Rute Almeida

A wavelet-based ECG delineator: evaluation on standard databases

In this paper, we developed and evaluated a robust single-lead electrocardiogram (ECG) delineation system based on the wavelet transform (WT). In a first step, QRS complexes are detected. Then, each QRS is delineated by detecting and identifying the peaks of the individual waves, as well as the complex onset and end. Finally, the determination of P and T wave peaks, onsets and ends is performed. We evaluated the algorithm on several manually annotated databases, such as MIT-BIH Arrhythmia, QT, European ST-T and CSE databases, developed for validation purposes. The QRS detector obtained a sensitivity of Se=99.66% and a positive predictivity of P+=99.56% over the first lead of the validation databases (more than 980,000 beats), while for the well-known MIT-BIH Arrhythmia Database, Se and P+ over 99.8% were attained. As for the delineation of the ECG waves, the mean and standard deviation of the differences between the automatic and manual annotations were computed. The mean error obtained with the WT approach was found not to exceed one sampling interval, while the standard deviations were around the accepted tolerances between expert physicians, outperforming the results of other well known algorithms, especially in determining the end of T wave.

QT variability and HRV interactions in ECG: quantification and reliability

In this paper, a dynamic linear approach was used over QT and RR series measured by an automatic delineator, to explore the interactions between QT interval variability (QTV) and heart rate variability (HRV). A low-order linear autoregressive model allowed to separate and quantify the QTV fractions correlated and not correlated with HRV, estimating their power spectral density measures. Simulated series and artificial ECG signals were used to assess the performance of the methods, considering a respiratory-like electrical axis rotation effect and noise contamination with a signal-to-noise ratio (SNR) from 30 to 10 dB. The errors found in the estimation of the QTV fraction related to HRV showed a nonrelevant performance decrease from automatic delineation. The joint performance of delineation plus variability analysis achieved less than 20% error in over 75% of cases for records presenting SNRs higher than 15 dB and QT standard deviation higher than 10 ms. The methods were also applied to real ECG records from healthy subjects where it was found a relevant QTV fraction not correlated with HRV (over 40% in 19 out of 23 segments analyzed), indicating that an important part of QTV is not linearly driven by HRV and may contain complementary information.

Multilead ECG Delineation Using Spatially Projected Leads From Wavelet Transform Loops

In this paper, a novel multilead (ML) based automatic strategy for delineation of ECG boundaries is proposed and evaluated with respect to the QRS and T-wave boundaries. The ML strategy is designed from a single-lead (SL) wavelet-transform-based delineation system. It departs from three orthogonal leads and takes advantage of the spatial information provided using a derived lead better fitted for delineation. SL delineation is then applied over this optimal derived lead. The ML strategy produces a reduced error dispersion compared to SL results, thus providing more robust, accurate, and stable boundary locations than any electrocardiographic lead by itself and outperforming strategies based on lead selection rules after SL delineation.

Respiration Effect on Wavelet-Based ECG T-Wave End Delineation Strategies

The main purpose of this paper is to study the influence of the mechanical effect of respiration over the T-wave end delineation. We compared the performance of an automatic delineation system based on the wavelet transform (WT), considering single lead (SL), global delineation locations obtained from SL annotations (SLR), and multilead (ML) approaches. The linear relation between the variations on T-wave end locations obtained with each of the methods and the mechanical effect of respiration was quantified using spectral coherence and ARARX modeling both in simulated signals and in real data. We also explored the evolution of the vectorcardiographic spatial loop using the projection on the main direction of the WT in the region close to the T-wave end ( ) and its relation with respiration. The dispersion of the additional T-wave end location error due to respiration was reduced by 15% using SLR with respect to SL, while ML allows for a reduction of around 40%. The percentage of that error correlated with respiration was in average 99% using SL while 82% and 72% using SLR and ML, respectively. Thus, results suggest that ML is the most adequate strategy for T-wave delineation, allowing the reduction of the instability of T-wave end location caused by respiration.

Improved QT variability quantification by multilead automatic delineation

In this work we evaluate the joint robustness of a multilead delineation and a parametric approach to study the relations between HRV and QTV. The performance of the automatic system is studied over simulated 3-lead ECG signals in order to quantify the improvement allowed by the multilead delineation. Respiratory effect and contamination with realistic noise extracted from a real ECG, rescaled to obtain SNR levels from 30 to 5 dB, were also considered. Compared with same parametric methods over RR and QT series measured from a single lead based approach, the multilead delineator allows to reduce the error in QTV quantification, in particular the error bias in signals at SNR=20 dB. It improves the joint performance facing realistic 3 lead noise at SNRges20 dB, remarkably around 20 dB, making it usable for ECG signals with QTV levels corresponding to a QT standard deviationges13 ms

Blind source separation using time-delayed signals

In this work a modified version of AMUSE, called dAMUSE, is proposed. The main modification consists in increasing the dimension of the data vectors by joining delayed versions of the observed mixed signals. With the new data a matrix pencil is computed and its generalized eigendecomposition is performed as in AMUSE. We will show that in this case the output (or independent) signals are filtered versions of the source signals. Some numerical simulations using artificially mixed signals as well as biological data (RR and QT intervals of Electrocardiogram) are presented.

