---
license: odbl
pretty_name: MIMIC-IV-ECG (Parquet mirror)
tags:
  - ecg
  - medical
  - physionet
  - mimic-iv
task_categories:
  - time-series-forecasting
---

# MIMIC-IV-ECG -- Parquet mirror

Unofficial Parquet re-encoding of **MIMIC-IV-ECG: Diagnostic
Electrocardiogram Matched Subset v1.0**, mirrored here from
[PhysioNet](https://physionet.org/content/mimic-iv-ecg/1.0/) for convenient loading in ML pipelines (one row
per 12-lead/500Hz/10s study instead of paired WFDB `.hea`/`.dat` files).

This is **not** an official PhysioNet distribution. No signal processing
was applied -- each row holds the original `p_signal` values per lead as
published, just reshaped into columnar Parquet
(`lead_<name>: list<float32>`, one file per source `pXXXX` folder,
zstd-compressed).

## Source

- PhysioNet project page: https://physionet.org/content/mimic-iv-ecg/1.0/
- DOI: https://doi.org/10.13026/4nqg-sb35
- License: **Open Data Commons Open Database License v1.0 (ODbL 1.0)** -- https://opendatacommons.org/licenses/odbl/1-0/

## License and attribution

The original MIMIC-IV-ECG dataset is licensed under ODbL 1.0. This
repository is a Share-Alike **Adaptation** under that license (ODbL
section 4.4): it remains licensed under **ODbL 1.0**, and any further use,
redistribution, or adaptation of this repository must likewise remain
under ODbL 1.0 and retain the attribution below (ODbL section 4.3).

If you use this data, please cite the original dataset and PhysioNet as
requested by the authors and by PhysioNet's citation policy:

```bibtex
@article{PhysioNet-mimic-iv-ecg-1.0,
  author = {Gow, Brian and Pollard, Tom and Nathanson, Larry A and
    Johnson, Alistair and Moody, Benjamin and Fernandes, Chrystinne and
    Greenbaum, Nathaniel and Waks, Jonathan W and Eslami, Parastou and
    Carbonati, Tanner and Chaudhari, Ashish and Herbst, Elizabeth and
    Moukheiber, Dana and Berkowitz, Seth and Mark, Roger and Horng, Steven},
  title = {{MIMIC-IV-ECG: Diagnostic Electrocardiogram Matched Subset}},
  journal = {PhysioNet},
  year = {2023},
  month = sep,
  note = {Version 1.0},
  doi = {10.13026/4nqg-sb35},
  url = {https://doi.org/10.13026/4nqg-sb35}
}
```

```bibtex
@article{PhysioNet2026Platform,
  author = {Pollard, Tom and Moody, Benjamin and Lehman, Li-wei and others},
  title = {PhysioNet as a Global Platform for Biomedical Research},
  journal = {Nature Health},
  year = {2026},
  doi = {10.1038/s44360-026-00096-z}
}
```

```bibtex
@article{Goldberger2000PhysioBank,
  author = {Goldberger, Ary L and Amaral, Luis A N and Glass, Leon and
    Hausdorff, Jeffrey M and Ivanov, Plamen Ch and Mark, Roger G and
    Mietus, Joseph E and Moody, George B and Peng, Chung-Kang and
    Stanley, H Eugene},
  title = {PhysioBank, PhysioToolkit, and PhysioNet: Components of a New
    Research Resource for Complex Physiologic Signals},
  journal = {Circulation},
  volume = {101},
  number = {23},
  pages = {e215--e220},
  year = {2000},
  url = {http://circ.ahajournals.org/content/101/23/e215.full}
}
```

## Schema

Each row is one study: `subject_id`, `study_id`, `fs` (500.0), `sig_len`
(5000), and one `lead_<name>: list<float32>` column per of the 12 leads
(e.g. `lead_I`, `lead_II`, `lead_V1`, ...).