Early prediction of sudden cardiac death risk with Nested LSTM based on electrocardiogram sequential features.

Electrocardiogram (ECG) signals are very important for heart disease diagnosis. In this paper, a novel early prediction method based on Nested Long Short-Term Memory (Nested LSTM) is developed for sudden cardiac death risk detection. First, wavelet denoising and normalization techniques are utilized for reliable reconstruction of ECG signals from extreme noise conditions. Then, a nested LSTM structure is adopted, which can guide the memory forgetting and memory selection of ECG signals, so as to improve the data processing ability and prediction accuracy of ECG signals. To demonstrate the effectiveness of the proposed method, four different models with different signal prediction techniques are used for comparison. The extensive experimental results show that this method can realize an accurate prediction of the cardiac beat's starting point and track the trend of ECG signals effectively. This study holds significant value for timely intervention for patients at risk of sudden cardiac death.

Cardiopulmonary Coupling Spectrogram as an Ambulatory Method for Assessing Sleep Disorders in Patients With Autoimmune Encephalitis.

BACKGROUND AND OBJECTIVES: Sleep disorders are a common and important clinical feature in patients with autoimmune encephalitis (AE); however, they are poorly understood. We aimed to evaluate whether cardiopulmonary coupling (CPC), an electrocardiogram-based portable sleep monitoring technology, can be used to assess sleep disorders in patients with AE. METHODS: Patients fulfilling the diagnostic criteria of AE were age- and sex-matched with recruited healthy control subjects. All patients and subjects received CPC testing between August 2020 and December 2022. Demographic data, clinical information, and Pittsburgh Sleep Quality Index (PSQI) scores were collected from the medical records. Data analysis was performed using R language programming software. RESULTS: There were 60 patients with AE (age 26.0 [19.8-37.5] years, male 55%) and 66 healthy control subjects (age 30.0 [25.8-32.0] years, male 53%) included in this study. Compared with healthy subjects, patients with AE had higher PSQI scores (7.00 [6.00-8.00] vs 3.00 [2.00-4.00], p < 0.001), lower sleep efficiency (SE 80% [71%-87%] vs 92% [84%-95%], p < 0.001), lower percentage of high-frequency coupling (25% [14%-43%] vs 45% [38%-53%], p < 0.001), higher percentage of REM sleep (19% ± 9% vs 15% ± 7%, p < 0.001), higher percentage of wakefulness (W% 16% [11%-25%] vs 8% [5%-16%], p = 0.074), higher low-frequency to high-frequency ratio (LF/HF 1.29 [0.82-2.40] vs 0.91 [0.67-1.29], p = 0.001), and a higher CPC-derived respiratory disturbance index (9.78 [0.50-22.2] vs 2.95 [0.40-6.53], p < 0.001). Follow-up evaluation of 14 patients showed a decrease in the PSQI score (8.00 [6.00-9.00] vs 6.00 [5.00-7.00], p = 0.008), an increased SE (79% [69%-86%] vs 89% [76%-91%], p = 0.030), and a decreased W% (20% [11%-30%] vs 11% [8%-24], p = 0.035). Multiple linear regression indicated that SE (-7.49 [-9.77 to -5.21], p < 0.001) and LF/HF ratio (0.37 [0.13-0.6], p = 0.004) were independent factors affecting PSQI scores in patients with AE. DISCUSSION: Sleep disorders with autonomic dysfunction are common in patients with AE. Improvements in the PSQI score and SE precede the restoration of sleep microstructural disruption in the remission stage. CPC parameters may be useful in predicting sleep disorders in patients with AE.

Trait impulsivity influences behavioural and physiological responses to threat in a virtual environment.

