Database Credentialed Access
Stanford Sleep Bench
Magnus Ruud Kjaer , Rahul Thapa , Gauri Ganjoo , Hyatt Moore , Poul Jennum , M Brandon Westover , James Zou , Bryan He , Andreas Brink-Kjaer , Emmanuel Mignot
Published: Nov. 20, 2025. Version: 1.0
When using this resource, please cite:
(show more options)
Kjaer, M. R., Thapa, R., Ganjoo, G., Moore, H., Jennum, P., Westover, M. B., Zou, J., He, B., Brink-Kjaer, A., & Mignot, E. (2025). Stanford Sleep Bench (version 1.0). Brain Data Science Platform. https://doi.org/10.60508/0ta5-v163.
Abstract
The Sleep-Bench dataset is a collection of clinical polysomnography (PSG) recordings from Stanford Sleep Clinic. It consists of 17.467 PSG recordings spanning 163,650 hours. The dataset is accompanied by a codebase for benchmarking models in the Sleep-Bench GitHub repository.
Background
The Sleep-Bench dataset has been used to develop and systematically evaluate sleep foundation models (SleepFM) resulting in prediction of future disease onset directly from PSG data.
The dataset contains canonical information on sleep including AHI and annotated sleep stages. Additionally there are labels for 13 outcomes with onset time and censoring, ideal for prediction tasks.
Methods
The dataset comprises 6 channels of electroencephalography (EEG) (C3-M2, C4-M1, O1-M2, O2-M1, FP1-M2, FP2-M1), electroculography (EOG) (E1-M2, E2-M1), electromyography (EMG) (Chin (avg(L−ctr,R−ctr))), Leg(RAT−LAT)), echocardiogram (EKG) (EKGL - EKGR), and respiratory channels (Chest (thoracic effort), Abd (abdominal effort), Nasal (nasal airflow), Oral(oral airflow),SpO2 (oxygen saturation)). All signals were resampled to 128 Hz after appropriate filtering and derived from data in the Human Sleep Project.
Data Description
The Sleep-Bench dataset includes 17.467 PSG studies conducted on 12,794 distinct patients
- Sleep stages were annotated by certified sleep technologists as part of routine clinical care, according to the American Academy of Sleep Medicine (AASM) manual for the scoring of sleep. Stages were annotated in 30 second contiguous intervals, and include: wakefulness, (W) non-REM stage 1 (N1), non-REM stage 2 (N2), non-REM stage 3 (N3), and rapid eye movement (REM) sleep.
- AHI is noted for all subjects
- Time to outcomes
- Mortality
- Angina
- Atrial fibrillation and flutter
- Chronic kidney disease
- Dementia
- General atherosclerosis
- Heart failure
- Hypertension
- Hypotension
- Ischemic heart disease
- Myocardial infarction
- Pulmonary heart disease
- Type 2 diabetes
Usage Notes
Dataset Folder Structure:
Description:
Metadata File
|
Column Name |
Description |
|
SiteID |
Unique identifier of the hospital where the PSG was recorded. |
|
BDSPPatientID |
Unique identifier of the patient. |
|
CreationTime |
De-identified timestamp when the PSG was recorded. |
|
BidsFolder |
Folder where studies for a specific patient are available in the BDSP OpenData Repository. |
|
SessionID |
Folder in the BDSP OpenData Repository containing a specific study and its auxiliary files for a particular patient. |
|
PreSleepQuestionnaire |
Flag indicating if the study has a pre-sleep questionnaire. |
|
HasAnnotations |
Flag indicating if the study has annotations. |
Ethics
Data collection and sharing for Sleep-Bench is performed under Stanford University Institutional Review Board (protocol number: 69873). Sleep-Bench data was generated as part of usual patient care. All data is deidentified.
Acknowledgements
This work was supported by a grant from NIH/NHLBI (R01HL161253).
Conflicts of Interest
Dr. Westover is a co-founder, scientific advisor, consultant to, and has personal equity interest in Beacon Biosignals.
Parent Projects
Access
Access Policy:
Only credentialed users who sign the DUA can access the files.
License (for files):
BDSP Credentialed Health Data License 1.5.0
Data Use Agreement:
BDSP Credentialed Health Data Use Agreement
Required training:
CITI Data or Specimens Only Research
Discovery
DOI:
https://doi.org/10.60508/0ta5-v163
Topics:
foundation model
sleep
Project Website:
https://github.com/RuudeResearch/SleepBench
Corresponding Author
Files
- be a credentialed user
- complete required training:
- CITI Data or Specimens Only Research You may submit your training here.
- sign the data use agreement for the project