An open-access simultaneous electrocardiogram and phonocardiogram database

Objective.The EPHNOGRAM project aimed to develop a low-cost, low-power device for simultaneous electrocardiogram (ECG) and phonocardiogram (PCG) recording, with additional channels for environmental audio to enhance PCG through active noise cancellation. The objective was to study multimodal electro-mechanical activities of the heart, offering insights into the differences and synergies between these modalities during various cardiac activity levels.Approach.We developed and tested several hardware prototypes of a simultaneous ECG-PCG acquisition device. Using this technology, we collected simultaneous ECG and PCG data from 24 healthy adults during different physical activities, including resting, walking, running, and stationary biking, in an indoor fitness center. The data were annotated using a robust software that we developed for detecting ECG R-peaks and PCG S1 and S2 components, and overseen by a human expert. We also developed machine learning models using ECG-based, PCG-based, and joint ECG-PCG features, like R-R and S1-S2 intervals, to classify physical activities and analyze electro-mechanical dynamics.Main results.The results show a significant coupling between ECG and PCG components, especially during high-intensity exercise. Notable micro-variations in S2-based heart rate show differences in the heart's electrical and mechanical functions. The Lomb-Scargle periodogram and approximate entropy analyses confirm the higher volatility of S2-based heart rate compared to ECG-based heart rate. Correlation analysis shows stronger coupling between R-R and R-S1 intervals during high-intensity activities. Hybrid ECG-PCG features, like the R-S2 interval, were identified as more informative for physical activity classification through mRMR feature selection and SHAP value analysis.Significance.The EPHNOGRAM database, is available on PhysioNet. The database enhances our understanding of cardiac function, enabling future studies on the heart's mechanical and electrical interrelationships. The results of this study can contribute to improved cardiac condition diagnoses. Additionally, the designed hardware has the potential for integration into wearable devices and the development of multimodal stress test technologies.

Analysis of an adaptive lead weighted ResNet for multiclass classification of 12-lead ECGs

Background. Twelve lead ECGs are a core diagnostic tool for cardiovascular diseases. Here, we describe and analyse an ensemble deep neural network architecture to classify 24 cardiac abnormalities from 12 lead ECGs.Method. We proposed a squeeze and excite ResNet to automatically learn deep features from 12-lead ECGs, in order to identify 24 cardiac conditions. The deep features were augmented with age and gender features in the final fully connected layers. Output thresholds for each class were set using a constrained grid search. To determine why the model made incorrect predictions, two expert clinicians independently interpreted a random set of 100 misclassified ECGs concerning left axis deviation.Results. Using the bespoke weighted accuracy metric, we achieved a 5-fold cross-validation score of 0.684, and sensitivity and specificity of 0.758 and 0.969, respectively. We scored 0.520 on the full test data, and ranked 2nd out of 41 in the official challenge rankings. On a random set of misclassified ECGs, agreement between two clinicians and training labels was poor (clinician 1:κ= -0.057, clinician 2:κ= -0.159). In contrast, agreement between the clinicians was very high (κ= 0.92).Discussion. The proposed prediction model performed well on the validation and hidden test data in comparison to models trained on the same data. We also discovered considerable inconsistency in training labels, which is likely to hinder development of more accurate models.

Automated JTpeak analysis by BRAVO

Using BRAVO algorithm (AMPS-LLC, NY, v4.4.0), 5223 ECGs from a publicly available annotated dataset from a randomized clinical trial on four different compounds and placebo were analyzed. ECGs were automatically processed and JTp interval was computed on: 12 standard ECG leads, Vector Magnitude (VM), and root mean square (RMS) leads. On VM and RMS, JTp intervals were nearly identical (228 ± 29 vs. 227 ± 30 ms respectively, with correlation of 0.99, p < 0.0001). On lead II, JTp interval was about 10 ms longer, but highly correlated with that measured on VM (0.94, p < 0.0001). Similarly, on lead V5, JTp was about 8 ms longer than on VM, with a correlation of 0.95, p < 0.0001. When compared to the public available annotations, JTp by BRAVO generated longer (about 8 ms) measurement and evidenced outliers conducible to both the T-wave peak (in few ECGs presenting notched shapes) and, to a lesser degree, to the J point, due to variability of the two algorithms. Differences on the drug-induced effect from the four compounds were negligible.

UnoViS: the MedIT public unobtrusive vital signs database

While PhysioNet is a large database for standard clinical vital signs measurements, such a database does not exist for unobtrusively measured signals. This inhibits progress in the vital area of signal processing for unobtrusive medical monitoring as not everybody owns the specific measurement systems to acquire signals. Furthermore, if no common database exists, a comparison between different signal processing approaches is not possible. This gap will be closed by our UnoViS database. It contains different recordings in various scenarios ranging from a clinical study to measurements obtained while driving a car. Currently, 145 records with a total of 16.2 h of measurement data is available, which are provided as MATLAB files or in the PhysioNet WFDB file format. In its initial state, only (multichannel) capacitive ECG and unobtrusive PPG signals are, together with a reference ECG, included. All ECG signals contain annotations by a peak detector and by a medical expert. A dataset from a clinical study contains further clinical annotations. Additionally, supplementary functions are provided, which simplify the usage of the database and thus the development and evaluation of new algorithms. The development of urgently needed methods for very robust parameter extraction or robust signal fusion in view of frequent severe motion artifacts in unobtrusive monitoring is now possible with the database.

An Open-source Toolbox for Analysing and Processing PhysioNet Databases in MATLAB and Octave

The WaveForm DataBase (WFDB) Toolbox for MATLAB/Octave enables integrated access to PhysioNet's software and databases. Using the WFDB Toolbox for MATLAB/Octave, users have access to over 50 physiological databases in PhysioNet. The toolbox provides access over 4 TB of biomedical signals including ECG, EEG, EMG, and PLETH. Additionally, most signals are accompanied by metadata such as medical annotations of clinical events: arrhythmias, sleep stages, seizures, hypotensive episodes, etc. Users of this toolbox should easily be able to reproduce, validate, and compare results published based on PhysioNet's software and databases.

Ventricular cycle length characteristics estimative of prolonged RR interval during atrial fibrillation

BACKGROUND

When atrial fibrillation (AF) is incessant, imaging during a prolonged ventricular RR interval may improve image quality. It was hypothesized that long RR intervals could be predicted from preceding RR values.

METHODS

From the PhysioNet database, electrocardiogram RR intervals were obtained from 74 persistent AF patients. An RR interval lengthened by at least 250 ms beyond the immediately preceding RR interval (termed T0 and T1, respectively) was considered prolonged. A two-parameter scatterplot was used to predict the occurrence of a prolonged interval T0. The scatterplot parameters were: (1) RR variability (RRv) estimated as the average second derivative from 10 previous pairs of RR differences, T13-T2, and (2) Tm-T1, the difference between Tm, the mean from T13 to T2, and T1. For each patient, scatterplots were constructed using preliminary data from the first hour. The ranges of parameters 1 and 2 were adjusted to maximize the proportion of prolonged RR intervals within range. These constraints were used for prediction of prolonged RR in test data collected during the second hour.

RESULTS

The mean prolonged event was 1.0 seconds in duration. Actual prolonged events were identified with a mean positive predictive value (PPV) of 80% in the test set. PPV was >80% in 36 of 74 patients. An average of 10.8 prolonged RR intervals per 60 minutes was correctly identified.

CONCLUSIONS

A method was developed to predict prolonged RR intervals using two parameters and prior statistical sampling for each patient. This or similar methodology may help improve cardiac imaging in many longstanding persistent AF patients.