Irena Jekova

• Prof., PhD
• Managing Director at Bulgarian Academy of Sciences

GPU-based deep neural networks (DNNs) are powerful for electrocardiogram (ECG) processing and rhythm classification. Although questions often arise about their practical application in embedded systems with low computational resources, few studies have investigated the associated challenges. This study aims to show a useful workflow for deploying a...

Objective: This study involving automated external defibrillators (AEDs) in early treatment of refibrillation aims to evaluate the performance of a new shock advisory system (SAS) during chest compressions (CC) in out-of-hospital cardiac arrest (OHCA) patients. Methods: This work focuses on AED SAS performance as a secondary outcome of DEFI 2022 c...

The aim of this study is to address the challenge of 12-lead ECG delineation by different encoder–decoder architectures of deep neural networks (DNNs). This study compares four concepts for encoder–decoders based on a fully convolutional architecture (CED-Net) and its modifications with a recurrent layer (CED-LSTM-Net), residual connections between...

Ambulatory 24–72 h Holter ECG monitoring is recommended for patients with suspected arrhythmias, which are often transitory and might remain unseen in resting standard 12-lead ECG. Holter manufacturers provide software diagnostic tools to assist clinicians in evaluating these large amounts of data. Nevertheless, the identification of short arrhythm...

This study aims to present a novel deep learning algorithm for a sliding shock advisory decision during cardiopulmonary resuscitation (CPR) and its performance evaluation as a function of the cumulative hands-off time. We retrospectively used 13,570 CPR episodes from out-of-hospital cardiac arrest (OHCA) interventions reviewed in a period of intere...

Chest compressions (CC) during cardiopulmonary resuscitation (CPR) produce strong artifacts in the electrocardiogram (ECG) via defi-pads. Heart rhythm can hardly be determined visually, but also by the shock-advisory algorithms in automated external defibrillators (AED). This study aims to investigate the potential of deep neural networks (DNN) as...

This study aims to explore a new deep learning strategy for electrocardiogram (ECG) denoising under adverse conditions of non-stationary power-line interference (PLI) with amplitude changes or nominal frequency deviations. The study presents an exhaustive training strategy of deep convolutional autoencoder (CAE), while input with one of the largest...

This study investigates the use of atrioventricular (AV) synchronization as an important diagnostic criterion for atrial fibrillation and flutter (AF) using one to twelve ECG leads. Heart rate, lead-specific AV conduction time, and P-/f-wave amplitude were evaluated by three representative ECG metrics (mean value, standard deviation), namely RR-int...

Patients’ telemetry deals with remote follow-up of vital parameters and clinical conditions without continuous direct contact between physician and patient. This study presents a telemetry system for real-time monitoring of cardiac patients and investigates its accuracy, reliability and applicability. The system includes personal analyzer (PA) for...

Considering the significant burden to patients and healthcare systems globally related to atrial fibrillation (AF) complications, the early AF diagnosis is of crucial importance. In the view of prominent perspectives for fast and accurate point-of-care arrhythmia detection, our study optimizes an artificial neural network (NN) classifier and ranks...

High performance of the shock advisory analysis of the electrocardiogram (ECG) during cardiopulmonary resuscitation (CPR) in out-of-hospital cardiac arrest (OHCA) is important for better management of the resuscitation protocol. It should provide fewer interruptions of chest compressions (CC) for non-shockable organized rhythms (OR) and Asystole, o...

The InterCriteria Analysis (ICrA) is based on the mathematical formalisms of index matrices and intuitionistic fuzzy sets. It has been elaborated to discern possible similarities in the behavior of criteria pairs when multiple objects are considered, allowing also the accounting of information uncertainty. The focus of this study is to validate the...

Objective The aim of this study was to present new combination of algorithms for rhythm analysis during cardiopulmonary resuscitation (CPR) in automated external defibrillators (AED), called Analyze Whilst Compressing (AWC), designed for decreasing pre-shock pause and early stopping of chest compressions (CC) for treating refibrillation. Methods T...

Electrocardiogram (ECG) analysis is important for the detection of pace pulse artifacts, since their existence indicates the presence of a pacemaker. ECG gives information on the proper functionality of the device and could help to evaluate the reaction of the heart. Beyond the challenges related to the diversity of ECG arrhythmias and pace pulses,...

