Search

Cardiac arrhythmias that can lead to sudden cardiac death are common. Electrocardiograms (ECGs) offer valuable information about cardiac status and play a crucial role in evaluating patients with arrhythmia in clinical diagnosis. The development of machine learning technologies has made the computer-aided diagnosis of cardiac arrhythmias possible, which can improve the efficiency and quality of medical treatment. In this study, a new method for the automatic classification of heartbeats is developed. This paper presents a novel approach to detect arrhythmias in the inter-patient paradigm using a convolutional neural network (CNN) that integrates multiscale convolutional blocks, frequency convolutional block attention (FCBA) modules, and RR interval features. The proposed method also includes oversampling for the heartbeats of the minority class and the addition of random noise to increase training samples and alleviate the problem of data imbalance. We used the MIT-BIH-AR arrhythmia database , which is a globally recognized ECG database, to evaluate the classification performance of the proposed model in this paper. Based on the experiments, the sensitivities of the normal beats (N), supraventricular ectopic beats (SVEB), and ventricular ectopic beats (VEB) were 96.9%, 89.3%, and 93.3%, respectively. In the inter-patient heartbeat classification paradigm, the overall accuracy of the proposed method is 95.60% in classifying heartbeats. The study shows that the proposed scheme outperforms other published schemes in terms of classification results. • A heartbeat classification method combining deep neural networks and medical features is proposed. • The use of multiscale convolutional fusion allows the network to extract richer electrocardiogram (ECG) features. • A frequency convolutional block attention (FCBA) module is proposed to adjust the wight relationships. • Medical features, namely RR intervals, are embedded into our deep model to further improve the performance of heartbeat classification. • Accuracy of 95.6% was achieved under inter-patient scheme. [ABSTRACT FROM AUTHOR]

Accurate and early detection of cardiac arrhythmia present in an electrocardiogram (ECG) can prevent many premature deaths. Cardiac arrhythmia arises due to the improper conduction of electrical impulses throughout the heart. In this paper, we propose an improved RR interval-based cardiac arrhythmia classification approach. The Discrete Wavelet Transform (DWT) and median filters were used to remove high-frequency noise and baseline wander from the raw ECG. Next, the processed ECG was segmented after the determination of the QRS region. We extracted the primary feature RR interval and other statistical features from the beats to classify the Normal, Premature Ventricular Contraction (PVC), and Premature Atrial Contraction (PAC). The K-Nearest Neighbour (k-NN), Support Vector Machine (SVM), Decision Tree (DT), Naïve Bayes (NB), and Random Forest (RF) classifier were utilised for classification. Overall performance of SVM with Gaussian kernel achieved Se % = 99.28, Sp % = 99.63, +P % = 99.28, and Acc % = 99.51, which is better than the other classifiers used in this method. The obtained results of the proposed method are significantly better and more accurate. [ABSTRACT FROM AUTHOR]

This quantitative study delineates the influence of respiratory rate on cardiovascular signal in healthy subjects using granger causality approach. Electrocardiogram ( RR ), arterial blood pressure ( SBP ) and respiration ( RESP ) were simultaneously recorded for 5 min from 20 subjects during normal (13–20 cycles/min) and deep breathing (5 cycles/min) with equal inspiration and expiration time. During deep breathing mean RR remains same but the variance increases. Also deep breathing lowers blood pressure, increases baroreflex sensitivity and improves oxygen saturation. The traditional frequency domain methods based power spectrum analysis and coherence analysis lacks to measure coupling changes among physiological subsystems. Therefore, frequency domain Granger causality method based on directed coherence is proposed to detect the changes in coupling strength among cardiovascular signals under different respiration rates. Directed coherence spectrum can separate RESP , SBP and RR components from each recorded signals. The RESP component of both RR ( RR RESP ) and SBP ( SBP RESP ) increases (coherence > 0.5) significantly during deep breathing indicate that respiration affects both cardiac and vascular system but at normal breathing only cardiac system ( RR RESP ) get affected by respiration. Also a significant increase in coherence is observed on baroreflex direction ( RR SBP ) indicating that deep breathing controls blood pressure. Hence, the observed directed coherence spectrum helps to detect coupling changes among cardiac, vascular and respiratory signal during autonomic regulation. [ABSTRACT FROM AUTHOR]

