PubMed Health⌕ Search

Biomedical subjects

N V Thakor

Publications and source records attributed to N V Thakor.

At least 19 recordsLinked to original sources
PubMed1992

Adaptive filter for event-related bioelectric signals using an impulse correlated reference input: comparison with signal averaging techniques.

Many bioelectric signals result from the electrical response of physiological systems to an impulse that can be internal (ECG signals) or external (evoked potentials). In this paper an adaptive impulse correlated filter (AICF) for event-related signals that are time-locked to a stimulus is presented. This filter estimates the deterministic component of the signal and removes the noise uncorrelated with the stimulus, even if this noise is colored, as in the case of evoked potentials. The filter needs two inputs: the signal (primary input) and an impulse correlated with the deterministic component (reference input). We use the LMS algorithm to adjust the weights in the adaptive process. First, we show that the AICF is equivalent to exponentially weighted averaging (EWA) when using the LMS algorithm. A quantitative analysis of the signal-to-noise ratio improvement, convergence, and misadjustment error is presented. A comparison of the AICF with ensemble averaging (EA) and moving window averaging (MWA) techniques is also presented. The adaptive filter is applied to real high-resolution ECG signals and time-varying somatosensory evoked potentials.

Algorithms↗
PubMed1991

Applications of adaptive filtering to ECG analysis: noise cancellation and arrhythmia detection.

Several adaptive filter structures are proposed for noise cancellation and arrhythmia detection. The adaptive filter essentially minimizes the mean-squared error between a primary input, which is the noisy ECG, and a reference input, which is either noise that is correlated in some way with the noise in the primary input or a signal that is correlated only with ECG in the primary input. Different filter structures are presented to eliminate the diverse forms of noise: baseline wander, 60 Hz power line interference, muscle noise, and motion artifact. An adaptive recurrent filter structure is proposed for acquiring the impulse response of the normal QRS complex. The primary input of the filter is the ECG signal to be analyzed, while the reference input is an impulse train coincident with the QRS complexes. This method is applied to several arrhythmia detection problems: detection of P-waves, premature ventricular complexes, and recognition of conduction block, atrial fibrillation, and paced rhythm.

Algorithms↗
PubMed1991

A massively parallel computer model of propagation through a two-dimensional cardiac syncytium.

A computer model of electrical propagation through a two-dimensional (2D) sheet of cardiac tissue has been developed to run on the massively parallel processor Connection Machine (CM-2) computer. The transmembrane ionic currents in each of 16,384 (128 x 128) 100 x 100 microns 2 patches of cardiac tissue are described by modified Beeler-Reuter membrane equations. These equations, along with the parabolic differential equation derived from 2D cable theory, are solved in parallel to study normal and abnormal 2D propagation. The sheet is paced with planar waves at a basic cycle length of 500 msec (control). When a premature ectopic stimulus of sufficient strength and appropriate timing is then applied to a local region of the syncytium, one of two types of reentry is observed: (a) stable figure-of-eight reentry, or (b) unstable but self-sustaining "fibrillation-like" reentry. During this fibrillatory activity, action potential durations are 79.8 +/- 36.8 msec (control = 244.9 +/- 0.9 msec) and coupling intervals average 96.7 +/- 31.3 msec (control = 500 +/- 0 msec). We also observed that passive electrotonically-induced depolarization of already refractory tissue extended the refractory period of that tissue, and that the duration of this extension depended on the magnitude of the electrotonic effect.

Action Potentials↗
PubMed1990

Noise reduction in biological step signals: application to saccadic EOG.

A weighted filter for noise reduction in nonrecurrent step signals where adaptive filtering cannot be applied is described. An optimal correction of a conventional finite impulse response (FIR) filter is achieved by using a priori knowledge of noise variance and a continuous estimation of the error signal's power. The weighted filter provides an optimal compromise between noise filtering and distortionless tracking. The prior knowledge required is that of the noise power and the lowest frequency in the noise spectrum. Application of the weighted filter to the saccadic electro-oculogram (EOG) results in better estimations of saccade duration and velocity.

Electricity↗
PubMed1990

Low-pass differentiators for biological signals with known spectra: application to ECG signal processing.

Digital low-pass filtering and differentiation (LPD) are useful in real-time processing of many biomedical signals. A general method is presented for determining the coefficients of a differentiator that maximizes the signal-to-noise ratio or minimizes the error between actual and ideal LPD filters, when signal and noise spectra are known. Several examples of digital filters suitable for QRS complex and P-T wave processing in ECG are presented.

Electrocardiography↗
PubMed1990

Ventricular tachycardia and fibrillation detection by a sequential hypothesis testing algorithm.