Atrial fibrillation (AF) is associated with significant risk of heart failure and consequent death. Its episodic appearance, the wide variety of arrhythmias exhibiting irregular AF-like RR intervals and noises accompanying the ECG acquisition, impede the reliable AF detection. Therefore, the Computing in Cardiology Challenge 2017 organizers encoura...

Deep neural networks (DNN) are state-of-the-art machine learning algorithms that can be learned to self-extract significant features of the electrocardiogram (ECG) and can generally provide high-output diagnostic accuracy if subjected to robust training and optimization on large datasets at high computational cost. So far, limited research and opti...

Electrode reversal errors in standard 12-lead electrocardiograms (ECG) can produce significant ECG changes and, in turn, misleading diagnoses. Their detection is important but mostly limited to the design of criteria using ECG databases with simulated reversals, without Wilson's central terminal (WCT) potential change. This is, to the best of our k...

Objective This study aims to validate the 12-lead electrocardiogram (ECG) as a biometric modality based on two straightforward binary QRS template matching characteristics. Different perspectives of the human verification problem are considered, regarding the optimal lead selection and stability over sample size, gender, age, heart rate (HR). Meth...

The Archive contains all data related to the measurements of the pattern matching features in 12-lead ECG database, including all pairwise combinations between S1 and S2 sessions of the whole population, with clusterization to the subject’s identity (equal/different), data subset (training/test), age, gender, HR. (ZIP)

Human identification (ID) is a biometric task, comparing single input sample to many stored templates to identify an individual in a reference database. This paper aims to present the perspectives of personalized heartbeat pattern for reliable ECG-based identification. The investigations are using a database with 460 pairs of 12-lead resting electr...

The objective of biometrics is to identify subjects based on physiological or behavioral characteristics. This paper considers the spatial P-QRS-T loops of the vector-cardiogram (VCG), aiming to identify the most reliable VCG-based features for human verification. We analyze clinical standard 12-lead resting electrocardiograms (ECGs) from 460 non-c...

Background: Electrocardiogram (ECG)-based biometrics relies on the most stable and unique beat patterns, i.e. those with maximal intra-subject and minimal inter-subject waveform differences seen from different leads. We investigated methodology to evaluate those differences, aiming to rank the most prominent single and multi-lead ECG sets for biom...

Minimizing the impact of artifacts prior to the start of the ECG recording is an approach for providing a diagnostically reliable data. A solution to this problem is a continuous feedback of the ECG quality and prompt start of the recording at potentially the best quality. The real-time lead quality monitoring library (LQMLib) is introduced to trig...

This study investigates the potential of a set of ECG morphological features for person verification/identification. The measurements are done over 145 pairs of ECG recordings from healthy subjects, acquired 5 years apart (T1, T2 = T1+5 years). Time, amplitude, area and slope descriptors of the QRS-T pattern are analysed in 4 ECG leads, forming qua...

False intensive care unit (ICU) alarms induce stress in both patients and clinical staff and decrease the quality of care, thus significantly increasing both the hospital recovery time and rehospitalization rates. In the PhysioNet/CinC Challenge 2015 for reducing false arrhythmia alarms in ICU bedside monitor data, this paper validates the applicat...

Background: Electrocardiogram (ECG) biometrics is an advanced technology, not yet covered by guidelines on criteria, features and leads for maximal authentication accuracy. Objective: This study aims to define the minimal set of morphological metrics in 12-lead ECG by optimization towards high reliability and security, and validation in a person...

Traditional means for identity validation (PIN codes, passwords), and physiological and behavioral biometric characteristics (fingerprint, iris, and speech) are susceptible to hacker attacks and/or falsification. This paper presents a method for person verification/identification based on correlation of present-to-previous limb ECG leads: I ( r I...

This study presents a 2-stage heartbeat classifier of supraventricular (SVB) and ventricular (VB) beats. Stage 1 makes computationally-efficient classification of SVB-beats, using simple correlation threshold criterion for finding close match with a predominant normal (reference) beat template. The non-matched beats are next subjected to measuremen...

False Intensive Care Unit (ICU) alarms induce stress in both patients and clinical staff and decrease the quality of care, thus significantly increasing both the hospital recovery time and re-hospitalization rates. Therefore, PhysioNet/CinC Challenge 2015 encourages the development of algorithms for the analysis of bedside monitor data for robust d...