Background: The electrocardiogram (ECG) is one of the most non-invasive techniques to give support to the atrial fibrillation (AF) diagnosis. Several authors use the temporal difference between two consecutive R waves, a method known as RR interval, to perform the AF diagnosis. However, RR interval-based analysis does not detect distortions on the other ECG waves.Purpose: Thus, the present work proposes a diagnostic decision support systems for AF based on higher order spectrum analysis of the voltage variation on the ECG..Methods: The proposed method was used aiming AF classifying. The classifier is composed by two screening stages: one based on the average and another on the average deviation of kurtosis of the ECG signals. Heartbeat obtained from the MIT-BIH atrial fibrillation and MIT-BIH normal were used.Results: ECG signal featured by kurtosis outperforms second order statistics based metrics in up to 476 times, and up to 110 times above the RR interval. The screening methods obtained sensitivity equal to 100% and specificity is up to 84.04%. The two screening methods combined provided an AF classifier with an accuracy rate at diagnosis of 100%. The results presented take into account windows of up to five heartbeats and a 99.73% confidence interval.Conclusion: The results obtained by the proposed method can be used to support decision-making in clinical practices with a diagnostic accuracy rate of 90.04% to 100%. [ABSTRACT FROM AUTHOR]

Objectives To assess the influence of 2nd-degree AV blocks (AVB) on RR interval-based heart rate variability (HRV) variables; to investigate the effect of using PP interval time series and of artifact filtering on HRV analyses; to investigate the influence of electrocardiogram (ECG) recording length and time of recording; and to calculate day-to-day variability and reference intervals of HRV variables. Animals Thirty healthy adult horses. Methods RR and PP interval time series were extracted from 10-h Holter ECGs and an automated filter was applied to the RR time series (RR f ). Time-domain HRV variables were calculated based on RR, PP, and RR f time series and their relation to the number of AVBs was assessed. Hourly 10- and 60-min segments were extracted to investigate the influence of segment length and recording time on HRV variables. Day-to-day variability and reference intervals of HRV variables were calculated. Results Variables of short-term HRV were significantly influenced by the number of AVBs when based on RR, but not when based on PP- and RR f time series. PP- and RR f -based HRV variables were in good agreement. The majority of HRV variables were influenced by recording time and ECG segment length. Day-to-day variability of HRV variables was low when based on 10-h ECG recordings but moderate to high when based on 60-min and 10-min recordings. Conclusions Second-degree AVBs significantly influence conventional RR-based, but not PP- and RR f -based time-domain HRV variables. However, PP and RR f analyses have limitations and recording length and time of recording must be considered. [ABSTRACT FROM AUTHOR]

• In this article, the CPSC 2018 dataset was used to train a ResRNN model. To classify the diseases into nine categories: normal, Arrhythmia (AF), I-atrioventricular block (AVB), left bundle branch block (LBBB), right bundle branch block (RBBB), premature atrial contraction (PAC), premature ventricular contractions (PVC), ST-segment depression (STD), ST-segment elevation (STSE) /(STE). • The ResRNN outperforms in identifying eight types of anomalies in 12-lead ECG data. • When level 4 rough coefficients with the Symlet-8 (Sym8) channel were used for organization, the best degree of exactness was obtained. • The ResRNN classifier has a normal accuracy of 91 percent when using ECG signals, which is much higher than the deep multiscale fusion neural network (DesNet) at 83 % and inception ResNet V2 at 80 %. Automated electrocardiogram (ECG) analysis cannot be employed in clinical practice due to the accuracy of the present models. Deep Neural Networks (DNNs) are models made up of stacking transformations that learn tasks by example. This technology has lately demonstrated remarkable performance in several activities, and its potential to improve clinical practice is highly anticipated. In this article, the China physiological signal challenge (CPSC)- 2018 dataset was used to train a ResRNN model. The ResRNN outperforms in identifying eight types of anomalies in 12-lead ECG data. When level 4 rough coefficients with the Symlet-8 (Sym8) channel were used for organization, the best degree of exactness was obtained. The ResRNN classifier has a normal accuracy of 91% when using ECG signals, which is much higher than the deep multiscale fusion neural network (DesNet) at 83 % and Inception ResNet V 2 at 80 %. As a result, the given methodology is far better than others. After preprocessing the signals, the ResRNN model classifies the diseases into nine categories: normal, Atrial fibrillation (AF), first degree-atrioventricular block (AVB), left bundle branch block (LBBB), right bundle branch block (RBBB), premature atrial contraction (PAC), premature ventricular contractions (PVC), ST-segment depression (STD), and ST-segment elevation (STSE) /(STE). [ABSTRACT FROM AUTHOR]