Trait impulsivity represents a tendency to take action without forethought or consideration of consequences. This trait is multifaceted and can be decomposed into attentional, motor and non-planning subtypes of impulsivity. The purpose of the current study was to investigate how subtypes of trait impulsivity responded to different degrees of threat within room-scale virtual reality (VR) with respect to behaviour and level of physiological activation. Thirty-four participants were required to negotiate a virtual environment (VE) where they walked at height with the continuous threat of a virtual 'fall.' Behavioural measures related to the speed of movement, interaction frequency and risk were collected. Participants also wore ambulatory sensors to collect data from electrocardiogram (ECG) and electrodermal activity (EDA). Our results indicated that participants who scored highly on non-planning impulsivity exhibited riskier behaviour and higher skin conductance level (SCL). Participants with higher motor impulsivity interacted with more objects in the VE when threat was high, they also exhibited contradictory indicators of physiological activation. Attentional impulsivity was associated with a greater number of falls across the VE. The results demonstrate that subtypes of trait impulsivity respond to threats via different patterns of behaviour and levels of physiological activation, reinforcing the multifaceted nature of the trait.

An interesting electrocardiogram caused by lead reversal.

BACKGROUND: During normal sinus rhythm, atrial depolarization is conducted from right atrium to left atrium through Bachmann's bundle, and a normal P wave axis which is measured on the frontal plane is between 0º and + 75º. The change of P wave polarity is helpful for the analysis of origin point. CASE PRESENTATION: We report a patient with negative P wave in lead I. The characteristics of QRS complex in leads V1 to V6 are helpful to preliminarily differential diagnosis. The 12-lead electrocardiogram (ECG) with correct limb leads (right arm-left arm) placement shows sinus rhythm with complete right bundle branch block (RBBB). CONCLUSIONS: The change of P wave polarity as well as characteristics of QRS complex can help identify limb-lead reversals.

Multi-branch myocardial infarction detection and localization framework based on multi-instance learning and domain knowledge.

Objective. Myocardial infarction (MI) is a serious cardiovascular disease that can cause irreversible damage to the heart, making early identification and treatment crucial. However, automatic MI detection and localization from an electrocardiogram (ECG) remain challenging. In this study, we propose two models, MFB-SENET and MFB-DMIL, for MI detection and localization, respectively.Approach. The MFB-SENET model is designed to detect MI, while the MFB-DMIL model is designed to localize MI. The MI localization model employs a specialized attention mechanism to integrate multi-instance learning with domain knowledge. This approach incorporates handcrafted features and introduces a new loss function called lead-loss, to improve MI localization. Grad-CAM is employed to visualize the decision-making process.Main Results.The proposed method was evaluated on the PTB and PTB-XL databases. Under the inter-patient scheme, the accuracy of MI detection and localization on the PTB database reached 93.88% and 67.17%, respectively. The accuracy of MI detection and localization on the PTB-XL database were 94.89% and 85.83%, respectively.Significance. Our method achieved comparable or better performance than other state-of-the-art algorithms. The proposed method combined deep learning and medical domain knowledge, demonstrates effectiveness and reliability, holding promise as an efficient MI diagnostic tool to assist physicians in formulating accurate diagnoses.

Patient-independent, MHD-robust R-peak detection for retrospective gating in cardiac MRI imaging.

Objective.In cardiovascular magnetic resonance imaging, synchronization of image acquisition with heart motion (calledgating) is performed by detecting R-peaks in electrocardiogram (ECG) signals. Effective gating is challenging with 3T and 7T scanners, due to severe distortion of ECG signals caused by magnetohydrodynamic effects associated with intense magnetic fields. This work proposes an efficient retrospective gating strategy that requires no prior training outside the scanner and investigates the optimal number of leads in the ECG acquisition set.Approach.The proposed method was developed on a data set of 12-lead ECG signals acquired within 3T and 7T scanners. Independent component analysis is employed to effectively separate components related with cardiac activity from those associated to noise. Subsequently, an automatic selection process identifies the components best suited for accurate R-peak detection, based on heart rate estimation metrics and frequency content quality indexes.Main results.The proposed method is robust to different B0 field strengths, as evidenced by R-peak detection errors of 2.4 ± 3.1 ms and 10.6 ± 15.4 ms for data acquired with 3T and 7T scanners, respectively. Its effectiveness was verified with various subject orientations, showcasing applicability in diverse clinical scenarios. The work reveals that ECG leads can be limited in number to three, or at most five for 7T field strengths, without significant degradation in R-peak detection accuracy.Significance.The approach requires no preliminary ECG acquisition for R-peak detector training, reducing overall examination time. The gating process is designed to be adaptable, completely blind and independent of patient characteristics, allowing wide and rapid deployment in clinical practice. The potential to employ a significantly limited set of leads enhances patient comfort.