ECGs of 59 patients undergoing hemodialysis (HD): 52% males, age 59±13 years, renal disease duration 9.7±6.7 years, hemodialysis duration 5.2±4.4 years were recorded. Serum electrolytes (potassium-K, sodium-Na, phosphorus-Ph and calcium-Ca), urea and creatinine levels were evaluated before and after HD. ECG analysis on an average P-QRS-T interval i...

This study presents a method for automated detection of misplaced supplementary precordial leads, including the right-sided V3R, V4R and the posterior V8, V9 leads. Considering their uncommon use in clinical routine, a lead reversal is quite probable and could result in erroneous diagnosis and treatment. The method allows real-time implementation b...

This study investigates the potential of ECG morphological feature set for person identification. The measurements are done over 145 pairs of ECG recordings from healthy subjects, acquired 5 years apart. Time, amplitude, area and slope descriptors of the QRS-T pattern are analyzed in 4 ECG leads, forming quasi-orthogonal lead system (II&III, V1, V5...

This paper presents a high-resolution 16-channel ECG acquisition module with 24-bit amplitude resolution and sampling rate of 2kHz. The module is applied for collection of ECG database for the aims of development and testing of methods for person authentication via ECG. Such database could support the definition of optimal number of ECG leads and t...

ECGs of 59 patients undergoing hemodialysis (HD): 52% males, age 59±13 years, renal disease duration 9.7±6.7 years, hemodialysis duration 5.2±4.4 years were recorded. Serum electrolytes (potassium-K, sodium-Na, phosphorus-Ph and calcium-Ca), urea and creatinine levels were evaluated before and after HD. ECG analysis on an average P-QRS-T interval i...

The reliability of automatic person authentication has become critical, considering the necessary security for the cases of financial transactions, access control, travelling, etc. The traditional means for identity validation (PIN codes, passwords, etc.), as well as physiological and behavioural biometric characteristics (fingerprint, iris, speech...

This paper presents an ECG database, named 'PacedECGdb' (available at http://biomed.bas.bg/bioautomation/2014/vol_18.4/files/PacedECGdb.zip), which contains different arrhythmias generated by HKP (Heidelberger Praxisklinik) simulator, combined with artificially superimposed pacing pulses that cover the wide ranges of rising edge (from <10 μs to 100...

A new generation single-chip low-power ECG analog front-end module (ADAS1000) is used to develop a system for 12-lead high-resolution ECG and impedance-based respiration acquisition. The developed respiration rate (RR) measurement algorithm is validated on a test respiration database, including more than 650 reference 15s-window sliding of 5s with...

A pacemaker is a small battery-operated medical device that delivers electrical impulses to the heart in order to guarantee regular contractions. Depending on the number of active leads the pacemakers are single chamber, dual chamber and bi-ventricular. According to their programming the devices could be with fixed-rate, 'on demand' and raterespons...

This study presents a two-stage heartbeat classifier. The first stage makes initial assignment of beats towards continuously updated beat templates of the predominant rhythm, and calculates a set of features, tracking the morphology and RR-interval variation, and correlation to noise robust average beat templates. The second stage implements a deci...

This study presents a multichannel ECG quality monitoring system, which continuously scans the leads' status (validilead-off) and quality (0-100%), according to the ECG components in the low, medium and high frequency bands. The system aims to detect the optimal moment to start the record of a 10s resting ECG within the 1st minute of signal acquisi...

This study presents methods for automated detection of interchanged precordial and orthogonal ECG leads that may prevent from incorrect diagnosis and treatment. For precordial leads V1-V6, correlation coefficients of QRS-T patterns and time-alignment of R and S-peaks are assessed. For orthogonal leads (X,Y,Z), analysis of QRS loops in the frontal p...

This study aims to validate a shock advisory system in automated external defibrillators (AEDs) dedicated for ECG analysis during chest compressions (CC), guiding the rescuer to stop CC for rhythms which should be terminated by a defibrillation shock and to continue CC for non-shockable rhythms. The test-validation on a large database of out-of-hos...

This study presents a system for prediction of the weaning outcome based on classification trees and linear discriminant analysis. The design of several classification models involves anthropometric and diagnostic indicators, metabolic, ventilation, hemodynamic and heart rate variability indices, measured just before the weaning attempt and/or duri...