• In this article, the CPSC 2018 dataset was used to train a ResRNN model. To classify the diseases into nine categories: normal, Arrhythmia (AF), I-atrioventricular block (AVB), left bundle branch block (LBBB), right bundle branch block (RBBB), premature atrial contraction (PAC), premature ventricular contractions (PVC), ST-segment depression (STD), ST-segment elevation (STSE) /(STE). • The ResRNN outperforms in identifying eight types of anomalies in 12-lead ECG data. • When level 4 rough coefficients with the Symlet-8 (Sym8) channel were used for organization, the best degree of exactness was obtained. • The ResRNN classifier has a normal accuracy of 91 percent when using ECG signals, which is much higher than the deep multiscale fusion neural network (DesNet) at 83 % and inception ResNet V2 at 80 %. Automated electrocardiogram (ECG) analysis cannot be employed in clinical practice due to the accuracy of the present models. Deep Neural Networks (DNNs) are models made up of stacking transformations that learn tasks by example. This technology has lately demonstrated remarkable performance in several activities, and its potential to improve clinical practice is highly anticipated. In this article, the China physiological signal challenge (CPSC)- 2018 dataset was used to train a ResRNN model. The ResRNN outperforms in identifying eight types of anomalies in 12-lead ECG data. When level 4 rough coefficients with the Symlet-8 (Sym8) channel were used for organization, the best degree of exactness was obtained. The ResRNN classifier has a normal accuracy of 91% when using ECG signals, which is much higher than the deep multiscale fusion neural network (DesNet) at 83 % and Inception ResNet V 2 at 80 %. As a result, the given methodology is far better than others. After preprocessing the signals, the ResRNN model classifies the diseases into nine categories: normal, Atrial fibrillation (AF), first degree-atrioventricular block (AVB), left bundle branch block (LBBB), right bundle branch block (RBBB), premature atrial contraction (PAC), premature ventricular contractions (PVC), ST-segment depression (STD), and ST-segment elevation (STSE) /(STE). [ABSTRACT FROM AUTHOR]

Background Knowledge of the onset of atrial fibrillation (AF) episodes in patients with paroxysmal atrial fibrillation (PAF) will enable them to better manage this condition. Current advances in mobile technology allow RR interval data to be obtained in real time. An analysis technique using RR interval data is presented with a view to alert a subject before a PAF episode. Method The method is based on a time series of standard deviation and 0.99 quantile values of the spectral entropy, constructed from RR data. The RR data are taken from three time periods. The first time period has no occurrences of AF for 45 min to either side of the time period. The second time period just precedes an AF attack. Both of these are of thirty minutes duration. The third time period of approximately 5 min follows the second, and is when AF occurs. Results Twenty-two PAF subjects were studied and in all cases there was a steady increase in the values of these indices as the onset of the AF attack approached. Conclusion This method of analysis of RR interval data shows potential use to alert a PAF subject before the onset of an AF episode. [ABSTRACT FROM AUTHOR]

