A. Casaleggio

Study of the correlation dimension of ECG signals based on MIT-BIH Arrhythmia Data Base ECGs

The correlation dimensions of random noise and electrocardiograms of both normal and sick subjects have been analyzed and compared. The noisy signal has been obtained by using the random function of a DEC VAX computer; ECG data have been obtained from the MIT-BIH Arrhythmia Data Base. It was expected that a lower dimensionality of the ECG signal would be obtained with respect to the noisy one. This hypothesis has been confirmed by the analysis. An attempt was made to evaluate differences in the dimensionality of sick and normal subjects. The complexity and time-consuming aspects of the computations made it possible to examine only four patients. Up to now, there has not been enough evidence of significant differences between sick and normal subjects, and thus it is necessary to increase the number of analyzed cases. In order to obtain a good saturation of the correlation dimension, the embedding space has been spanned from 2 to 50.<<ETX>>

Computational model of erratic arrhythmias in a cardiac cell network: the role of gap junctions.

Cardiac morbidity and mortality increases with the population age. To investigate the underlying pathological mechanisms, and suggest new ways to reduce clinical risks, computational approaches complementing experimental and clinical investigations are becoming more and more important. Here we explore the possible processes leading to the occasional onset and termination of the (usually) non-fatal arrhythmias widely observed in the heart. Using a computational model of a two-dimensional network of cardiac cells, we tested the hypothesis that an ischemia alters the properties of the gap junctions inside the ischemic area. In particular, in agreement with experimental findings, we assumed that an ischemic episode can alter the gap junctions of the affected cells by reducing their average conductance. We extended these changes to include random fluctuations with time, and modifications in the gap junction rectifying conductive properties of cells along the edges of the ischemic area. The results demonstrate how these alterations can qualitatively give an account of all the main types of non-fatal arrhythmia observed experimentally, and suggest how premature beats can be eliminated in three different ways: a) with a relatively small surgical procedure, b) with a pharmacological reduction of the rectifying conductive properties of the gap-junctions, and c) by pharmacologically decreasing the gap junction conductance. In conclusion, our model strongly supports the hypothesis that non-fatal arrhythmias can develop from post-ischemic alteration of the electrical connectivity in a relatively small area of the cardiac cell network, and suggests experimentally testable predictions on their possible treatments.

Linear and non-linear indices of heart rate variability in chronic heart failure: mutual interrelationships and prognostic value

We computed 3 linear and 20 nonlinear HRV indexes on 24-h Holter recordings from 200 stable CHF patients (age 52plusmn9 yrs, NYHA II-III, LVEF 24plusmn6%) with the aim to assess i) the mutual interrelationships between these indexes and ii) their prognostic value towards cardiac death. We found high correlations between variables, with potential bias in fitting survival models. To overcome this problem a clustering procedure was used, obtaining 11 clusters. Cox analysis showed that seven clusters were significantly associated with the study outcome (p<0.05) but, after adjustment for major clinical prognostic parameters, significance persisted only in 2 of them (both composed by nonlinear variables). Our results indicate that composite scores derived from nonlinear indices contain significant prognostic information independent of classical clinical predictors, highlighting the importance of non linear HRV parameters in prognostic stratification of CHF patients

Analysis of implantable cardioverter defibrillator signals for non conventional cardiac electrical activity characterization

Implantable cardioverter defibrillators (ICDs) can store intracardiac electrograms (EGMs) in sinus rhythm (SR), at the onset of spontaneous ventricular tachyarrhythmias (VT) or during their course. This allows the investigation of unknown features of the heart electrical activity associated with different cardiac rhythms. In this study we propose a non conventional cardiac electrical activity characterization (CEAC) that extracts quantitative information about the power spectrum wideness and variability of the beat-by-beat morphology. We analyze 293 EGMs from 40 patients who underwent implantation of St Jude Medical–Ventritex ICDs that allow the storage of EGMs with two different modes of recording: bipolar (BIP) and unipolar or far-field (FF). The EGMs are studied with this CEAC by (1) exploring differences between the CEAC measured from FF and BIP EGMs during similar cardiac rhythms, and (2) investigating the mode of recording that allows a better separation between SR and VT rhythms. Results show that, with similar cardiac rhythm, the CEACs from FF or BIP recordings are different (for SR rhythm: sensitivity 81.5%, specificity 93.6%; for VT rhythm: sensitivity and specificity 100%); thus FF and BIP EGMs analyze different aspects of cardiac activity.The CEAC applied to FF EGMs distinguishes better EGMs obtained during SR from VT rhythms (VT vs SR with sensitivity 92.7% and specificity 79.7%) than when it is applied to BIP signals (VT vs SR with sensitivity 60% and specificity 73.3%).

