Raul Llinares

Semi-blind source extraction of atrial activity by combining statistical and spectral features

Atrial fibrillation is the most common human arrhythmia. During atrial fibrillation episodes, the surface electrocardiogram contains the linear superposition of the atrial and ventricular rhythms in addition to other non-cardiac artifacts. Since these signals can be considered statistically independent, a Blind Source Separation (BSS) approach fits the problem properly. The signal that contains useful clinical information is the atrial one. We present a solution that focuses on the extraction of the atrial activity, enforcing simultaneously the statistical and temporal properties of the atrial signal. In addition, we propose the use of kurtosis as a parameter to measure the quality of the extraction. The algorithm is applied successfully to synthetic and real data. It improves the extraction of the atrial signal in comparison to other BSS methods, recovers only the interesting atrial rhythm using the information contained in all the leads and reduces the computational cost. The results obtained are shown to be highly satisfactory, with an average of 53.9% of spectral concentration, -0.04 of kurtosis value, 2.98 of ventricular residua and 4.77% of significant QRS residua over a database of thirty patients.

Application of constrained independent component analysis algorithms in electrocardiogram arrhythmias

OBJECTIVES The extraction of the atrial activity in atrial fibrillation episodes is a must for clinical purposes. During atrial fibrillation arrhythmia, the independent atrial and ventricular signals are superposed in the electrocardiogram, fulfilling the independent component analysis (ICA) model. We propose three new algorithms that constrain the classical ICA solution to fit the spectral content of the atrial component. This constraint allows the statement of the problem in terms of semiblind source extraction instead of blind source separation (BSS), in the sense that we only recover one source and we exploit the prior information about the sources in the extraction process. METHODS AND MATERIALS The methods used are extensions of classical BSS methods based on second and higher order statistics. We exploit the prior assumption about the sources in order to obtain the source extraction algorithms that are focused on the extraction of the atrial component. The material corresponds to 10 synthetic recordings in order to measure and compare the quality of the different algorithms and 66 real recordings coming from two different databases, one public database from Physionet and one database from the Clinical University Hospital, Valencia, Spain. RESULTS We have analyzed the performance of the three new algorithms and compared it with the performance of the traditional ICA algorithms. In the case of the synthetic data, it is possible to obtain the mean square error, so the comparison is easier. The new methods outperform the non-constrained versions in addition to simplifying the solution, since they do not need to recover all the components in order to estimate the atrial activity, i.e., the new methods are focused on the extraction of the atrial activity, so the extraction is stopped after the atrial signal is recovered. CONCLUSIONS We have shown that the ICA only version of the algorithms can be improved and adapted to fulfill the prior information about the characteristics of the atrial activity. This modification allows us to obtain new algorithms that have the following advantages compared to ICA only based solutions: they exploit prior information during the extraction, not in the postprocessing identification of the atrial signal; they extract only the interesting clinical signal instead of all the components; they outperform the ICA only version of the algorithm, improving the estimation of the atrial signal.

Nonnegative Matrix Factorization of Laboratory Astrophysical Ice Mixtures

We present an application of nonnegative matrix factorization (NMF) in astrophysics. It consists of the study of ice mixtures obtained in the laboratory. They simulate real astrophysical ices. The goal is to identify the molecules that are present in the ice mixtures. The data correspond to the infrared absorption spectra of ices formed by different combinations of molecules. The spectra and abundances are nonnegative, allowing the application of NMF. In addition, the statistics of the spectra correspond to supergaussian signals, so a sparseness restriction can be added. We review some NMF algorithms imposing sparseness in a Bayesian framework. We perform several simulations with artificial mixtures of ices in order to compare the algorithms and with real mixtures to show the usefulness of the approach. NMF is revealed as a powerful technique to analyze large databases in order to determine the compounds present in every ice.

Independent Component Analysis of Body Surface Potential Mapping Recordings with Atrial Fibrillation

Body surface potential mapping BSPM is an advanced electrocardiographic technique aimed at estimating the global potential distribution by recording potentials at multiple sites on the thoracic surface. In patients with atrial fibrillation the recorded potentials can be modelled as a linear mixture of the statistically independent atrial and ventricular activities. In order to extract atrial activity from surface recordings, we propose the independent component analysis of 64-lead BSPM recordings with atrial fibrillation. We apply also ICA to the 12-lead classical ECG and compare the results. This analysis shows that increasing the number of leads not only improves the quality of the extraction of the atrial activity but it allows an optimization of the localization of the leads. We introduce also a new index to measure the quality of the extraction. The BSPM recordings can also be used to obtain the best locations in order to select the minimum number of leads required to extract the AA, an important feature for long time recordings.