Synchronization provides an insight into underlying the interaction mechanisms among the bivariate time series and has recently become an increasing focus of interest. In this study, we proposed a new cross entropy measure, named cross fuzzy measure entropy (C-FuzzyMEn), to detect the synchronization of the bivariate time series. The performances of C-FuzzyMEn, as well as two existing cross entropy measures, i.e., cross sample entropy (C-SampEn) and cross fuzzy entropy (C-FuzzyEn), were first tested and compared using three coupled simulation models (i.e., coupled Gaussian noise, coupled MIX( p ) and coupled Henon model) by changing the time series length, the threshold value for entropy and the coupling degree. The results from the simulation models showed that compared with C-SampEn, C-FuzzyEn and C-FuzzyMEn had better statistical stability and compared with C-FuzzyEn, C-FuzzyMEn had better discrimination ability. These three measures were then applied to a cardiovascular coupling problem, synchronization analysis for RR and pulse transit time (PTT) series in both the normal subjects and heart failure patients. The results showed that the heart failure group had lower cross entropy values than the normal group for all three cross entropy measures, indicating that the synchronization between RR and PTT time series increases in the heart failure group. Further analysis showed that there was no significant difference between the normal and heart failure groups for C-SampEn (normal 2.13 ± 0.37 vs. heart failure 2.07 ± 0.16, P = 0.36). However, C-FuzzyEn had significant difference between two groups (normal 1.42 ± 0.25 vs. heart failure 1.31 ± 0.12, P < 0.05). The statistical difference was larger for two groups when performing C-FuzzyMEn analysis (normal 2.40 ± 0.26 vs. heart failure 2.15 ± 0.13, P < 0.01). [ABSTRACT FROM AUTHOR]

Abstract: The study investigated whether the beat-to-beat QT interval variability relationship to the mean heart rate and the RR interval variability depended on the cardiovascular autonomic status changed by postural positioning. Repeated long-term 12-lead Holter recordings were obtained from 352 healthy subjects (mean age 32.7 ± 9.1 years, 176 females) while they underwent postural provocative tests involving supine, unsupported sitting and unsupported standing positions. Each recording was processed as a sequence of overlapping 10-second segments. In each segment, the mean RR interval, the coefficients of variance of the RR intervals (RRCV) and the QT intervals (QTCV) were obtained. In each subject, these characteristics, corresponding to different postural positions, were firstly averaged and secondly used to obtain within-subject correlation coefficients between the different characteristics at different postural positions. While the within-subject means of RRCV generally decreased when changing the position from supine to sitting and to standing (4.53 ± 1.95%, 4.12 ± 1.51% and 3.26 ± 1.56% in females and 3.99 ± 1.44%, 4.00 ± 1.24% and 3.53 ± 1.32% in males respectively), the means of QTCV systematically increased during these position changes (0.96 ± 0.40%, 1.30 ± 0.56% and 1.88 ± 1.46% in females and 0.85 ± 0.30%, 1.13 ± 0.41% and 1.41 ± 0.59% in males, respectively). The intra-subject relationship between QTCV, RRCV and mean RR intervals was highly dependent on postural positions. The study concludes that no universally applicable normalization of the QT interval variability for the heart rate and/or the RR interval variability should be assumed. In future studies of the QT variability, it seems preferable to report on the absolute values of QT variability, RR variability and mean heart rate separately. [Copyright &y& Elsevier]

Background: Although J-wave elevation in the inferolateral leads could be related to idiopathic ventricular fibrillation (IVF), little is known about the pathophysiologic characteristics of J-wave elevation in patients with IVF. Objective: This study aimed to determine the relationship between augmentation of J-wave elevation and changes in RR interval or autonomic nervous activities in patients with IVF. Methods: Eight patients with IVF and 22 controls with J-wave elevation (≥0.1 mV) in lead V5 were studied. The J-wave amplitude was automatically measured in lead CM5 of a digital Holter electrocardiogram, and the J–RR relationship was determined. Based on the analysis of heart rate variability, the relationship between the J-wave amplitude and the natural logarithm of high-frequency (HF) components (J–ln HF relationship) or the ratio of low frequency (LF) components to HF components (J–LF/HF relationship) was also determined. Results: The J–RR slope (mm/s) was greater in patients with IVF than in controls (3.5 ± 0.7 vs 2.4 ± 0.8; P <.01), as was J-wave amplitude (mm) at an RR interval of 1.2 seconds (2.8 ± 0.9 vs 2.0 ± 0.6; P <.05). The J-wave amplitude was correlated positively with ln HF and negatively with LF/HF, and the slopes of both J–ln HF and J–LF/HF regression lines were greater in patients with IVF than in controls. During an entire 24-hour period, there was no difference between the 2 groups in either HF or LF/HF. Nine of the total 11 episodes (82%) of spontaneous ventricular fibrillation occurred between 18:00 and 6:00. Conclusions: In patients with IVF as compared with control subjects, J-wave elevation was more strongly augmented during bradycardia and was associated with an increase in vagal activity. This could be related to the occurrence of ventricular fibrillation predominantly at night in patients with IVF. [ABSTRACT FROM AUTHOR]