Forward problem of electrocardiography based on cardiac source vector orientations.

To localize the unusual cardiac activities non-invasively, one has to build a prior forward model that relates the heart, torso, and detectors. This model has to be constructed to mathematically relate the geometrical and functional activities of the heart. Several methods are available to model the prior sources in the forward problem, which results in the lead field matrix generation. In the conventional technique, the lead field assumed the fixed prior sources, and the source vector orientations were presumed to be parallel to the detector plane with the unit strength in all directions. However, the anomalies cannot always be expected to occur in the same location and orientation, leading to misinterpretation and misdiagnosis. To overcome this, the work proposes a new forward model constructed using the VCG signals of the same subject. Furthermore, three transformation methods were used to extract VCG in constructing the time-varying lead fields to steer to the orientation of the source rather than just reconstructing its activities in the inverse problem. In addition, the unit VCG loop of the acute ischemia patient was extracted to observe the changes compared to the normal subject. The abnormality condition was achieved by delaying the depolarization time by 15ms. The results involving the unit vectors of VCG demonstrated the anisotropic nature of cardiac source orientations, providing information about the heart's electrical activity.

Utility of RAND/UCLA appropriateness method in validating multiple-choice questions on ECG.

OBJECTIVES: This study aimed to investigate the utility of the RAND/UCLA appropriateness method (RAM) in validating expert consensus-based multiple-choice questions (MCQs) on electrocardiogram (ECG). METHODS: According to the RAM user's manual, nine panelists comprising various experts who routinely handle ECGs were asked to reach a consensus in three phases: a preparatory phase (round 0), an online test phase (round 1), and a face-to-face expert panel meeting (round 2). In round 0, the objectives and future timeline of the study were elucidated to the nine expert panelists with a summary of relevant literature. In round 1, 100 ECG questions prepared by two skilled cardiologists were answered, and the success rate was calculated by dividing the number of correct answers by 9. Furthermore, the questions were stratified into "Appropriate," "Discussion," or "Inappropriate" according to the median score and interquartile range (IQR) of appropriateness rating by nine panelists. In round 2, the validity of the 100 ECG questions was discussed in an expert panel meeting according to the results of round 1 and finally reassessed as "Appropriate," "Candidate," "Revision," and "Defer." RESULTS: In round 1 results, the average success rate of the nine experts was 0.89. Using the median score and IQR, 54 questions were classified as " Discussion." In the expert panel meeting in round 2, 23% of the original 100 questions was ultimately deemed inappropriate, although they had been prepared by two skilled cardiologists. Most of the 46 questions categorized as "Appropriate" using the median score and IQR in round 1 were considered "Appropriate" even after round 2 (44/46, 95.7%). CONCLUSIONS: The use of the median score and IQR allowed for a more objective determination of question validity. The RAM may help select appropriate questions, contributing to the preparation of higher-quality tests.

A Smartphone-Based M-Health Monitoring System for Arrhythmia Diagnosis.

Deep learning technology has been widely adopted in the research of automatic arrhythmia detection. However, there are several limitations in existing diagnostic models, e.g., difficulties in extracting temporal information from long-term ECG signals, a plethora of parameters, and sluggish operation speed. Additionally, the diagnosis performance of arrhythmia is prone to mistakes from signal noise. This paper proposes a smartphone-based m-health system for arrhythmia diagnosis. First, we design a cycle-GAN-based ECG denoising model which takes real-world noise signals as input and aims to produce clean ECG signals. In order to train its two generators and two discriminators simultaneously, we explore an unsupervised pre-training strategy to initialize the generator and accelerate the convergence speed during training. Second, we propose an arrhythmia diagnosis model based on the time convolution network (TCN). This model can identify 34 common arrhythmia events using eight-lead ECG signals, and we deploy such a model on the Android platform to develop an at-home ECG monitoring system. Experimental results have demonstrated that our approach outperforms the existing noise reduction methods and arrhythmia diagnosis models in terms of denoising effect, recognition accuracy, model size, and operation speed, making it more suitable for deployment on mobile devices for m-health monitoring services.

