Gianfranco Passariello

Atrial activity enhancement by Wiener filtering using an artificial neural network

Describes a novel technique for the cancellation of the ventricular activity for applications such as P-wave or atrial fibrillation detection. The procedure was thoroughly tested and compared with a previously published method, using quantitative measures of performance. The novel approach estimates, by means of a dynamic time delay neural network (TDNN), a time-varying, nonlinear transfer function between two ECG leads. Best results were obtained using an Elman TDNN with 9 input samples and 20 neurons, employing a sigmoidal tangencial activation in the hidden layer and one linear neuron in the output stage. The method does not require a previous stage of QRS detection. The technique was quantitatively evaluated using the MIT-BIH arrhythmia database and compared with an adaptive cancellation scheme proposed in the literature. Results show the advantages of the proposed approach, and its robustness during noisy episodes and QRS morphology variations.

Multisensor fusion for atrial and ventricular activity detection in coronary care monitoring

Information management for critical care monitoring is still a very difficult task. Medical staff are often overwhelmed by the amount of data provided by the increased number of specific monitoring devices and instrumentation, and the lack of an effective automated system. Specifically, a basic task such as arrhythmia detection still produce an important amount of undesirable alarms, due in part to the mechanistic approach of current monitoring systems. In this work, multisensor and multisource data fusion schemes to improve atrial and ventricular activity detection in critical care environments are presented. Applications of these schemes are quantitatively evaluated and compared with current methods, showing the potential advantages of data fusion techniques for event detection in noise corrupted signals.

An algorithm for microprocessor-based QRS detection

QRS detection is fundamental for intensive care cardiac instrumentation, such as in arrhythmia recognition or ambulatory recording of cardiac events. Reliable detection of the QRS is necessary in order to obtain accurate measurements of the R-R intervals and QRS width. This paper describes a real-time algorithm for QRS detection based upon the theory of maxima and minima to locate peaks in the digitized ECG. A three-point sliding window is used to determine the presence of peaks and valleys on the signal. Analysis of the difference in amplitude and the time interval between consecutive peaks and valleys allows QRS detection. Thresholds used for detection are adapted according to changes in the amplitude and duration of the previous QRS complexes. Testing was performed with a normal ECG recording contaminated with different levels of added Gaussian noise, giving an average detection rate from 99.7 percent True Positive for a signal-to-noise ratio of 19 dB to 86.5 percent for a low SNR of 7 dB.QRS detection is fundamental for intensive care cardiac instrumentation, such as in arrhythmia recognition of ambulatory recording of cardiac events. This paper describes a real-time algorithm for QRS detection based upon the theory of maxima and minima to locate peaks in the digitized ECG. A three-point sliding window is used to determine the presence of peaks and valleys on the signal. Analysis of the difference in amplitude and the time interval between consecutive peaks and valleys allows QRS detection. Thresholds used for detection are adapted according to changes in the amplitude and duration of the previous QRS complexes. Testing was performed with a normal ECG recording contaminated with different levels of added Gaussian noise, giving an average detection rate from 99. 7 percent True Positives for a signal-to-noise ratio of 19 db to 86. 5 percent for a low snr of 7 db.

Stress ECG and Laboratory Database for the Assessment of Diabetic Cardiovascular Autonomic Neuropathy

Development of a diabetic patient database in order to study cardiovascular autonomic neuropathy (CAN) using as a primary source, stress ECG is presented. The selected platform (ecgML) allows user-friendly environment to analyze and interpret graphs, signals and data. It also allows the ability to perform annotations and reports done by users from different fields. In order to feed the database, the input data is codify using MatLab. The database is composed by two populations: 1) Type 2 Diabetes mellitus group and 2) a control group with no medical history of cardiovascular disease. At the present, there are 62 records available from these two groups. The database also contains laboratory parameters, concurrent medical diagnoses reports verified by cardiologists and other clinicians, automatic annotations for each beat and trend series from parameters extracted from the ECG signals such as RR intervals and ST segment measurements. All this information will become very useful for CAN investigations.