Validation of heart rate monitor Polar RS800 for heart rate variability analysis during exercise.

Abstract Hernando, D, Garatachea, N, Almeida, R, Casajús, JA, and Bailón, R. Validation of heart rate monitor Polar RS800 for heart rate variability analysis during exercise. J Strength Cond Res 32(3): 716–725, 2018—Heart rate variability (HRV) analysis during exercise is an interesting noninvasive tool to measure the cardiovascular response to the stress of exercise. Wearable heart rate monitors are a comfortable option to measure interbeat (RR) intervals while doing physical activities. It is necessary to evaluate the agreement between HRV parameters derived from the RR series recorded by wearable devices and those derived from an electrocardiogram (ECG) during dynamic exercise of low to high intensity. Twenty-three male volunteers performed an exercise stress test on a cycle ergometer. Subjects wore a Polar RS800 device, whereas ECG was also recorded simultaneously to extract the reference RR intervals. A time–frequency spectral analysis was performed to extract the instantaneous mean heart rate (HRM), and the power of low-frequency (PLF) and high-frequency (PHF) components, the latter centered on the respiratory frequency. Analysis was done in intervals of different exercise intensity based on oxygen consumption. Linear correlation, reliability, and agreement were computed in each interval. The agreement between the RR series obtained from the Polar device and from the ECG is high throughout the whole test although the shorter the RR is, the more differences there are. Both methods are interchangeable when analyzing HRV at rest. At high exercise intensity, HRM and PLF still presented a high correlation (&rgr; > 0.8) and excellent reliability and agreement indices (above 0.9). However, the PHF measurements from the Polar showed reliability and agreement coefficients around 0.5 or lower when the level of the exercise increases (for levels of O2 above 60%).

dAMUSE---A new tool for denoising and blind source separation

In this work a generalized version of AMUSE, called dAMUSE is proposed. The main modification consists in embedding the observed mixed signals in a high-dimensional feature space of delayed coordinates. With the embedded signals a matrix pencil is formed and its generalized eigendecomposition is computed similar to the algorithm AMUSE. We show that in this case the uncorrelated output signals are filtered versions of the unknown source signals. Further, denoising the data can be achieved conveniently in parallel with the signal separation. Numerical simulations using artificially mixed signals are presented to show the performance of the method. Further results of a heart rate variability (HRV) study are discussed showing that the output signals are related with LF (low frequency) and HF (high frequency) fluctuations. Finally, an application to separate artifacts from 2D NOESY NMR spectra and to denoise the reconstructed artefact-free spectra is presented also.

Microgravity effects on ventricular response to heart rate changes

The effect of simulated microgravity on ventricular repolarization (VR) has been evaluated on healthy volunteers by a 5-day Head Down (-6°) Bed Rest (HDBR) maneuver. QT to RR and QTp (measured until the peak of the T wave) to RR hystereses have been measured during a tilt table test, and differences between them have been studied to better understand possible changes in the final part of the repolarization. To characterize the hystereses, two indices have been computed: M90, quantifying adaptation lag in beats, and α evaluating the slope of parabolic regression fitting. Significant differences between QT and QTp were found before, but not after HDBR. Specifically, before HDBR was considerable lower for QTp than for QT, while α was significantly higher. After HDBR, M90 and a took essentially the same values for QT and QTp. This fact evidenced the different effect of HDBR on QT to RR and QTp to RR adaptations, and suggest HDBR could lead to an impairment in ventricular repolarization dispersion.

A parametric model approach for quantification of short term QT variability uncorrelated with heart rate variability

In this work we propose to assess the relation between HRV and QTV measured by an automatic delineator. A low order linear autoregressive model on RR versus QT interactions was used to explore short term relations and quantify the fractions of QTV correlated and not correlated with HRV. Power spectral density measures were estimated from the total QTV and from the two separated fractions using the proposed model. Simulated RR and QT series were used to quantify the error bounds associated to the method performance. ECG records of young normal subjects were processed to obtain the RR and QT series. The high QTV fraction not correlated with RR found in these records (over 40% in 98% of the segments) indicates that an important part of QTV can be driven by other factors rather than by heart rate, and may contain complementary information.