An algorithm for detecting ventricular fibrillation (VF) and ventricular tachycardia (VT) by the method of sequential hypothesis testing is presented. The algorithm first generates a binary sequence by comparing the signal to a threshold. The probability distribution of the time intervals of the binary sequence is obtained, and Wald's sequential hypothesis testing procedure is next employed to discriminate the arrhythmias. Sequential hypothesis testing of 85 cases resulted in identification of 1) 97.64% VF and 97.65% VT episodes after 5 s, and 2) 100% identification of both VF and VT after 7 s. The desired false positive and false negative error probabilities can be preprogrammed into the algorithm. An important feature of the sequential method is that extra time for detection can be traded off for improved accuracy, and vice versa.

Algorithms↗
PubMed1990

The defibrillation success rate versus energy relationship: Part I--Curve fitting and the most efficient defibrillation energy.

The effect of applying an energy pulse to the heart during ventricular fibrillation is described by the probability of successful defibrillation or success rate. Seven to ten (8.60 +/- 0.84: mean +/- standard deviation) defibrillation trials per energy were randomly attempted at energies which span the defibrillation success rate versus energy curve. We obtained 70.0 +/- 8.4 episodes per dog. We fit the defibrillation success rate versus energy relationship from ten dogs (20.5 +/- 1.5 kg) to four types of curves: linear, exponential, probit transformed linear, and logit transformed linear. The correlation coefficients for each fit are 0.917 +/- 0.057, 0.944 +/- 0.014, 0.926 +/- 0.51, and 0.889 +/- 0.098, respectively. We therefore conclude that the exponential curve best describes the DSRE relationship. This suggests the existence of an energy below which defibrillation does not occur. At higher energies, the exponential curve asymptotically approaches a 100% success rate, which indicates that increasing the energy produces a diminishing benefit to defibrillation success rate. The estimated energies with a 0% defibrillation success rate are surprisingly consistent among dogs, with 2.072 +/- 0.553 J. The estimated energy with an 80% defibrillation success rate is 5.217 +/- 1.091 J. The estimated defibrillation success rate corresponding to the defibrillation threshold of 3.59 +/- 1.06 J is consistent with 0.516 +/- 0.144. The estimated energies with a 0% success rate correlate well with the defibrillation thresholds with R = 0.772; P = 0.0088. Since implantable defibrillators have a limited energy supply, we determined energy efficiency by dividing defibrillation success rate by the applied energy and energy consumption by dividing the applied energy by the defibrillation success rate. The most efficient defibrillation energy occurs at the maximum energy efficiency and the minimum energy consumption. The most efficient defibrillation energy of 4.34 +/- 0.97 J determined from the exponential fit has a success rate of 0.70 +/- 0.06. The most efficient defibrillation energy can be predicted from the defibrillation threshold. Clinically, a 70% success rate may not be adequate. We, therefore, compared the energy efficiency and consumption of energies with 90% and 95% success rates to the most efficient defibrillation energy. About a 50% increase in energy from the most efficient defibrillation energy is necessary for a 90% success rate which results in about a 13% loss in energy efficiency and about a 16% increase in energy consumption. About an 84% energy increase is necessary for a 95% success rate which results in about a 24% loss in energy efficiency and about a 33% increase in energy consumption.(ABSTRACT TRUNCATED AT 400 WORDS)

Animals↗
PubMed1990

The defibrillation success rate versus energy relationship: Part II--Estimation with the "bootstrap".

Seventy or so defibrillation trials were typically attempted to determine the relationship between defibrillation success rate and energy (DSRE). Clinically, it may be desirable to estimate the DSRE relationship with fewer trials. We used the statistical resampling technique called the "bootstrap" to determine the number of defibrillation trials necessary for an accurate estimation of the DSRE relationship. The bootstrap technique assumes that the observed database is the maximum likelihood sample of the estimated population. The observed database is repeatedly resampled to produce a large bootstrap data-base and the bootstrap best estimate of a statistic is determined. DSRE data were obtained from ten dogs (20.5 +/- 1.5 kg). We bootstrapped our experimental DSRE data by two methods: (1) randomly choosing with replacement a specified number of defibrillation trials per energy; and (2) randomly choosing with replacement a specified number of defibrillation trials per bootstrap replication. For both bootstrap techniques, 100 replications were made. We performed a linear regression analysis on the bootstrap success rates and the observed success rates determined from 71.0 +/- 6.8 defibrillation attempts from each of the ten dogs. We concluded that 28 defibrillation trials are necessary to estimate the observed DSRE relationship with a correlation coefficient of 0.95.

Animals↗
PubMed1989

Success rate versus defibrillation energy: temporal profile and the most efficient defibrillation threshold.

