BACKGROUND Conventional mapping of ventricular tachycardia (VT) after myocardial infarction is limited in patients with hemodynamically untolerated or noninducible VT.OBJECTIVES The purpose of this study was to develop a unique strategy using noncontact unipolar mapping to define infarct substrate and VT circuits.METHODS Dynamic substrate mapping (DSM) was performed in seven pigs with healed anterior myocardial infarction. This technique defined substrate as the intersection of low-voltage areas identified in sinus rhythm and during pacing around the infarct. Pacing was also performed within the substrate to determine exit sites.RESULTS Anteroapical transmural scar was identified in all animals. A mean of three pacing sites was used for substrate definition. The mean area (+/- SD) was 18.4 +/- 8.8 cm(2) by DSM and 15.4 +/- 6.9 cm(2) by pathology (P > .5). A mean of 4.5 sites was paced within substrate. Ten of 18 paced wavefronts exited substrate adjacent to the pacing area, seven exited at distant areas, and one had two exits. VT was induced in five animals (1.6 morphologies per animal). Except for one VT, circuit exit sites were identified at substrate borders on the endocardium. VT exit sites were at (n = 6) or near (it = 3) a pacing exit site. Electrogram voltages differed significantly between substrate, border, and nonsubstrate areas in infarcted animals and in comparison with control animals. No Substrate was identified in two control animals.CONCLUSION DSM is a reliable method for infarct substrate localization in this model. Pacing within substrate can predict VT exit sites and may prove useful for ablation of unmappable VT after myocardial infarction.

We have designed a multirate digital signal processing algorithm to detect heartbeats in the electrocardiogram (ECG), The algorithm incorporates a filter bank (FB) which decomposes the ECG into subbands with uniform frequency bandwidths, The FB-based algorithm enables independent time and frequency analysis to be performed on a signal. Features computed from a set of the subbands and a heuristic detection strategy are used to fuse decisions from multiple one-channel beat detection algorithms. The overall beat, detection algorithm has a sensitivity of 99.59% and a positive predictivity of 99.56% against the MIT/BIH database. Furthermore this is a real-time algorithm since its beat detection latency Is minimal. The FB-based beat detection algorithm also inherently lends itself to a computationally efficient structure since the detection logic operates at the subband rate, The FB-based structure is potentially useful for performing multiple ECG processing tasks using one set of preprocessing filters.

The classification of heart beats is important for automated arrhythmia monitoring devices. The study describes two different classifiers for the identification of premature ventricular complexes (PVCs) in surface ECGs. A decision-tree algorithm based on inductive learning from a training set and a fuzzy rule-based classifier are explained in detail. Traditional features for the classification task are extracted by analysing the heart rate and morphology of the heart beats from a single lead. In addition, a novel set of features based on the use of a filter bank is presented. Filter banks allow for time-frequency-dependent signal processing with low computational effort. The performance of the classifiers is evaluated on the MIT-BIH database following the AAMI recommendations. The decision-tree algorithm has a gross sensitivity of 85.3% and a positive predictivity of 85.2%, whereas the gross sensitivity of the fuzzy rule-bassed system is 81.3%, and the positive predictivity is 80.6%.

The authors propose a fuzzy logic system for the classification of premature ventricular beats (PVCs) in ECGs. The classifier uses novel features extracted from a time-frequency analysis performed by a filter bank in addition to measurements related to the timing of R-R intervals and morphology analysis. The performance of the algorithm is evaluated on the MIT-BIH Arrhythmia Database following the AAMI recommendations. The achieved sensitivity is 81.34% and the positive predictivity is 80.64%.

The proposed algorithm uses a Filter Bank (FB) based method to detect heart beats in the electrocardiogram (ECG). The FB-based method allows for time and frequency dependent analysis to be performed on a signal. The beat detection algorithm operates at a reduced rate as compared to the input signal rate. The beat detection accuracy is comparable to that of other algorithms reported in the literature.

We propose an algorithm which can potentially perform multiple ECG processing tasks using a filter bank (FB). One set of filters in the FB decomposes the ECG into uniform frequency subbands. Since the subbands have a narrower bandwidth than the input ECG, they can be downsampled to get a lower rate subband signal. Time and frequency dependent processing can be performed at a lower rate and hence reduce the computation cost. A beat detection algorithm is presented which has minimal detection latency and good beat detection performance on the MIT/BIH database. The ECG can be enhanced by processing the subbands to remove noise. Features computed from the subbands can be used to distinguish between some ventricular and sinus beats. The FB offers a strategy to perform multiple tasks on the ECG using one set of filters operating at a computationally efficient rate.

