Robust moving averages,with Hopfield neural network implementation,for monitoring evoked potential signals

This technical note describes a robust version of moving averages, that enables reliable monitoring of the evoked potential (EP) signals. A cluster analysis (CA) procedure is introduced to robustify the signal averaging (SA). It is implemented via a Hopfield neural network (HNN), which performs selection of the trials forming a cluster around the current state of the EP signal. The core of this cluster serves as an estimate of the instantaneous EP. The effectiveness of the method, indicated by application to real data, and its computation efficiency, due to the use of simple matrix operations, makes it very promising for clinical observations.     

Single evoked potential reconstruction by means of wavelet transform

We would like to propose a method of single evoked potential (EP) extraction free from assumptions and based on a novel approach — the wavelet representation of the signal. Wavelets were introduced by Grossman and Morlet in 1984. The method is based on the multiresolution signal decomposition. Wavelets are already used for speech recognition, geophysics investigations and fractal analysis. This method seems to be a useful improvement upon Fourier Transform analysis, since it provides simultaneous information on frequency and time localization of the signal. We would like to introduce wavelet formalism for the first time to brain signal analysis. One of the most important problems in this field is the analysis of evoked potentials. This signal has an amplitude several times smaller than EEG, therefore stimulus-synchronized averaging is commonly used. This method is based on several assumptions. Namely it is postulated that: 1) EP are characterized by a deterministic repeatable pattern, 2) EEG has purely stochastic character, 3) EEG and EP are independent. These assumptions have been challenged e.g. the variability of the EP pattern was demonstrated by John (1973) by means of factor analysis. In view of the works of Sayers et al. (1974) and Baar (1988) EP reflects the reorganization of the spontaneous activity under the influence of a stimulus and it is connected with the redistribution of EEG phases. Several attempts to overcome the limitation of the averaging method have been made. Heintze and Künkel (1984) used an autoregressive moving average (ARMA) model to extract evoked potentials from 2 segments. This was possible under two condiitons: high signal to noise ratio and clear separation of the EEG and EP spectra. These assumptions are not easy to fulfill, though. Cerutti et al. (1987) modeled background EEG activity by means of an AR process and event related brain activity by ARMA. In this way they were able to find a filter extracting single EP. Nevertheless, their method was not quite free of assumptions, since they since they used averaged EP to define their ARMA filter. In the following we shall briefly describe the method of the multiresolution decomposition and we will apply it to the analysis and reconstruction of single evoked potentials.     

Noise and signal power and their effects on evoked potential estimation

Signal power, noise power and their ratio (SNR) are important variables underlying estimation of evoked potential signals, yet, they are rarely explicitly considered in the design or analysis of EP experiments. A model is developed which relates the reliability of the average evoked potential (AEP) wave form to signal power, noise power, SNR, and the number of single trials included in the average. Measurements taken from auditory and visual EP experimental in elderly subjects show that noise power is highly reliable across experimental conditions and probably reflects global CNS anatomic or physiologic factors. In contrast, signal power and SNR are variable across conditions and sensory modalities, but are stable across replications. Thus signal power reflects CNS processes specific to the experimental paradigm. These results have importance for EP estimation. The expected reliability of the AEP cannot be adequately predicted from estimates of a subject's noise power, or from SNR estimated under different experimental conditions. These findings suggest the need for on-line estimation of SNR during data acquisition to ensure adequate reliability of AEPs.     

Exploratory analysis of longitudinal trials with staggered intervention times

Longitudinal trials involving surgical interventions commonly have subject-specific intervention times, due to constraints on the availability of surgeons and operating theatres. Moreover, the intervention often effects a discontinuous change in the mean response. We propose a nonparametric estimator for the mean response profile of longitudinal data with staggered intervention times and a discontinuity at the times of intervention, as an exploratory tool to assist the formulation of a suitable parametric model. We use an adaptation of the standard generalized additive model algorithm for estimation, with smoothing constants chosen by a cross-validation criterion. We illustrate the method using longitudinal data from a trial to assess the effect of lung resection surgery in the treatment of emphysema patients.     

Few-sweep estimation of evoked potential based on a generalized subspace approach

A generalized subspace approach is proposed for single channel brain evoked potential (EP) extraction from background electroencephalogram (EEG) signal. The method realizes the optimum estimate of EP signal from the observable noisy signal. The underlying principle is to project the signal and noise into signal and noise coefficient subspace respectively by applying projection matrix at first. Secondly, coefficient weighting matrix is achieved based on the autocorrelation matrices of the noise and the noisy signal. With the coefficient weighting matrix, we can remove the noise projection coefficients and estimate the signal ones. EP signal is then obtained by averaging the signals estimated with the reconstruction matrix. Given different signal-to-noise ratio (SNR) conditions, the algorithm can estimate the EP signal with only two sweeps observable noisy signals. Our approach is shown to have excellent capability of estimating EP signal even in poor SNR conditions. The interference of spontaneous EEG has been eliminated with significantly improved SNR. The simulation results have demonstrated the effectiveness and superior performance of the proposed method.     