To determine the temporal profile of the energy requirement for defibrillation, shocks were delivered to canine hearts after 5, 10, or 20 seconds from the onset of fibrillation with a combination of patch and catheter electrodes. A total of 956 fibrillation-defibrillation sequences were performed at one of four energy levels appropriately selected for each period of fibrillation in 10 anesthetized dogs. The energy values related to 50% (E50) and 80% (E80) of the predicted success were calculated from a logistic regression curve. The E50 and E80 values at 10 seconds after the onset of fibrillation were less than those at 20 seconds after the onset by 7.1% +/- 18.3% and 9.7% +/- 21.4%, respectively; differences were not significant. At 5 seconds after the onset, the differences were 15.3% +/- 14.2% (p less than 0.02) and 16.4% +/- 12.7% (p less than 0.01), respectively. The defibrillation energy efficiency was assessed by dividing the success rate (SR) of fibrillation by the applied energy (E). The maximal SR/E at 5, 10, and 20 seconds of fibrillation was achieved at the energy corresponding to the SRs of 88.8% +/- 4.5%, 90.4% +/- 3.9%, and 88.1% +/- 4.6%, respectively. We conclude that the energy requirement for defibrillation increases with the duration of fibrillation, even shortly after the onset of fibrillation, and the maximal energy efficiency is attained at the energy associated with the SR of approximately 90%.

Animals↗
PubMed1989

The effect of an unsuccessful subthreshold shock on the energy requirement for the subsequent defibrillation.

The effect of an unsuccessful subthreshold shock on the energy requirement for the subsequent defibrillation was studied in 10 anesthetized dogs. Defibrillation was achieved with a spring catheter electrode in the superior vena cava and a patch electrode on the anteroapical ventricular wall. Success rates of defibrillation 20 seconds from the onset of ventricular fibrillation were determined at three energy levels with and without a preceding subthreshold shock. Altogether, 637 episodes of fibrillation-defibrillation were performed (63.7 +/- 6.7 per dog). Predicted energy levels for defibrillation success rates of 50% and 80% (E50 and E80) acquired from a logistic regression curve were 0.0303 +/- 0.0064 and 0.0367 +/- 0.0069 joule/gm, respectively, without subthreshold shocks. E50 and E80 with an unsuccessful subthreshold shock resulted in comparable values (E50: 0.0325 +/- 0.0041 joule/gm; E80: 0.0.380 +/- 0.0100 joule/gm). Our results suggest that an unsuccessful low-energy shock does not alter the energy requirement for subsequent defibrillation with an implantable defibrillator.

Animals↗
PubMed1989

Three-dimensional computer model of the heart: fibrillation induced by extrastimulation.

We present a three-dimensional (3D) computer model that simulates electrical activity in the heart during fibrillation. A real dog heart is discretized to form 1473 interconnected cubic elements. The model exhibits normal activation and recovery from pacing. Five or more extrastimuli induce a self-sustaining tachyarrhythmia that soon degenerates into a fibrillatory rhythm. The extrastimuli increase the excitability of the myocardial cell population. The result is a rapid re-excitation of cells that allows for only a partial recovery of cell action potential. This suggests that a dispersion of refractory states of the cell population is the cause of fibrillation in this computer model.

Animals↗
PubMed1989

Adaptive Fourier estimation of time-varying evoked potentials.

An estimation procedure for dealing with time-varying evoked potentials is presented here. The evoked response is modeled as a dynamic Fourier series and the Fourier coefficients are estimated adaptively by the least mean square algorithm. Approximate expressions have been developed for the estimation error and time constant of adaptation. A procedure for optimizing the estimator performance is also presented. The effectiveness of the estimator in dealing with simulated as well as actual evoked responses is demonstrated.

Evoked Potentials, Visual↗
PubMed1987

ECG waveform analysis by significant point extraction. I. Data reduction.

We present a new technique for automatic data reduction and pattern recognition of time-domain signals such as electrocardiogram (ECG) waveforms. Data reduction is important because only a few significant features of each heart beat are of interest in pattern analysis, while the patient data collection system acquires an enormous number of data samples. We present a significant point extraction algorithm, based on the analysis of curvature, that identifies data samples that represent clinically significant information in the ECG waveform. Data reduction rates of up to 1:10 are possible without significantly distorting the appearance of the waveform. This method is unique in that common procedures help in both data reduction as well as pattern recognition. Part II of this work deals specifically with pattern analysis of normal and abnormal heart beats.

Algorithms↗
PubMed1987

ECG waveform analysis by significant point extraction. II. Pattern matching.

From a set of significant points which characterizes the ECG waveform, the pattern matching algorithm detects and classifies QRS complexes. R waves are detected from the analysis of global curvature. Next, the morphology of the QRS complex is determined. QRS complexes with different morphologies are classified by a correlation algorithm. This method is sensitive to changes in shape, such as that of abnormal QRS complexes. The algorithm should be useful in automated analysis of waveforms, such as ECG signals recorded in clinical environments.

Algorithms↗