Measuring the functioning of the human brain is one of the most formidable scientific/engineering endeavors ever undertaken. It is difficult to extract information about any particular processing function from brain electromagnetic signals (BEMS) since, at any instant, only a small fraction of the brain's hundreds of simultaneously active major systems might be performing processing related to the function being studied. With recent developments, a new era of research is dawning based on an interdisciplinary approach in which advanced signal processing methods are focused on increasingly more specific neuroanatomical, neurophysiological, and neuropsychological research questions and clinical applications. This brief review highlights the major accomplishments of the last several decades in human BEMS analysis and discusses obstacles to progress. Five main topics are addressed: 1) the historical problem of developing a computerized expert clinical electroencephalogram (EEG) system; 2) advances in signal processing methods, including primary analysis, feature extraction, and statistical hypothesis testing and pattern classification; 3) integrated computing systems for BEMS analysis; 4) biophysical, basic science, practical and conceptual obstacles to progress; and 5) the long-term goal of developing a device for measuring the functional integrity of major neural systems, and the related topic of neurocybernetics. Cutting-edge issues discussed include measurement and modeling of nonstationary event-related signals, characterization of spatial processes, single-trial signal detection, location of the sources of scalp-recorded field distributions, and studies of the functional significance of BEMS.

Analysis of the Electromagnetic Signals of the Human Brain: Milestones, Obstacles, and Goals

Exploiting the intra-subject latency variability from single-trial event-related potentials in the P3 time range: A review and comparative evaluation of methods.

Weighted averages of brain evoked potentials (EPs) 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 SNRs that are up to 21% greater than the conventional ensemble average SNR. >

Weighted averaging of evoked potentials

An original method is presented for the single sweep analysis of visual evoked potentials (VEP's). The introduced algorithm bases upon an AutoRegressive with eXogenous input (ARX) modelling. A Least Squares procedure estimates the coefficients of the model and allows to obtain a complete black-box description of the signal generation mechanism, besides providing a filtered version of the single sweep potential. The performance of the algorithm is verified on proper simulation tests and the experimental results put into evidence the noticeable improvement of signal-to-noise ratio with a consequent better recognition of the classical parameters of the peaks (latencies and amplitudes). The possibility of measuring these parameters on a single sweep basis enables to evaluate the dynamics of the Central Nervous System response during the entire course of the examination. A classification of the estimated evoked potentials in a small number of subsets, on the basis of their morphology, is also possible.

Single sweep analysis of visual evoked potentials through a model of parametric identification

A posterior "Wiener" filtering is an estimation technique that has gained renewed interest in recent years, particularly in evoked potential studies. However, several misconceptions appear to exist with respect to the theory and its applicability. The purpose of this paper is to clarify these aspects and to explain the method in a comprehensive manner. Finally, the use of a more powerful time-varying filtering technique is recommended.

Facts and Fancies About a Posteriori "Wiener" Filtering

In order to obtain better estimates of the true evoked potential signal with fewer stimuli than is possible with classical signal averaging, the Wiener theory [16] for optimum linear filtering, as formulated by Walter [15] and Doyle [5], was implemented in four different versions of the Wiener filter?two a posteriori filters and two recursive filters. The effectiveness of the filters in separating signal from noise was tested in two simulations with known signal, and both simulated noise and real spontaneous brain wave activity. Correlation coefficients between the known test signals and filtered and unfiltered averages of test signal plus noise showed no significant difference between filtered and unfiltered averages.

Eloise H. Carlton

Papers

Is Wiener Filtering an Effective Method of Improving Evoked Potential Estimation