Estimation of the noise autocorrelation function in auditory evoked potential applications

This work presents a simple and accurate method to estimate the noise autocorrelation function in auditory evoked potential applications. It basically consists in applying a conventional correlation function estimator over the contaminated evoked potential signal processed by a comb filter. The main feature of the proposed technique is the possibility of obtaining information on large correlation lags without the need of extra time intervals, minimizing the estimation time. A theoretical analysis is provided showing that, under certain but achievable conditions, the correlation function of the processed signal approximates the real noise correlation function. Simulation results and an example with a single-trial evoked potential estimation technique illustrate the expected performance. The proposed method is of special interest to either single or small number of trials evoked potential estimation techniques in anaesthesia monitoring applications.     

An automated laboratory control system: collection and analysis of behavioral and electro-physiological data.

A system of man-machine interactive PDP-11 assembly language programs is described which presents stimuli to a subject and records and analyzes behavioral and evoked potential data. The system was designed for researchers with no knowledge of computer programming and enables the user to create complicated sequences of stimulus presentations ("trials") and sequences of successive trials ("runs"), with no new programming required. The system is written for DEC.s DECLAB 11/40 system.     

Haptic Exploratory Behavior During Object Discrimination: A Novel Automatic Annotation Method

In order to acquire information concerning the geometry and material of handheld objects, people tend to execute stereotypical hand movement patterns called haptic Exploratory Procedures (EPs). Manual annotation of haptic exploration trials with these EPs is a laborious task that is affected by subjectivity, attentional lapses, and viewing angle limitations. In this paper we propose an automatic EP annotation method based on position and orientation data from motion tracking sensors placed on both hands and inside a stimulus. A set of kinematic variables is computed from these data and compared to sets of predefined criteria for each of four EPs. Whenever all criteria for a specific EP are met, it is assumed that that particular hand movement pattern was performed. This method is applied to data from an experiment where blindfolded participants haptically discriminated between objects differing in hardness, roughness, volume, and weight. In order to validate the method, its output is compared to manual annotation based on video recordings of the same trials. Although mean pairwise agreement is less between human-automatic pairs than between human-human pairs (55.7% vs 74.5%), the proposed method performs much better than random annotation (2.4%). Furthermore, each EP is linked to a specific object property for which it is optimal (e.g., Lateral Motion for roughness). We found that the percentage of trials where the expected EP was found does not differ between manual and automatic annotation. For now, this method cannot yet completely replace a manual annotation procedure. However, it could be used as a starting point that can be supplemented by manual annotation.     

Method for recording short latency evoked potentials using an EKG artifact elimination procedure

This paper describes the application of an EKG elimination procedure¹, previously reported by the authors, to evoked potentials (EPs) recording using a non-cephalic reference. The method consists of three separate steps: data acquisition, EKG artifact elimination, and EP averaging. EKG artifacts are eliminated from the raw EEG by applying a four step procedure to the simultaneously recorded EEG, EKG and stimulus pulse. The steps are: synchronized partition of the raw EEG, EKG averaging, synchronized repetition, and synchronized subtraction of the EKG estimate from the raw EEG. Average EP values are then obtained by averaging the processed EEG using the stimulus pulse as a trigger. Somatosensory evoked potentials to a hand reference, averaged using the proposed method, were compared with those obtained by two conventional averaging methods, and were shown to be more clearly defined. The advantage of the proposed method for recording short latency EP values with a non-cephalic reference is that it requires fewer sweeps and thus takes less time than other methods. The proposed method may also be applicable to the recording of other EP values.     

基于小波变换的单次诱发电位信号时频分析

比较小波变换和平均叠加两种方法提取“模拟自然阅读”刺激模式下的诱发脑电信号,分析其时频特性,并进行脑功能源分布定位分析。结果显示,采用平均叠加法来提取和分析诱发电位信号,损失了某些重要的诱发电位成分,且其功能源分布定位反映的只是等效功能源的静态过程;而使用小波变换和脑功能源定位来提取和分析单次诱发电位信号,既能观察到丰富的诱发电位成分,又能反映脑功能源的实时动态活动过程。这表明,小波变换下的时频分析是脑电信号处理的一种可行的新方法。     

Sample size reestimation by Bayesian prediction

We review a Bayesian predictive approach for interim data monitoring and propose its application to interim sample size reestimation for clinical trials. Based on interim data, this approach predicts how the sample size of a clinical trial needs to be adjusted so as to claim a success at the conclusion of the trial with an expected probability. The method is compared with predictive power and conditional power approaches using clinical trial data. Advantages of this approach over the others are discussed.     