Wearable Ring-Shaped Biomedical Device for Physiological Monitoring through Finger-Based Acquisition of Electrocardiographic, Photoplethysmographic, and Galvanic Skin Response Signals: Design and Preliminary Measurements.

Wearable health devices (WHDs) are rapidly gaining ground in the biomedical field due to their ability to monitor the individual physiological state in everyday life scenarios, while providing a comfortable wear experience. This study introduces a novel wearable biomedical device capable of synchronously acquiring electrocardiographic (ECG), photoplethysmographic (PPG), galvanic skin response (GSR) and motion signals. The device has been specifically designed to be worn on a finger, enabling the acquisition of all biosignals directly on the fingertips, offering the significant advantage of being very comfortable and easy to be employed by the users. The simultaneous acquisition of different biosignals allows the extraction of important physiological indices, such as heart rate (HR) and its variability (HRV), pulse arrival time (PAT), GSR level, blood oxygenation level (SpO2), and respiratory rate, as well as motion detection, enabling the assessment of physiological states, together with the detection of potential physical and mental stress conditions. Preliminary measurements have been conducted on healthy subjects using a measurement protocol consisting of resting states (i.e., SUPINE and SIT) alternated with physiological stress conditions (i.e., STAND and WALK). Statistical analyses have been carried out among the distributions of the physiological indices extracted in time, frequency, and information domains, evaluated under different physiological conditions. The results of our analyses demonstrate the capability of the device to detect changes between rest and stress conditions, thereby encouraging its use for assessing individuals' physiological state. Furthermore, the possibility of performing synchronous acquisitions of PPG and ECG signals has allowed us to compare HRV and pulse rate variability (PRV) indices, so as to corroborate the reliability of PRV analysis under stationary physical conditions. Finally, the study confirms the already known limitations of wearable devices during physical activities, suggesting the use of algorithms for motion artifact correction.

Electrocardiographic, biochemical, and scintigraphic evidence for the cardioprotective effect of paricalcitol and vitamin D3 on doxorubicin-induced acute cardiotoxicity in rats.

AIM: We aimed to investigate the possible cardioprotective effects of paricalcitol (PR), its vitamin D receptor agonist, and vitamin D3 (VIT-D3) on an experimental model of doxorubicin (DX) cardiotoxicity by 99mTc-PYP scintigraphy, electrocardiographic (ECG) and biochemical methods. METHOD: Forty-two male Wistar/Albino rats (250‒300 g; aged 10‒12 weeks) were randomly separated into six groups, namely into control (CN), doxorubicin (DX), paricalcitol (PR), vitamin D3 (VIT-D3), paricalcitol + doxorubicin (PR+DX), and vitamin D3 + doxorubicin (VIT-D3+DX) groups. Cardiotoxicity was induced by three doses of DX (18 mg/kg, i.p.) at 24-hour intervals on days 18, 19 and 20. PR (0.5 ug/ kg, i.p) and VIT-D3 (5,000 IU/kg, i.p) were injected for 20 days before and after the application of DX (18 mg/kg, i.p.). On day 21 of the experiment, biochemical parameters [tumor necrosis factor TNF-alpha (TNF-α); interleukin-6 (IL-6), nitric oxide (NO), and cardiac troponin T (cTnT)], as well as ECG and scintigraphic (99mTc-PYP) features were assessed. RESULTS: Compared to CN, DX significantly raised TNF-α, IL-6, and NO in heart tissue, cTnT in serum, 99mTc-PYP uptake in the myocardium, and ECG parameters, specifically QRS complex duration, QT interval duration, and ST-segment amplitude, while also reducing heart rate (p<0.001). Pretreatment with PR and VIT-D3 mitigated these abnormalities produced by DX in the heart (p<0.001). CONCLUSION: Results show that vitamin D receptor agonist paricalcitol and vitamin D protect against DX-induced cardiotoxicity through anti-inflammatory and antioxidant effects (Fig. 4, Ref. 59). Text in PDF www.elis.sk Keywords: paricalcitol, doxorubicin, vitamin D, ECG, 99mTc-PYP scintigraphy, cardiotoxicity, inflammation.