A fixed point algorithm for extracting the atrial activity in the frequency domain

In this work we present a fixed point algorithm to extract the atrial rhythm in atrial tachyarrhythmias from the surface electrocardiogram (ECG). In the frequency domain the atrial signal is characterized by a concentration of power around a main peak in the bandwidth 3-10Hz. The proposed algorithm exploits this discriminative property of the atrial component in combination with the decoupling of the atrial and other components superposed in the ECG. It recovers only the interesting atrial rhythm in a simple, fast and reliable way using the information contained in all the leads and reducing the average computational time from 0.902s (FastICA) to 0.023s (the proposed method). The algorithm is applied successfully to synthetic and real data. In simulated ECGs, the correlation index obtained was 0.792. In real ECGs, the accuracy of the results was validated using spectral and temporal parameters. The average peak frequency and spectral concentration obtained were 5.354Hz and 59.4%, respectively, and the kurtosis was 0.065.

Two applications of independent component analysis for non-destructive evaluation by ultrasounds

This paper presents two novel applications of ICA in Non Destructive Evaluation by ultrasounds applied to diagnosis of the material consolidation status and to determination of the thickness material profiles in restoration of historical buildings. In those applications the injected ultrasonic pulse is buried in backscattering grain noise plus sinusoidal phenomena; this latter is analyzed by ICA. The mixture matrix is used to extract useful information concerning to resonance phenomenon of multiple reflections of the ultrasonic pulse at non consolidated zones and to improve the signals by detecting interferences in ultrasonic signals. Results are shown by real experiments at a wall of a Basilica’s restored cupola. ICA is used as pre-processor to obtain enhanced power signal B-Scans of the wall.

Source separation of astrophysical ice mixtures

Infrared spectroscopy provides direct information on the composition of interstellar dust grains, which play a fundamental role in the evolution of interstellar medium ISM, from cold, quiet and low density molecular clouds to warm, active and dense protostellar ones. The determination of these components is fundamental to predict under the appropriate environmental conditions their evolution, including the appearance of new molecules, radicals and complex organics. The absorption spectrum of the astrophysical ice can be considered the additive linear absorption spectra of the multiple molecules present in the ice, so a linear instantaneous ICA mixture model is appropriate. We present the ICA statement of the problem, discussing the convenience of the model and its advantages in front of supervised methods. We obtain the MAP estimate of the mixing matrix, including its non-negative entries as a prior. We present the results carried out with an ice analogs database, confirming the suitability of the ICA approach.

Comparison of three artificial models of the magnetohydrodynamic effect on the electrocardiogram

The electrocardiogram (ECG) is often acquired during magnetic resonance imaging (MRI), but its analysis is restricted by the presence of a strong artefact, called magnetohydrodynamic (MHD) effect. MHD effect is induced by the flow of electrically charged particles in the blood perpendicular to the static magnetic field, which creates a potential of the order of magnitude of the ECG and temporally coincident with the repolarisation period. In this study, a new MHD model is proposed by using MRI-based 4D blood flow measurements made across the aortic arch. The model is extended to several cardiac cycles to allow the simulation of a realistic ECG acquisition during MRI examination and the quality assessment of MHD suppression techniques. A comparison of two existing models, based, respectively, on an analytical solution and on a numerical method-based solution of the fluids dynamics problem, is made with the proposed model and with an estimate of the MHD voltage observed during a real MRI scan. Results indicate a moderate agreement between the proposed model and the estimated MHD model for most leads, with an average correlation factor of 0.47. However, the results demonstrate that the proposed model provides a closer approximation to the observed MHD effects and a better depiction of the complexity of the MHD effect compared with the previously published models, with an improved correlation (), coefficient of determination () and fraction of energy () compared with the best previous model. The source code will be made freely available under an open source licence to facilitate collaboration and allow more rapid development of more accurate models of the MHD effect.

Ambiguous Keyboards and Scanning: The Relevance of the Cell Selection Phase

This paper focuses on the relevance of the cell selection phase in the overall performance of a text entry system based on scanning and with an ambiguous keyboard. Most of the layouts are designed trying only to minimize the ambiguity of the keyboard, and taking into consideration only the disambiguation process when entering text. Nevertheless, the number of scan cycles necessary for selecting the cells has great importance in the overall performance. As we show, the performance depends on the number of cells and the linguistic model used in the cell selection phase.

An informed source separation of astrophysical ice analogs

The analysis of astrophysical ices and the determination of the compounds that are present in the molecular clouds play a fundamental role in order to predict the future evolution of the cloud, e.g., its transformation to protostellar bodies or the appearance of new radicals and molecules. Because of the difficulties of obtaining satellite data, the process is simulated first in the laboratory generating ice analogs under well controlled variables. In this case, the ice mixture is carried on allocating the different components in the appropriate concentrations and recording the spectrum of the aggregated ice. This process tries to simulate the real process of forming ice mantles under the environmental conditions of the interstellar medium. The spectrum of each ice can be modeled as the linear instantaneous superposition of the spectrum of the different compounds, so a source separation approach is proper. In addition, some priors about the sources and the mixing matrix entries can be assumed, obtaining an informed Bayesian approach to the problem. We present the results obtained with the variational Bayesian approach for simulated and real mixtures, showing the good performance of the algorithm.