Cross-trial correlation analysis of evoked potentials reveals arousal-related attenuation of thalamo-cortical coupling

We describe a computational method for assessing functional connectivity in sensory neuronal networks. The method, which we term cross-trial correlation, can be applied to signals representing local field potentials (LFPs) evoked by sensory stimulations and utilizes their trial-to-trial variability. A set of single trial samples of a given post-stimulus latency from consecutive evoked potentials (EPs) recorded at a given site is correlated with such sets for all other latencies and recording sites. The results of this computation reveal how neuronal activities at various sites and latencies correspond to activation of other sites at other latencies. The method was used to investigate the functional connectivity of thalamo-cortical network of somatosensory system in behaving rats at two levels of alertness: habituated and aroused. We analyzed potentials evoked by vibrissal deflections recorded simultaneously from the ventrobasal thalamus and barrel cortex. The cross-trial correlation analysis applied to the early post-stimulus period (<25 ms) showed that the magnitude of the population spike recorded in the thalamus at 5 ms post-stimulus correlated with the cortical activation at 6–13 ms post-stimulus. This correlation value was reduced at 6–9 ms, i.e. at early postsynaptic cortical response, with increased level of the animals’ arousal. Similarly, the aroused state diminished positive thalamo-cortical correlation for subsequent early EP waves, whereas the efficacy of an indirect cortico-fugal inhibition (over 15 ms) did not change significantly. Thus we were able to characterize the state related changes of functional connections within the thalamo-cortical network of behaving animals.     

Improved processing of the steady-state evoked potential

Two related procedures for estimating the parameters of steady-state evoked potentials (SSEPs) are introduced. The first procedure involves an initial stage of digital bandpass filtering followed by a Discrete Fourier Transform analysis. In the second method, a high resolution method based on parametric modelling is applied to the filtered data. The digital pre-filter consists of a non-phase shifting Chebychev bandpass filter. The parametric modelling method considers the evoked-response-plus-noise distribution to consist of a set of exponentially damped sinusoids. The frequency, amplitude, phase and damping factors of these components are estimated by calculating the mean of the forward and backward prediction filters and linear regression.We compared the signal-to-noise ratio (SNR) of the new procedures to the conventional Discrete Fourier Transform method for Monte Carlo simulations utilizing known sinusoids buried in white noise, known sinusoids buried in human EEG noise and for a sample of visual evoked potential data. Both of the new methods produce substantially more accurate and less variable estimates of test sinusoid amplitude. For VEP recording, the EEG background noise level is reduced by 5–6 dB over that obtained with the DFT. The new methods also provide approximately 5 dB better SNR than the DFT for detection of sinusoids based on the Rayleigh statistic. The parametric modelling approach is particularly suited for the analysis of very short data records including cycle-by-cycle analysis of the SSEP.     

Inferring synaptic inputs given a noisy voltage trace via sequential Monte Carlo methods

We discuss methods for optimally inferring the synaptic inputs to an electrotonically compact neuron, given intracellular voltage-clamp or current-clamp recordings from the postsynaptic cell. These methods are based on sequential Monte Carlo techniques ("particle filtering"). We demonstrate, on model data, that these methods can recover the time course of excitatory and inhibitory synaptic inputs accurately on a single trial. Depending on the observation noise level, no averaging over multiple trials may be required. However, excitatory inputs are consistently inferred more accurately than inhibitory inputs at physiological resting potentials, due to the stronger driving force associated with excitatory conductances. Once these synaptic input time courses are recovered, it becomes possible to fit (via tractable convex optimization techniques) models describing the relationship between the sensory stimulus and the observed synaptic input. We develop both parametric and nonparametric expectation-maximization (EM) algorithms that consist of alternating iterations between these synaptic recovery and model estimation steps. We employ a fast, robust convex optimization-based method to effectively initialize the filter; these fast methods may be of independent interest. The proposed methods could be applied to better understand the balance between excitation and inhibition in sensory processing in vivo.     

Signal detection in averaged evoked potentials: Monte Carlo comparison of the sensitivity of different methods

Many clinical and research applications rely on detecting evoked potential (EP) signal or EP differences between conditions. Statistical methods for objective signal detection should be sensitive to the presence of signal, but must provide the user strict control on tolerated false alarm rate. The respective sensitivities of 6 signal detection methods were compared through several Monte Carlo simulations involving 2 autocorrelation structures, 5% and 1% significance levels, 8, 10 or 12 replications per study, and increasing signal to noise ratio. The signal detection methods compared were: (1) the Record Orthogonality Test by Permutations (ROT-p), a variant of the Residual Orthogonality Test (Achim et al., 1988), that provides an unbiased estimate of the energy of the signal present in the averaged data, (2) the Tsum² permutation test of Karniski et al. (1994), (3) a Principal Component Analysis method (PC1) consisting of a t test on the weights of the first principal component, (4) multiple t tests on amplitudes with empirical adjustment for global false alarm rate, and (5–6) the test of Guthrie and Buchwald (1991) on length of consecutive t tests significant at P < 0.05 or 0.01 per-test. The first 3 methods did not exceed their nominal false alarm rate and clearly outperformed the last 3, with the ROT-p method being significantly more sensitive than all others under almost all conditions.     

