detect candidate epileptiform transients on individual channels, while an expert system is used to detect focal and nonfocal multichannel epileptiform events. Considerable use of spatial and temporal contextual information present in the EEG aids both in the detection of epileptiform events and in the rejection of artifacts and background activity as events. Classification of events as definite or probable overcomes, to some extent, the problem of maintaining high detection rates while eliminating false detections. So far, the system has only been evaluated on development data but, although this does not provide a true measure of performance, the results are nevertheless impressive. Data from 11 patients, totaling 180 minutes of sixteen-channel bipolar EEG's, have been analyzed. A total of 45-71% (average 58%) of epileptiform events reported by the human expert in any EEG were detected as definite with no false detections (i.e., 100% selectivity) and 60-100% (average 80%) as either definite or probable but at the expense of up to nine false detections per hour. Importantly, the highest detection rates were achieved on EEG's containing little epileptiform activity and no false detections were made on normal EEG's.

Topographic mapping of single sweep evoked potentials in the brain

Single trial analysis of brain-evoked potentials via stochastic parametric identification and filtering is here extended to multichannel recordings, leading to the topographic mapping of the brain activity elicited by a single stimulus, instead of the usual averaged mapping. The temporal dynamics of the subsequent sweeps in the protocol of a neurophysiologic experiment can thus be recovered and quantified also on its spatial characteristic.

Time-frequency representation of electrocorticograms in temporal lobe epilepsy

Three time-frequency distributions are evaluated in terms of their efficacy in representing nonstationary electrocorticograms (ECoG's) in human temporal lobe epilepsy. The results of a new method, the exponential distribution, are compared with those of the spectrogram and the Wigner distribution. It is shown that the exponential distribution represents a considerable improvement over the spectrogram in terms of resolution and markedly reduces cross-terms present in the Wigner distribution. Exponential distribution representations of ECoG's from different stages of an epileptic record are developed as contour maps. These high-resolution representations offer a lucid display of temporal-spectral features of the rapidly varying signals that constitute ECoG's recorded in temporal lobe epilepsy.

Processing asymmetries for complex sounds: comparisons between behavioral ear advantages and electrophysiological asymmetries based on quantitative electroencephalography

This experiment extends our earlier work on individual differences in ear advantages for complex sounds (Lauter 1982, 1983, 1984) to examine the results of combined behavioral and qEEG testing in the same subjects. Results include: (1) between-subject differences in absolute values together with between-subject agreements in terms of relative values, observed both for ear advantages (EAs) and hemisphere advantages (HAs); (2) within-subject agreement between behavioral (EAs) and physiological (HAs) measures of asymmetries; and (3) preliminary findings related to the interpretation of qEEG asymmetry data, such as the influence of hand movements on auditory-cortex qEEG recordings, and persistence of activation effects in which asymmetries evoked during a stimulation condition may be reflected in resting asymmetries observed during a subsequent control condition.

Weighted averaging of evoked potentials

Weighted averages of brain evoked potentials (EP's) are obtained by weighting each single EP sweep prior to averaging. These weights are shown to maximize the signal-to-noise ratio (SNR) of the resulting average if they satisfy a generalized eigenvalue problem involving the correlation matrices of the underlying signal and noise components. The signal and noise correlation matrices are difficult to estimate and the solution of the generalized eigenvalue problem is often computationally impractical for real-time processing. Correspondingly, a number of simplifying assumptions about the signal and noise correlation matrices are made which allow an efficient method of approximating the maximum SNR weights. Experimental results are given using actual auditory EP data which demonstrate that the resulting weighted average has estimated SNR's that are up to 21% greater than the conventional ensemble average SNR.

The relationship between prestimulus-alpha amplitude and visual evoked potential amplitude

Root-mean-square (RMS) amplitude derived from power spectral measures in the alpha band of the 1 s prestimulus EEG were related to the peak-to-peak amplitude of the N1 and P2 components (N1P2PP) of the visual evoked potential (VEP) in 7 male subjects. Stimuli were low intensity flashes delivered randomly between 2 and 6 whole seconds. Trials were rank ordered according to the levels of prestimulus alpha amplitude and were partitioned into groups of 40 trials each (25 groups per data set). Averaged VEPs were computed from these groups and scattergrams of N1P2PP and enhancement factor (following the approach by Başar, 1980) vs. prestimulus alpha amplitude were produced. There was a correlation of 0.74 (p less than .0001) between prestimulus alpha amplitude and N1P2PP, and all seven subjects displayed a general inverse relationship between VEP enhancement and prestimulus alpha amplitude, replicating the results of Başar. However, we observed an exponential relationship, rather than the linear relationship reported by Başar.

