2026-10-14T03:26:56Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33942946 2026-09-21T13:22:35Z category_24463 item_type_3 month_year_09_2026

A publicly available 12-lead ECG database for ventricular arrhythmia research covering 11,467 records Huihui Chang (25076341) Cardiology (incl. cardiovascular diseases) Ventricular arrhythmias <p dir="ltr">Ventricular arrhythmias are one of the major electrophysiological mechanisms underlying sudden cardiac death, while electrocardiography (ECG) is an important tool for their identification and phenotypic analysis. However, the abrupt onset and low incidence of high-risk ventricular arrhythmias make ECG recordings of these events difficult to obtain, and existing public datasets primarily focus on general arrhythmia or cardiovascular disease classification.</p><p dir="ltr">To address this gap, we present ZZU-VADB, a large, publicly available clinical ECG database for ventricular arrhythmia research. It comprises 11,467 12-lead ECG recordings from 4,345 patients, collected at the First Affiliated Hospital of Zhengzhou University between January 2015 and December 2025, including 6,623 ventricular-positive and 4,844 ventricular-negative recordings. The dataset standardizes original clinical diagnostic statements and organizes them into a hierarchical label taxonomy comprising six non-mutually exclusive ventricular superclasses and 44 subclasses, with mappings to standardized SCP, AHA, and CHN codes. The combination of diagnostic likelihoods, signal quality indices, and the multilabel structure of clinical ECG annotations, as well as predefined patient-level folds for training, validation, and testing, makes ZZU-VADB a resource for ventricular arrhythmia detection, multilabel phenotype analysis, and uncertainty modelling.</p> 2026-09-21T13:22:35Z Dataset Dataset 10.6084/m9.figshare.33942946.v2 https://figshare.com/articles/dataset/A_publicly_available_12-lead_ECG_database_for_ventricular_arrhythmia_research_covering_11_467_records/33942946 CC BY 4.0