In the present work an attempt is made to evaluate objectively the ventilated patients' condition from the monitored parameters (standard physiological parameters, parameters of the ventilation and respiratory mechanics, parameters of the gas exchange and energy expenditure) in order to determine their readiness for weaning from mechanical ventilat...

BACKGROUND: This study aims to test the Automated external defibrillators (AED) transthoracic impedance cardiogram (EICG) as a potential sensor for detection of compromised hemodynamics in different arrythmias (atrial fibrillation (AFIB), atrial flutter (AFL), ventricular tachycardia (VT)) vs. sinus rhythm (SR). METHOD: ECG and ICG recordings via p...

The aim of the study is to investigate whether and how QRS-complex and T-wave heterogeneity is influenced by different cardiac risk factors and clinical data. Digital ECG during stress test was acquired in 106 patients (age 63±10 years, 45 males). Two indices obtained by Principal Component Analysis (PCA): complexity (PCA 1) and non-linear componen...

This study tracks the global tendency of the heart rate variability (HRV) profile over five ventilation modes, aiming to find evidences on the hypothesis that patients who failed during the weaning process manifest important differences in the autonomic nervous system activity. This preliminary study enrols 17 patients (7 successful (S), 10 failure...

This study presents a simple algorithm for pulse wave PW detection dedicated to real-time pulse sensing devices in an emergency. Two basic principles are implemented – identification of extrema by time-amplitude criteria, and validation of the most prominent rising edges preceding the systolic peak according to criteria for slope and similarity to...

This study aims to test the usability of the transthoracic impedance cardiogram (ICG) for assessment of the quality of myocardial contractions in atrial fibrillation (AFIB) vs. sinus rhythm (SR), using signals recorded via defibrillation pads during external cardioversion (ECV). Data from 88 patients with persistent AFIB who received planned ECV ar...

The aim of the study is to investigate whether and how QRS-complex and T-wave heterogeneity is influenced by different cardiac risk factors and clinical data. Digital ECG during stress test was acquired in 106 patients (age 63±10 years, 45 males). Two indices obtained by Principal Component Analysis (PCA): complexity (PCA 1) and non-linear componen...

Minimum “hands-off” intervals during cardiopulmonary resuscitation (CPR) are required to improve the success rate of defibrillation. In support of such life-saving practice, a shock advisory system (SAS) for automatic analysis of the electrocardiogram (ECG) contaminated by chest compression (CC) artefacts is presented. Ease of use for the automated...

This paper presents a system for detection of the most common noise types seen on the electrocardiogram (ECG) in order to evaluate whether an episode from 12-lead ECG is reliable for diagnosis. It implements criteria for estimation of the noise corruption level in specific frequency bands, aiming to identify the main sources of ECG quality disrupti...

This study investigates volume and pressure signals during four phases of volume controlled and pressure support ventilation, aiming to find predictors of the weaning outcome. The work of breathing to overcome the resistive properties of the respiratory system, the airway occlusion pressure (P0.1) and the maximal inspiratory pressure during sniff t...

This paper investigates the usability of the transthoracic impedance cardiogram (ICG) for providing information about the different quality of myocardial contraction in sinus rhythm (SR), asystole, 3 supraventricular and 3 ventricular arrhythmias when using the signal recorded via the defibrillation pads during external cardioversion. All arrhythmi...

Shortening hands-off intervals can improve benefits from defibrillation. This study presents the performance of a shock advisory system (SAS), which aims to decrease the pre-shock pauses by triggering fast rhythm analysis at minimal delay after end of chest compressions (CC). The SAS is evaluated on a database of 1301 samples from 311 out-of-hospit...

BACKGROUND Long pre-shock ‘hands-off’ intervals without chest compressions (CC) are associated with defibrillation failure. Current guidelines recommend shortening the ‘hands-off’ intervals. The aim of this study is to present the performance of a Shock advisory System (SAS) which is designed for triggering a fast ECG analysis at minimal delay afte...

This study aims to contribute to the scarce data available about the abilities of untrained lay persons to perform hands-only cardio-pulmonary resuscitation (CPR) on a manikin and the improvement of their skills during training with an autonomous CPR feedback device. The study focuses on the following questions: (i) Is there a need for such a CPR t...