Abstract: Although heart rate dependency of QT interval is well known, the relationship of other electrocardiogram (ECG) parameters to heart rate has been researched less intensively. This study investigated the heart rate dependencies of QT interval, PQ interval, and QRS width in 40 healthy subjects (18 women; mean age, 30.4 ± 8.1 years). In each subject, 3 long-term (approximately 13 hours) 12-lead ECGs were obtained for 3 day-time periods with gaps of 2 to 3 weeks between repeated recordings. In each recording, approximately 230 ECG measurements of QT interval, PQ interval, and QRS width were made, each preceded by stable heart rate. For each recording, linear regression slopes of QT/RR, PQ/RR, and QRS/RR relationships were obtained. Intrasubject SDs of individual values were compared with intersubject SDs of intrasubject means to test the individuality of the relationships. The intrasubject means of the slope values were also compared between sexes. The individual SDs of the QT/RR regression slopes were 0.0116 ± 0.0065, whereas the population SD of intrasubject means was 0.0245 (P = 3.6 × 10−15). For the PQ/RR slopes and QRS/RR slopes, these values were 0.0085 ± 0.0050 vs 0.0314 (P = 7.9 × 10−28) and 0.001 89 ± 0.001 57 vs 0.005 50 (P = 2.6 × 10−17), respectively. The QT/RR slopes were steeper in women than in men (0.194 ± 0.019 vs 0.168 ± 0.022, 0.0005), whereas the QRS/RR slopes were, on average, negative in women while positive in men (−0.001 38 ± 0.0045 vs 0.003 35 ± 0.0054, P = .005). There were no sex differences in the PQ/RR slopes (0.054 ± 0.032 in women vs 0.055 ± 0.031 in men, P = .95). Thus, not only the heart rate dependency of QT interval but also the rate dependencies of PQ interval and of QRS width show high intrasubject stability with substantial intersubject differences. [Copyright &y& Elsevier]

Background: Circadian QTc changes have been reported, with conflicting results. Spontaneous QTc variability is important for pharmaceutical cardiac safety studies. Objective: The purpose of this study was to investigate QTc variability in accurately measured and heart-rate corrected daytime data of healthy subjects. Methods: Continuous 12-lead ECGs were recorded in 53 healthy volunteers. For each recording, ∼320 ECG samples of 10 seconds were obtained throughout the daytime period, all preceded by stable heart rates. In each ECG sample, QT interval was measured on superimposed 12 leads by two independent cardiologists and reconciled. Four RR-interval expressions were used: (a) average of the first three RR of the ECG sample, (b) RR average of the 10-second sample, (c) average of RR intervals in a 2-minute history, and (d) RR intervals of an independently established individual QT/RR hysteresis profile. For all RR-interval expressions, QT intervals were corrected using the Fridericia formula and individually optimized curvature correction. QTc variability was measured by intraindividual QTc standard deviations. Results: With Fridericia correction and the RR expressions (a) to (d), QTc variability obtained was (a) 9.45 ± 1.70 ms, (b) 7.80 ± 1.48 ms, (c) 6.37 ± 1.64 ms, and (d) 5.81 ± 1.75 ms. With individualized correction, QTc variability was (a) 8.16 ± 1.71 ms, (b) 6.71 ± 1.41 ms, (c) 5.22 ± 1.13 ms, and (d) 4.56 ± 1.18 ms. All differences (b) vs (a), (c) vs (b), and (d) vs (c) were highly statistically significant (P <10−12 in all cases). Conclusion: Previously reported large QTc variability largely results from methodologic imprecision. Little QTc variability is present in daytime recordings of healthy subjects. Consequently, QT-related pharmaceutical cardiac safety studies can be made smaller without decreasing their power. [Copyright &y& Elsevier]