On cardiac activity characterization from Implantable Cardioverter Defibrillator electrogram analysis: is far-field better?

The new generation of Implantable Cardioverter Defibrillators (ICDs) allows storage of intracardiac electrograms (EGMs) recorded immediately before the onset of malignant ventricular tachyarrhythmias (VTs) and during their course. Although all devices base most of diagnostic parameters on the bipolar EGMs, some of them allow two types of EGM recordings: bipolar (potential is obtained between two electrodes at the tip of the electrocatheter) and far-field (potential is recorded between the tip of the electrocatheter and the active can device). The main objective of this work is to study which differences can be observed from the analysis of bipolar or far-field EGMs when deterministic chaos methods are being applied to bipolar or far-field EGM recordings. For this analysis we considered 20 EGMs obtained from 6 patients with ICDs. Some of the EGMs contain VT episodes, other are collected in basal heart conditions. Within the limited set of EGM considered in this work, results indicate that far field recordings, monitoring the whole heart activity, allow better characterization of heart substrate, and they are more indicated when deterministic chaos analysis would be applied.

Neural network for automatic anomalous QRS complex detection

An application of the back-propagation (BP) neural network (NN) for the discrimination between normal and pathological electrocardiogram (ECG) complexes is presented. The BP is used as a part of an unsupervised method: the network output has not been used to discriminate normal and pathological complexes, but only to extract the prototype complex of the analyzed ECG. An attempt is made to automatically individualize a pathological QRS morphology on those ECGs where anomalous premature ventricular contraction (PVC) beats were less than 15%. Results show a sensitivity of 0.991 and a specificity of 0.985.<<ETX>>

Study on the influence of a noisy environment on the ECG correlation dimension determination; possible use for noise estimation

It is shown how the determination of the correlation dimension (D/sub 2/), an algorithm which allows the estimation of a nonlinear system degrees of freedom, is affected by the presence of noise. The poor accuracy in computing D/sub 2/ is related to the noise fraction of the signal measurement. A very simple empirical way to adjust the D/sub 2/ value is presented. Some preliminary results on a rough way for the noise fraction estimation are given. This study has been done mainly on a Henon attractor, and D/sub 2/ adjustments have been tried on electrocardiogram (ECG) signals, yielding D/sub 2/ values ranging between 1 and 2, in healthy subjects.<<ETX>>

Low complexity methods for intracardiac atrial electrogram compression

This paper studies methods for intracardiac atrial electrograms compression suitable with implementation on implantable devices. Algorithms are based an piecewise linear approximation, beat detection, and their combination. Bipolar intracardiac electrograms were obtained during electrophysiological studies and divided into 3 rhythm groups. The authors analyzed 2196 seconds of sinus rhythm (SR), 786 sec. of atrial fibrillation (AF) and 1793 sec. of atrial flutter (AFL). Performance over the whole data base reach average compressed data rate (CDR) ranging between 500 and 850 (bits/second), and the percent of root mean square difference (PRD) varies from 2 to 12% depending on the choice of methods and accepted errors. Preliminary results show that time consumption can be reduced to enable real time compression with implanted devices.

Effect of constant intervals of a time series on the computation of the correlation dimension

The author discusses the computation of the correlation dimension (D/sub 2/) of a time series in the particular case of a signal which is constant for a significant fraction of its length. Two examples are investigated, a periodical signal obtained from a synthesized electrocardiogram (ECG) and the z variable of a Rossler nonlinear system. Such signals are chosen because their correlation dimension values are known. It is shown that the D/sub 2/ obtained by using the algorithm of P. Grassberger and I. Procaccia yields, in both cases, an underestimation of the correct D/sub 2/ value. This problem is discussed and a slight modification of the basic algorithm to reduce this error is proposed.<<ETX>>

Some results on a fractal-like behaviour of ECG signals

Electrocardiograms (ECGs) of two normal subjects were studied using methods of nonlinear dynamics, both in supine rest and standing up. For each signal, the authors computed power spectrum, phase-space portrait, and correlation dimension. They found that the power spectrum shows a 1/f behavior at high frequencies, and that the correlation dimension is 4.7+or-0.3, suggesting fractal behavior of ECG signals.<<ETX>>