There are two predominant types of noise that contaminate the electrocardiogram (EGG) acquired during a stress test: the baseline wander noise (BW) and electrode motion artifact, and electromyogram-induced noise (EMG). BW noise is at a lower frequency, caused by respiration and motion of the subject or the leads. The frequency components of BW noise are usually below 0.5 Hz, and extend into the frequency range of the ST segment during a stress test. EMG noise, on the other hand, is predominantly at higher frequencies, caused by increased muscle activity and by mechanical forces acting on the electrodes. The frequency spectrum of the EMG noise overlaps that of the ECG signal and extends even higher in the frequency domain. In this article, the authors review some of the published ECG enhancing techniques to overcome the noise problems, and compare their performance on stress ECG signals under adverse noise scenarios. They also describe the filter bank-based ECG enhancing algorithm.

The algorithm presented is based on subband processing of the electrocardiogram (EGG), using Filter Banks (FB), and removes noise from stress-type ECGs. For Gaussian noise the FB-based method improves the signal-to-noise ratio (SNR), significantly better than the Mean and Median averaging methods. For muscle noise the FB-based method improves the SNR comparatively better than the Mean and Median averaging methods. The FB-based algorithm offers a way to process specific time periods in the heart beat cycle and remove noise in specific frequency bands.

The short time Fourier transform (STFT), smoothed pseudo Wigner Ville distribution (SPWVD), and cone-shaped kernel distribution (CKD) have been used to compare the time-frequency distribution of normal sinus rhythm, ventricular tachycardia, ventricular flutter, and ventricular fibrillation signals. This work is a pilot study to illustrate that the CKD and SPWVD have better time and frequency resolution than the STFT. It demonstrates that accurate methods of computing the time-frequency domain should be found for ECG signals. Only then should future work be done to design discriminatory features and classifiers for arrhythmias

The authors computed the discrete Wigner distribution (DWD) of the instantaneous heart rate for six male subjects. The power of the ventilation frequency decreased during the segment of the experimental protocol that included no breathing. Parameters extracted from the time-frequency distribution corresponding to respiratory sinus arrhythmia may be useful in conjunction with a typical apnea detection algorithm that is prone to false alarms due to movement artifacts in the ventilation signal

The authors have investigated potential applications of artificial neural networks for electrocardiographic QRS detection and beat classification. For the task of QRS detection, the authors used an adaptive multilayer perceptron structure to model the nonlinear background noise so as to enhance the QRS complex. This provided more reliable detection of QRS complexes even in a noisy environment. For electrocardiographic QRS complex pattern classification, an artificial neural network adaptive multilayer perceptron was used as a pattern classifier to distinguish between normal and abnormal beat patterns, as well as to classify 12 different abnormal beat morphologies. Preliminary results using the MIT/BIH (Massachusetts Institute of Technology/Beth Israel Hospital, Cambridge, MA) arrhythmia database are encouraging.

In this paper, we determined which electrode types, sizes, and locations were best suited for impedance-based ventilation measurement. Optimal electrodes provide high signal-to-(motion) artifact ratio (SAR) and reliability by meeting the following criteria: 1) low baseline impedance, 2) high adhesion, 3) good physical stability, 4) large effective area, 5) thin with high flexibility. We compared 14 electrodes from two main groups: adhesive-gel and conductive rubber electrodes. Adhesive-gel electrodes are easy to apply, make good body contact, and do not slip during the course of an experiment. We found that higher SAR's are obtained when electrode area is increased by connecting several small electrodes together rather than by using a single electrode with a larger area. The peak SAR is achieved when two electrode arrays (area = 70 cm2) are centered at the 8th intercostal spaces on opposite midaxillary lines. To determine the optimal electrode locations, we placed 32 electrodes on the trunk and recorded impedance between 171 electrode combinations on ten normal adult subjects. Based on these data, we conclude that the SAR's are highest when one electrode is placed on the midpoint between the left and right second intercostal spaces on the sternum and the other electrode is placed in the opposite position on the back.