The quantitative extraction and topographic mapping of the abnormal components in the clinical EEG

A method is described which seems to be effective for extracting the abnormal components from the clinical EEG. The approach involves the use of a set a spatial patterns which are common to recorded and 'normal' EEGs and which can account for maximally different proportions of the combined variances in both EEGs. These spatial factors are used to decompose the EEG into orthogonal temporal wave forms which can be judged by the expert electroencephalographer to be abnormal, normal or of artifactual origin. The original EEG is then reconstructed using only the abnormal components and principal component analysis is used to present the spatial topography of the abnormal components. The effectiveness of the method is discussed along with its value for localization of abnormal sources. It is suggested, in conclusion, that the approach described may be optimal for interpretation of the clinical EEG since it allows what is best in terms of quantitative analysis of the EEG to be combined with the best that is available in terms of expert qualitative analysis.

Computerized seizure detection of complex partial seizures

In this study, we describe a computerized method that uses 3 quantified EEG features and discriminant analysis to automatically detect seizure EEG. The quantified EEG features were relative amplitude, dominant frequency and rhythmicity. Using EEGs recorded from intracranial electrodes, the seizure detection method was applied to consecutive non-overlapping 2-channel EEG epochs. A seizure detection sensitivity, ranging from 90% to 100%, was associated with a false positive detection rate of 1.5-2.5/h. The performance of the seizure detection method remained stable for EEG recorded over variable time periods.

Artificial neural networks for source localization in the human brain

Source localization in the brain remains an ill-posed problem unless further constraints about the type of sources and the head model are imposed. Human head is modeled in various ways depending critically on the computing power available and/or the required level of accuracy. Sophisticated and truly representative models may yield more accurate results in general, but at the cost of prohibitively long computer times and huge memory requirements. In conventional source localization techniques, solution source parameters are taken as those which minimize an index of performance, defined relative to the model-generated and clinically measured voltages. We propose the use of a neural network in the place of commonly employed minimization algorithms such as the Simplex Method and the Marquardt algorithm, which are iterative and time consuming. With the aid of the error-backpropagation technique, a neural network is trained to compute source parameters, starting from a voltage set measured on the scalp. Here we describe the methods of training the neural network and investigate its localization accuracy. Based on the results of extensive studies, we conclude that neural networks are highly feasible as source localizers. A trained neural network's independence of localization speed from the head model, and the rapid localization ability, makes it possible to employ the most complex head model with the ease of the simplest model. No initial parameters need to be guessed in order to start the calculation, implying a possible automation of the entire localization process. One may train the network on experimental data, if available, thereby possibly doing away with head models.

Inversion of simulated evoked potentials to charge distribution inside the human brain using an algebraic reconstruction technique

An analytic method is presented to estimate the evolution of electrical charge distribution inside the human brain related to the evoked potentials observed on the head surface. A three-layer concentric spherical human head model is adopted to express the relation between the observed potentials on the head surface and the spatial charge distribution inside the brain. An integral equation associated with the three-layer concentric head model Green's function is employed. Assuming the electric potentials are measured on the head surface, the charge distributions inside the human brain are computed by solving an inverse problem. The Green's function integral equation is inverted by using an algebraic reconstruction technique widely employed in X-ray tomography imaging. The accuracy of the proposed technique is examined by employing computer simulations and by checking the self-consistency of the algorithm.

Topographic segmentation of waking EEG in medication-free schizophrenic patients

Lehmann has demonstrated that EEG topography can be used to segment EEG map series into a sequence of spatially stationary segments characterized by location of potential maxima and minima. We employed topographic segmentation techniques to study 9 channel EEGs recorded from 11 medication-free schizophrenic patients and 10 normal controls during resting and active task conditions, retesting 8 patients after neuroleptic treatment. To define EEG segments, average reference potential maps corresponding to global field power peaks in theta, alpha, and low beta activity were classified according to locations of extreme minimum and maximum values. Normals and schizophrenics did not differ in the number or types of switches between segments, or the frequency of hemisphere crossing of potential extrema. However, EEGs of normal subjects were characterized by significantly more (P less than 0.003) unused theta segment types (of a theoretically possible 36). Moreover, medication significantly (P less than 0.02) increased the number of unused theta segment types in EEGs of schizophrenics. We interpret these findings as evidence of increased spatial variability of brain electrical activity in schizophrenics and discuss their functional implications.