Daniel H. Lange

Modeling and estimation of single evoked brain potential components

Presents 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, the authors propose an estimation approach based on identification of evoked potential components on a single-trial basis. The estimation process is performed in 2 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 2 assumed signal contributions, namely the ongoing brain activity and the single evoked brain response. The estimators 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, 2 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 2 overlapping components throughout the experimental session is detected and tracked.

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

Current estimators for single-trial evoked potentials (EPs) require a signal-to-noise ratio (SNR) of 0 dB or better to obtain high quality estimations, yet many types of EPs suffer from substantially lower SNRs. This paper presents a robust-evoked-potential-estimator (REPE) facilitating high quality estimations of single movement related EPs 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.

Segmented matched filtering of single event related evoked potentials

A fast segmentation-based matched filtering (MF) technique of single trial evoked potentials (EPs) is presented. MF improves the signal-to-noise ratio of single EPs, 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.<<ETX>>

Variable single-trial evoked potential estimation via principal component identification

Current single-trial Evoked Potential (EP) estimators assume deterministic signal waveforms embedded in the background electroencephalographic brain activity. Identification of morphological changes of the evoked responses has been suggested, requiring however a skilled operator to predetermine the location of the variable components. In this paper the authors propose an alternative approach for the identification of variable single-trial EPs, based on reconstruction of the EPs from the Principal Components of the data correlation matrix and thus eliminating the requirement of a-priori knowledge of the variable component locations. The reconstruction performance is demonstrated via simulations and application to experimental EP data.

Modern Techniques in ERP Research

Evoked Potentials (EP) are defined as averaged electric responses of the nervous system to sensory stimulation (Gevins 1984). They consist of a sequence of transient waveforms, each with its own morphology, latency and amplitude. In clinical settings, EPs are elicited by visual or auditory stimulation, or by electric stimulation of sensory nerves (Chiappa 1983). These EPs are usually recorded from the scalp, although in special cases like during brain surgery, electrodes may be placed on the surface of the brain or even deep in brain tissue. The term Event Related Potential (ERP) is now commonly used to denote both EP as well as other brain responses that are the result of cognitive processes accompanying and following external stimuli, or of preparatory mechanisms preceding motor action. However, due to historical reasons and to avoid confusion, we shall generally use the term EP for all types of brain responses including ERP.

Estimation of morphologically varying single evoked brain potentials

The authors present a new parametric estimator for single morphologically varying evoked potentials (MVEPs). The estimator can detect changes in specific components within the evoked potential complex, extending current methods which only compensate for global latency and gain variations with respect to the averaged evoked response. The authors demonstrate the performance of the estimator on simulated signals at low signal to noise ratios and on movement related brain potentials. The results presented confirm that single MVEPs can be successfully estimated from the noisy background activity, which enables tracking the transient qualities of single evoked responses throughout the experimental session.<<ETX>>

Modeling and Estimation of Amplitude and Time Shifts in Single Evoked Potential Components

It is well established that different components of an evoked potential complex may originate from different functional brain sites, and can be distinguished by their respective latencies and amplitudes. This emphasizes the need and advantage in the ability to estimate single evoked potential components, which would open a window into complex dynamic processes occurring in the central nervous system.

A model based decomposition method for temporally overlapping evoked potential components

The authors present a robust model-based decomposition method for temporally overlapping evoked potential (EP) components. The model assumes linear superposition of the EP constituent components, and Gaussian distributed firing instants of the neuronal population associated with each component. The decomposition provides a loss-less description of the EP complex, which is demonstrated via computer simulations as robust to violations of the model assumptions. The decomposition method is applied to experimental data, demonstrating its separation performance of severely overlapping EP components.

Unsupervised identification of event-related brain potentials via competitive learning

We present a novel approach to the problem of Event-Related Potential (ERP) identification, based on a competitive Artificial Neural Net (ANN). Our approach dismisses the need for stimulus- or event-related selective averaging, thus avoiding conventional assumptions on response invariability. The identifier 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.