Method for single-trial recordings of somatosensory evoked potentials

Evoked potentials (EPs) in response to stimuli are recorded from a human scalp contaminated with noise. To improve the signal-to-noise ratio, averaging methods have been widely used for the recorded data. However, when the waveforms of EP for each stimulus are not identical, the average waveform of the EP deteriorates. Variation of the EP waveform to each stimulus itself is important information for the EP. In this paper, a recording method for single somatosensory evoked potential (SEP) waveform is proposed, in which three kinds of band-pass filters were selectively used during three specific time sectors for each interstimulus interval. For the late section of the interval, an EEG waveform prediction method was applied to eliminate contaminated alpha rhythm components. By using the proposed method, we were successful in detecting the single SEP waveform.     

Statistical Tests Based on New Composite Hypotheses in Clinical Trials Reflecting the Relative Clinical Importance of Multiple Endpoints Quantitatively

In clinical trials, several endpoints (EPs) are often evaluated to compare treatments in some therapeutic area. Suppose that there are two EPs in a clinical trial. We propose a new set of composite hypotheses for continuous variables, taking the relative clinical importance of the EPs into account. The main hypotheses were formulated to show that a treatment is so superior to the control treatment, which is not necessarily a placebo, in one EP, that the possible non‐inferiority of the treatment by at most a certain value in the other EP can be compensated sufficiently, taking the clinical point of view into account. The maximum non‐inferiority margin of one EP might not be a biologically unimportant difference in exchange for much superiority of the other EP. This formulation leads to a new composite EP and a very simple test statistic. The intersection‐union principle was employed to derive the proposed test.     

Effect of local surface cooling of the rat cortex on the impulse activity of neurons assumed to be the sources of corticofugal influences

In an experiment on albino rats with electrodermal stimulation of the forepaw evoked potentials (EP) in the neostriatum (NS), the cortical primary response (PR), and impulse reactions of neurons (mainly of layers V and VI of the cortex) were recorded. The zone of leading-off of the potentials in the cortex was subjected to local surface cooling, which led to an increase in the PR amplitude. This facilitation was accompanied by a change in the time parameters of the impulse reactions of the cortical neurons: the latency and duration increased, and a rhythmic organization of activity appeared or intensified (if it was already present). The increase in the PR amplitude and number of spikes in the response of the cortical neurons to stimulus presentation was far less intensive than the sharp increase in EP amplitude in the NS, and did not correspond to it fully in time. The data suggest that the activating influence of the corticofugal signal on EP in the NS is determined not so much by the intensity of the descending signal as by its temporal organization.I. M. Sechenov Institute of Evolutionary Physiology and Biochemistry, Academy of Sciences of the USSR, Leningrad. Translated from Neirofiziologiya, Vol. 23, No. 2, pp. 181–189, March–April, 1991.     

PA-CRM: A continuous reassessment method for pediatric phase I oncology trials with concurrent adult trials

Pediatric phase I trials are usually carried out after the adult trial testing the same agent has started, but not completed yet. As the pediatric trial progresses, in light of the accrued interim data from the concurrent adult trial, the pediatric protocol often is amended to modify the original pediatric dose escalation design. In practice, this is done frequently in an ad hoc way, interrupting patient accrual and slowing down the trial. We developed a pediatric-continuous reassessment method (PA-CRM) to streamline this process, providing a more efficient and rigorous method to find the maximum tolerated dose for pediatric phase I oncology trials. We use a discounted joint likelihood of the adult and pediatric data, with a discount parameter controlling information borrowing between pediatric and adult trials. According to the interim adult and pediatric data, the discount parameter is adaptively updated using the Bayesian model averaging method. Numerical study shows that the PA-CRM improves the efficiency and accuracy of the pediatric trial and is robust to various model assumptions.     

Denoising functional magnetic resonance imaging time-series using anisotropic spatial averaging

We propose a novel iterative scheme for adaptive smoothing of functional MR images. The method estimates a signal model at every voxel in the time-series, which is subsequently used in determining the weights of the smoothing kernel. The method does not require any information about the test hypothesis and is well-suited as a preprocessing step for both hypothesis-driven and data-driven analysis techniques. We demonstrate the performance of the proposed method by applying it to preprocess both simulated and real fMRI data. The method is found to effectively suppress the noise while preserving the shapes of the active brain regions.