Evaluation of the introduction of a single-lead ECG device and digital cardiologist consultation platform among general practitioners in the Netherlands.

AIM: To evaluate the use of a single-lead electrocardiography (1L-ECG) device and digital cardiologist consultation platform in diagnosing arrhythmias among general practitioners (GPs). BACKGROUND: Handheld 1L-ECG offers a user-friendly alternative to conventional 12-lead ECG in primary care. While GPs can safely rule out arrhythmias on 1L-ECG recordings, expert consultation is required to confirm suspected arrhythmias. Little is known about GPs' experiences with both a 1L-ECG device and digital consultation platform for daily practice. METHODS: We used two distinct methods in this study. First, in an observational study, we collected and described all cases shared by GPs within a digital cardiologist consultation platform initiated by a local GP cooperative. This GP cooperative distributed KardiaMobile 1L-ECG devices among all affiliated GPs (n = 203) and invited them to this consultation platform. In the second part, we used an online questionnaire to evaluate the experiences of these GPs using the KardiaMobile and consultation platform. FINDINGS: In total, 98 (48%) GPs participated in this project, of whom 48 (49%) shared 156 cases. The expert panel was able to provide a definitive rhythm interpretation in 130 (83.3%) shared cases and answered in a median of 4 min (IQR: 2-18). GPs responding to the questionnaire (n = 43; 44%) thought the KardiaMobile was of added value for rhythm diagnostics in primary care (n = 42; 98%) and easy to use (n = 41; 95%). Most GPs (n = 36; 84%) valued the feedback from the cardiologists in the consultation platform. GPs experienced this project to have a positive impact on both the quality of care and diagnostic efficiency for patients with (suspected) cardiac arrhythmias. Although we lack a comprehensive picture of experienced impediments by GPs, solving technical issues was mentioned to be helpful for further implementation. More research is needed to explore reasons of GPs not motivated using these tools and to assess real-life clinical impact.

Machine learning based detection of T-wave alternans in real ambulatory conditions.

BACKGROUND AND OBJECTIVE: T-wave alternans (TWA) is a fluctuation in the repolarization morphology of the ECG. It is associated with cardiac instability and sudden cardiac death risk. Diverse methods have been proposed for TWA analysis. However, TWA detection in ambulatory settings remains a challenge due to the absence of standardized evaluation metrics and detection thresholds. METHODS: In this work we use traditional TWA analysis signal processing-based methods for feature extraction, and two machine learning (ML) methods, namely, K-nearest-neighbor (KNN) and random forest (RF), for TWA detection, addressing hyper-parameter tuning and feature selection. The final goal is the detection in ambulatory recordings of short, non-sustained and sparse TWA events. RESULTS: We train ML methods to detect a wide variety of alternant voltage from 20 to 100 µV, i.e., ranging from non-visible micro-alternans to TWA of higher amplitudes, to recognize a wide range in concordance to risk stratification. In classification, RF outperforms significantly the recall in comparison with the signal processing methods, at the expense of a small lost in precision. Despite ambulatory detection stands for an imbalanced category context, the trained ML systems always outperform signal processing methods. CONCLUSIONS: We propose a comprehensive integration of multiple variables inspired by TWA signal processing methods to fed learning-based methods. ML models consistently outperform the best signal processing methods, yielding superior recall scores.

[New frontiers in pacing: from myocardial pacing to conduction system pacing]. / Nuove frontiere dell'elettrostimolazione: dal pacing del miocardio comune al pacing del sistema di conduzione.

For many years, cardiac pacing has been based on the stimulation of right ventricular common myocardium to correct diseases of the conduction system. The birth and the development of cardiac resynchronization have led to growing interest in the correction and prevention of pacing-induced dyssynchrony. Many observational studies and some randomized clinical trials have shown that conduction system pacing (CSP) can not only prevent pacing-induced dyssynchrony but can also correct proximal conduction system blocks, with reduction of QRS duration and with equal or greater effectiveness than biventricular pacing. Based on these results, many Italian electrophysiologists have changed the stimulation target from the right ventricular common myocardium to CSP. The two techniques with greater clinical impact are the His bundle stimulation and the left bundle branch pacing. The latter, in particular, because of its easier implantation technique and better electric parameters, is spreading like wildfire and is representing a real revolution in the cardiac pacing field. However, despite the growing amount of data, until now, the European Society of Cardiology guidelines give a very limited role to CSP.