Index of vulnerability is a parameter based on ventricular gradient evaluating the risk of arrhythmia development. The index is derived from isointegral maps of the QT interval. Individual characteristics of isointegral maps are influenced by different factors, which contribute to the relatively high variability among measured parameters of maps in measured subjects. While several electrocardiographic indexes have been introduced, there are only few studies of their dependence on heart rate. In this study we set out to establish the dependence of vulnerability index on the RR interval or heart rate in healthy population. A positive linear correlation between RR intervals and mean and minimum values of vulnerability indexes was found. [ABSTRACT FROM AUTHOR]

• A low computational load and dynamic thresholding based QRS complex detection method. • The dynamic threshold value is computed for both upward and downward QRS complex simultaneously. • The proposed method was applied on five different full-length databases, MITBIH AD, FTD, ESTD, NSDT, and FTDB, has given AETR of 2.35 s, 3.8 s, 3.69 s, 1.99 s, 0.53 s, respectively. • The proposed method is tested on total 195 records and 672,472 beats of five different ECG databases, which yields the 99.70 and 99.69, average Se%, and +P% respectively. QRS-complex detection is a primitive step in the detection of cardiac disorder using electrocardiogram (ECG). Abnormal and varying peaks, baseline wander and other noise are the main challenges in accurate QRS-complex detection. In this paper, A window-based FIR filter is used to eliminate the high-frequency noise. Next, the ECG is enhanced to the power third after multiplication followed by normalization and moving average process to retain dynamic QRS-complex. Then baseline and root mean square (RMS) value of first three seconds of the signal are used for initial thresholding, later dynamic thresholding process was utilized to update the threshold value after the detection of four R-peaks. The threshold value was automatically updated using the previous threshold value, R-peak amplitude, RR-interval, and RR-intervals means. Kurtosis coefficient computation is used for discarding prominent T-wave and further this technique located the QRS-complex accurately in the raw ECG signal. The proposed method was applied to Massachusetts Institute of Technology-Beth Israel Hospital Arrhythmia Database (MIT-BIH AD), Fantasia Database (FTD), European ST-T Database (ESTD), MIT-BIH Noise Stress Test Database (NSTD), and Direct Fetal ECG Database (FTD) for its evaluation and validation. The overall sensitivity rate of 99.70% and positive predictivity rate of 99.69% have been achieved. The proposed method is based on dynamic thresholding and simple decision rules, which makes this method computationally efficient. Wide validation over five different databases proves the robustness of this method. [ABSTRACT FROM AUTHOR]

• A simple but efficient heuristic method for atrial fibrillation detection. • A novel AF feature extraction strategy based on Poincaré plot of first-order differential RR interval. • The advantage of distinguishing AF from non-AF with complicated arrhythmias. With the popularity of wearable mobile devices, increasing numbers of applications use single-lead electrocardiograms (ECGs), which provides information about heart rate and rhythm to detect cardiac diseases. Atrial fibrillation (AF) is the most common arrhythmia in clinical practice, it is very important to detect AF early. In this study, we propose a Poincaré plot-based method for AF detection that has the advantages of requiring few calculations and short detection segments. Our method first extracts the first-order differential RR interval (ΔRRI) sequences from a segmented ECG data, and then the polar coordinate transformation is performed on the Poincaré plot of the ΔRRI to obtain the phase-based distribution. Two features, the distribution width D w and the average distribution height D h are extracted from the phase-based distribution to classify the AF and non-AF episodes. Five PhysioNet databases were used for the assessment, which contain 3,843.3 hours of data from 229 subjects. Long-Term AF Database was selected as the training set, and the remaining four databases were used as the testing sets. In the testing sets, the results yield good sensitivity/specificity (97.91%/99.14%) with 60 second ECG signal segments, which are better than those of the existing the Poincaré plot-based methods, and the method also performs well with 20 second ECG signal segments (sensitivity/specificity of 97.28%/98.35%). The detection performance of this method is good under different condition, and has the advantage of easy implementation, which makes it promising for application in the detection and monitoring of AF. [ABSTRACT FROM AUTHOR]