MIT-BIH Arrhythmia Database P-Wave Annotations

@article{PhysioNet-pwave-1.0.0,
  author = {Nemcova, Andrea and Smisek, Radovan and Vitek, Martin and Smital, Lukas},
  title = {{MIT-BIH Arrhythmia Database P-Wave Annotations}},
  journal = {{PhysioNet}},
  year = {2018},
  month = jan,
  note = {Version 1.0.0},
  doi = {10.13026/C2108F},
  url = {https://doi.org/10.13026/C2108F}
}

@article{pollard_physionet_2026, title = {{PhysioNet} as a global platform for biomedical research}, volume = {1}, issn = {3005-0693}, url = {https://doi.org/10.1038/s44360-026-00096-z}, doi = {10.1038/s44360-026-00096-z}, abstract = {Twenty-five years after its first public repository was established, the PhysioNet resource for biomedical and clinical data has enabled tens of thousands of research papers. Further efforts are ongoing to advance secure data sharing, community engagement and scalable research infrastructure.}, number = {8}, journal = {Nature Health}, author = {Pollard, Tom and Moody, Benjamin E. and Lehman, Li-wei H. and Gow, Brian J. and Fernandes, Chrystinne and Xie, Chen and Johnson, Alistair and Mark, Roger G. and Heldt, Thomas}, month = aug, year = {2026}, pages = {792--795},}

Abstract

This database provides manually reviewed P-wave annotations for twelve ECG recordings from the MIT-BIH Arrhythmia Database [1]. The recordings were selected to support evaluation of P-wave detection across a range of rhythms, including pathological conditions that can make detection difficult. One expert annotated the recordings, a second expert reviewed the annotations, and ambiguous segments were discussed jointly. The annotations are supplied in standard WFDB format alongside the corresponding ECG signals and header files. They support the development and evaluation of automated P-wave detection methods, although annotation completeness and accuracy are not guaranteed.

Background

Reliable detection of P-waves is challenging in pathological ECG recordings. Algorithms evaluated primarily on sinus-rhythm recordings may perform less well when arrhythmias alter the timing or morphology of atrial and ventricular activity. Reference annotations for such recordings are therefore useful for evaluating P-wave detection under more demanding conditions [2].

This database supplements the MIT-BIH Arrhythmia Database [1] with P-wave annotations for twelve selected recordings. The accompanying study investigated P-wave detection during ventricular extrasystoles and evaluated the proposed method on recordings with and without ventricular extrasystoles [2].

Methods

Twelve recordings were selected from the MIT-BIH Arrhythmia Database [1]. The selection includes recordings with pathological features that complicate P-wave detection, as well as recordings used to evaluate detection during sinus rhythm [2].

P-waves were annotated manually using SignalPlant [3]. One expert produced the initial annotations, and a second expert checked them. Ambiguous portions of the recordings were reviewed jointly by both annotators.

Both annotators had more than five years of professional experience analyzing Holter ECG recordings and held Master's degrees in Biomedical Engineering and Bioinformatics from Brno University of Technology.

Data Description

The database contains twelve P-wave annotation files in standard WFDB format, each with the extension .pwave. Corresponding ECG waveform files (.dat) and header files (.hea) are also included. The source recordings are approximately half an hour long, contain two ECG channels, and were sampled at 360 samples per second per channel [1].

The following table lists the annotation files and the corresponding source records and subject information.

Usage Notes

The annotation files can be read using rdann from the WFDB Software Package, specifying pwave as the annotator name.

These annotations may be used to evaluate automated P-wave detection algorithms. However, not every P-wave is necessarily annotated, and some annotations may be incorrect. An unannotated interval should therefore not automatically be interpreted as evidence that no P-wave is present.

Users should report the recordings included in their evaluation, any excluded intervals, and the temporal tolerance used to match detections to reference annotations. Because the recordings overlap with the MIT-BIH Arrhythmia Database, this overlap should be considered when constructing training and evaluation sets.

When using these annotations, please cite the accompanying study [2]. When using the source ECG recordings, also cite the MIT-BIH Arrhythmia Database publication [1].

Acknowledgements

We acknowledge the developers of the MIT-BIH Arrhythmia Database for providing the source ECG recordings [1], the experts who manually annotated and reviewed the P-waves, and the developers of SignalPlant, which was used for annotation [3].

References

1. Moody GB, Mark RG. The impact of the MIT-BIH Arrhythmia Database. IEEE Engineering in Medicine and Biology Magazine. 2001;20(3):45–50. doi:10.1109/51.932724.
2. Maršánová L, Němcová A, Smíšek R, Goldmann T, Vítek M, Smital L. Automatic Detection of P Wave in ECG During Ventricular Extrasystoles. In: Lhotska L, Sukupova L, Lacković I, Ibbott GS, editors. World Congress on Medical Physics and Biomedical Engineering 2018. IFMBE Proceedings, vol. 68/2. Singapore: Springer; 2019. p. 381–385. doi:10.1007/978-981-10-9038-7_72.
3. Plesinger F, Jurco J, Halamek J, Jurak P. SignalPlant: an open signal processing software platform. Physiological Measurement. 2016;37(7):N38–N48. doi:10.1088/0967-3334/37/7/N38.