Border detection in echocardiography images using texture analysis

A tool for image texture analysis (TITA) has been developed for application to two-dimensional echocardiographic images. Through a user-friendly environment, spatio-temporal changes in texture can be analyzed to characterize the tissue or to provide segmentation and border detection. As an example of the application of TITA, the implementation of an inverse difference moment based algorithm for the extraction of endocardial boundaries from echocardiographic images is presented. The proposed method has been applied to apical four chamber and paraesternal short axis views, showing excellent results even in difficult images. It is possible to conclude that the difference image of echocardiographic frames contains sufficient information for accurate detection of borders and regions.<<ETX>>

A methodology for measuring the chronotropic and inotropic effects during autonomic tests

A methodology for autonomic testing, that reflects both chronotropic and inotropic effects and their relationship is presented. ECG and impedance cardiogram (ICG) are recorded from the subject during each test. The method provides plots of heart rate, pre-ejection period and blood ejection velocity. This non-invasive technique has been applied to normal subjects and patients with autonomic disorders. Some analysis techniques are proposed, as future research directions, to improve the clinical application of the method.

Real-time Detection And Quantification Of Ischemic ECG Changes

A real-time algorithm, for one lead ST-segment analysis, that can be implemented in intensive coronary care bedside monitors, has been developed. The algorithm allows QRS detection, beat labeling and alignment, averaging, b ase-line correction, and quantification of the ST-segment. A trend graphic plot of ST-deviation or ST-slope vs. time is generated as the output of the algorithm. In the absence of an adequate data base, we have tested the algorithm with selected exercise ECG recordings, records obtained during angioplasty, and four strips f rom the lnstituto d i Fisiologia Clinica di Pisa data base, with very encouraging results.

Protocol for cardiac assessment of recreational athletes

In this work, the development of a database on physical fitness is presented. As initial population to fill this database, people who practice recreational sports at the Universidad Simon Bolivar (USB) were chosen. The goal was studying individual physical fitness in order to structure exercise routines that gives certain benefits without risking the individual health, promoting a less sedentary way of life. Before the study, a low-cost noninvasive protocol was designed to determine the level of physical fitness. The methodology consisted of four steps: a) A review of existing protocols to propose a set of physical fitness (International Physical Activity Questionnaire (IPAQ)), cardiovascular (heart rate variability, heart rate recovery time and arterial blood pressure), anthropomorphic, aerobic (maximum oxygen consumption) and mood state (Profile of Mood State (POMS)) measurements, which allow sketching a complete profile on the sportsman physical fitness. b) Instrumental data collection. c) Electrocardiographic signal processing. d) Data post-processing using multivariate analysis. The database was composed of 26 subject from USB. Ten subjects were soccer players, ten were mountain climbers and six were sedentary people. Results showed that the heart rate recover time after 2-3 min, IPAQ and maximum oxygen consumption have higher weights for classifying individuals according to their habitual physical activity. Heart rate variability, as well as, POMS did not contribute greatly for discriminating recreational sport from sedentary persons.

Evaluación de dos Métodos para la Segmentación del Ancho de la Onda T en el ECG

In this work, two methods for T wave segmenta- tions were compared: classic method based on differentiator filter and the method based on slope estimation. This study was carried out on simulated electrocardiogram (ECG) signals with four simulated added noise types: (a) baseline wander due to breathing (0-0.5 Hz), (b) motion artifacts (3-5 Hz), (c) elec- tromyography and motion artifacts (Gaussian white noise), (c) mixed effects (sum of the previous noises). Four conditions of Signal to Noise Ratio (SNR) were computed: 5, 10, 15, 20 and 25 dB. The beginning, the end and the width of the T wave was determined with both methods and the mean absolute error was computed for all signals. The slope estimation method shows a T wave width larger than the obtained with the differ- entiator filter. Both methods showed that the end location is easier to find than the beginning of the T wave. The achieved performance shows a satisfying behavior of both methods in favorable conditions of SNR. However, the differentiator filter method shows better performance that the method base on slope estimation. The noise due to motion artifacts affect greatly the mean absolute error in the beginning and end loca- tion of the T wave. An important increase of T wave width variability induced by the artifacts was observed.

Estimation of the deformation field for the left ventricle walls in 4-d multislice computerized tomography

This paper describes a method for estimating the deformation field of the Left Ventricle (LV) walls from a 4–D Multi Slice Computerized Tomography (MSCT) database. The approach is composed of two stages: in the first, a 2–D non–rigid correspondence algorithm matches a set of contours on the LV at consecutive time instants. In the second, a 3–D curvature–based correspondence algorithm is used to optimize the initial approximate correspondence. The dense displacement field is obtained based on the optimized correspondence. Parameters like LV volume, radial contraction and torsion are estimated. The algorithm is validated on synthetic objects and tested using a 4–D MSCT database. Results are promising as the error of the displacement vectors is 2.69 ± 1.38 mm using synthetic objects and, when tested in real data, local parameters extracted agree with values obtained using tagged magnetic resonance imaging.