Abstract
"67,830 record sets of multi-channel physiologic waveforms (ECG, ABP, respiration, PPG) and vital sign time series from approximately 30,000 ICU patients, supporting clinical research and ML model development for critical care."
Description
Overview
The MIMIC-III Waveform Database contains thousands of recordings of multiple physiologic signals ('waveforms') and time series of vital signs ('numerics') collected from bedside patient monitors in adult and neonatal intensive care units (ICUs).
Data Collection
- 67,830 record sets for approximately 30,000 ICU patients, collected in a largely automated fashion from all bedside monitors in certain ICUs.
- Almost all record sets include a waveform record (typically ECG, ABP, respiration, PPG, and other signals digitized at 125 Hz) and a numerics record (periodic measurements such as heart rate, oxygen saturation, blood pressure sampled at 1 Hz or 1/min).
- Recording lengths vary; most are a few days in duration, but some are several weeks long.
Signals
- Waveforms: Up to 8 simultaneous channels including ECG (multiple leads), arterial blood pressure (ABP), photoplethysmogram (PPG), and respiration.
- Numerics: Heart rate, respiration rate, SpOâ‚‚, and systolic/mean/diastolic blood pressure, plus other vital signs as available.
Use Cases
- Developing and benchmarking patient monitoring algorithms for arrhythmia, hypotension, and respiratory distress.
- Training deep learning models for multi-signal physiologic event detection in critical care.
- Research on patient deterioration prediction and clinical decision support in ICUs.
📊 View Data Structure
To explore column names, data types, and sample rows, visit the official dataset page on PhysioNet.
Preview on PhysioNet
Cite This Dataset
Moody, Benjamin, Moody, George, Villarroel, Mauricio, Clifford, Gari D., & Silva, Ikaro (2020). MIMIC-III Waveform Database. [Dataset]. PhysioNet. https://doi.org/10.13026/c2607m
Select your preferred citation style above. The citation will automatically update and you can copy it to your clipboard.
Original source: PhysioNet (2020). Visit official page for more details.
Indexed by IoTDataset.com on Jan 30, 2026
Ready to Start Your Research?
Download this dataset directly from the official repository and start building your next breakthrough project.