This study evaluates the influence of analysis duration on the accuracy of AED shock advisory system (SAS), which is adapted to provide ‘Shock’/‘No Shock’ decision in real time at every second from 2s to 10s. MIT-BIH Malignant Ventricular Arrhythmia database is used for validation of the SAS accuracy on a computer. Four basic ECG criteria used in t...

The upgrade of mobile phones with applications for acquisition, pre-processing and transmitting the patient's ECG to a hospital unit would be of great benefit for prevention against the most frequent mortality caused by heart failure. This idea is promoted by the Computing in Cardiology Challenge 2011, which encourages the development of algorithms...

This study aims at validation of the specificity (Sp) of a shock advisory system (SAS) in automatic external defibrillators (AED) with non-shockable pediatric ECGs. Own pediatric ECG database is collected including lead II holter recordings from 46 children healthy and cardiac patients. A number of 10301 ten-second samples of non-shockable (N) rhyt...

Minimum "hands-off" intervals during cardiopulmonary resuscitation (CPR) are required to improve the success rate of defibrillation. In support of such life-saving practice, a shock advisory system (SAS) for automatic analysis of the electrocardiogram (ECG) contaminated by chest compression (CC) artefacts is presented. Ease of use for the automated...

Long interruptions of cardiopulmonary resuscitation (CPR) in case of a sudden cardiac arrest result in higher failure rate of resuscitation. The current work concerns the filtering of the chest compression (CC) artefacts during CPR, which is essential for the CPR continuation during electrocardiogram (ECG) analysis by automated external defibrillat...

Recent works are aimed at development of shock advisory systems (SAS) for automated external defibrillators (AEDs), which continuously analyze the electrocardiogram (ECG) during non-interrupted chest compressions (CC). Being also part of the cardiopulmonary resuscitation (CPR), small 'hands-off' intervals (CC pauses) for insufflations are interrupt...

Time-frequency domain features of chest compression (CC) artefacts, non-shockable (NShR) and shockable (ShR) rhythms were investigated. The aim was to provide reliable shock advisory analysis during CC by single channel electrocardiogram (ECG) processing. Three frequency bands were suggested to enhance specific components of the CC artefacts, NShR...

Cardio-pulmonary resuscitation (CPR) is a life-saving first aid which is part of the treatment given in case of sudden cardiac death. According to the American Heart Association (AHA) 2005 Guidelines for CPR, there are three key components related to the chest compressions which should be considered: (i) optimal compression depth between 3.8 and 5....

The morphological and rhythm analysis of the electrocardiogram (ECG) is based on ventricular beats detection, wave parameters measurement, as amplitudes, widths, polarities, intervals and relations between them, and a subsequent classification supporting the diagnostic process. Number of algorithms for detection and classification of the QRS comple...

This paper presents a bench study on a commercial automated external defibrillator (AED). The objective was to evaluate the performance of the defibrillation advisory system and its robustness against electromagnetic interferences (EMI) with central frequencies of 16.7, 50 and 60 Hz. The shock advisory system uses two 50 and 60 Hz band-pass filters...

The morphological and rhythm analysis of the electrocardiogram (ECG) is based on ventricular beats detection, wave parameters measurement, as amplitudes, widths, polarities, intervals and relations between them, and a subsequent classification supporting the diagnostic process. Number of algorithms for detection and classification of the QRS comple...

This study evaluates the efficacy of a Pulsed Biphasic Waveform (PBW) for treatment of out-of-hospital cardiac arrest (OHCA) patients in ventricular fibrillation (VF). Large database (2001-2006), collected with automated external defibrillators (AED), (FRED®, Schiller Medical SAS, France), is processed.In Study1 we compared the defibrillation effi...

The efficiency of a pulsed biphasic waveform (PBW) was compared with that of biphasic truncated exponential (BTE) waveforms. First defibrillation shock outcome was studied in a population of 104 out-of-hospital cardiac arrest patients in ventricular fibrillation as the presenting rhythm. The call to first shock time was 8.2+/-5.4 min. At 5s post-sh...

The most common way to diagnose cardiac dysfunctions is the ECG signal analysis, usually starting with the assessment of the QRS complex as the most significant wave in the electrocardiogram. Many methods for automatic heartbeats classification have been applied and reported in the literature but the use of different ECG features and the training a...