Output-dependent His bundle pacing: Unexpected His-Purkinje system pathology unmasking.

The anatomy of the His-Purkinje system has been studied, yet there remains a knowledge gap regarding the impact of His bundle pacing and its electrocardiographic implications. This case report highlights the presence of His-Purkinje system pathology without apparent clues on the surface electrocardiogram (EKG). By observing identical QRS morphology with varying HV intervals resulting from different pacing outputs, we demonstrate the presence of an electrical propagation block within the His bundle.

Clinical significance of R-wave amplitude in lead V1 and inferobasal myocardial infarction in patients with inferior wall myocardial infarction.

OBJECTIVE: To assess electrocardiogram (ECG) for risk stratification in inferior ST-elevation myocardial infarction (STEMI) patients within 24 h. METHODS: Three hundred thirty-four patients were divided into four ECG-based groups: Group A: R V1 <0.3 mV with ST-segment elevation (ST↑) V7-V9, Group B: R V1 <0.3 mV without ST↑ V7-V9, Group C: R V1 ≥0.3 mV with ST↑ V7-V9, and Group D: R V1 ≥0.3 mV without ST↑ V7-V9. RESULTS: Group A demonstrated the longest QRS duration, followed by Groups B, C, and D. ECG signs for right ventricle (RV) infarction were more common in Groups A and B (p < .01). ST elevation in V6, indicative of left ventricle (LV) lateral injury, was more higher in Group C than in Group A, while the ∑ST↑ V3R + V4R + V5R, representing RV infarction, showed the opposite trend (p < .05). The estimated LV infarct size from ECG was similar between Groups A and C, yet Group A had higher creatine kinase MB isoform (CK-MB; p < .05). Cardiac troponin I (cTNI) was higher in Groups A and C than in B and D (p < .05 and p = .16, respectively). NT-proBNP decreased across groups (p = .20), with the highest left ventricular ejection fraction (LVEF) observed in Group D (p < .05). Group A notably demonstrated more cardiac dysfunction within 4 h post-onset. CONCLUSIONS: For inferior STEMI patients, concurrent R V1 <0.3 mV with ST↑ V7-V9 suggests prolonged ventricular activation and notable myocardial damage. RV infarction's dominance over LV lateral injury might explain these observations.

Characteristics of Patients With Acute Coronary Syndrome and Normal Electrocardiogram.

Background: Patients with acute coronary syndrome (ACS) and normal electrocardiogram (ECG) may have an increased risk of late diagnosis and complications of the disease. Objective: To study the demographic, angiographic and echocardiographic characteristics of patients hospitalized for ACS in whom the ECG was normal on admission to the hospital. Methods: This retrospective study included patients who were hospitalized for ACS without ST-elevation between 2015 and 2023 and who had coronary artery disease (CAD) confirmed by coronary angiography. By further inspection of the electronic databases, patients with ACS who had a normal ECG on admission were filtered out and analyzed separately. Results: Of the total 3137 patients with suspected ACS without ST-elevation, 129 patients (4.1%) were diagnosed as having ACS with a normal ECG. In three patients a non-atherosclerotic cause for the ACS was found. A significantly higher proportion of patients had single-vessel (54.3%) compared to two-vessel (29.5%) and three-vessel (14%) CAD. In addition to a normal ECG, 5.7% of patients with single-vessel CAD and 3.5% of patients with multi-vessel CAD had normal troponin levels and normal regional LV systolic function on echocardiography. Conclusion: Less than 5% of hospitalized patients with ACS without ST-elevation had a normal ECG on admission. The majority of these patients have single-vessel CAD. In about 5% of patients with single-vessel CAD, neither elevated troponin levels nor LV asynergy are detected.