We use cookies and similar technologies to help personalize content, tailor and measure ads, and provide a better experience. By clicking "Accept All", you agree to our use of cookies.
You can customize your preferences or learn more in our .
- Epoch-by-epoch classification
- Visual timeline and distribution charts
- Detailed sleep architecture analysis
- Cardiovascular disease risk
- Neurological condition prediction
- Comprehensive risk categorization
How It Works
1. Upload Data
Upload your polysomnography recording in EDF or HDF5 format
2. AI Analysis
SleepFM model processes your data using advanced deep learning
3. View Results
Get detailed visualizations and insights from your sleep data
SleepFM: Decoding a Single Night's Sleep to Predict Future Disease
Published in Nature Medicine (2025) — A breakthrough multimodal foundation model that transforms sleep analysis
Rich data from standard sleep studies (polysomnography/PSG) is too complex for conventional analysis. Traditional methods struggle to extract meaningful health insights from the intricate patterns of brain activity (EEG), heart rhythms (ECG), breathing (respiratory), and muscle activity (EMG) recorded during sleep.
Brain (EEG)
Neural activity patterns
Heart (ECG)
Cardiac rhythms
Respiratory
Breathing patterns
Muscle (EMG)
Movement signals
SleepFM integrates multiple brain and body signals simultaneously, learning the complex relationships between sleep patterns and health outcomes. Like a language model learns from text, SleepFM learns from the "language" of sleep physiology.
Massive Training Dataset
585,000+
hours of sleep data from 65,000+ participants
Multimodal Integration
Simultaneously processes EEG, ECG, respiratory, and EMG signals to capture the full complexity of sleep physiology
SleepFM achieves a C-Index of at least 0.75 for 130 different medical conditions, demonstrating remarkable accuracy in predicting future disease risk from sleep patterns alone.
Key Disease Prediction Accuracy (C-Index)
Dementia
All-Cause Mortality
Myocardial Infarction
Heart Failure
Chronic Kidney Disease
Stroke
SleepFM is more accurate than models using only demographics or raw PSG data without pretraining. This breakthrough enables earlier disease detection, personalized risk assessment, and proactive intervention strategies.
5-17%
Improvement over baseline models
130+
Diseases predicted accurately
1 Night
Single sleep study required
Privacy & Security
All data is encrypted and securely stored. Your sleep recordings are processed with the highest standards of data protection and privacy.
Learn with Video Tutorials
Master advanced workflows like batch processing, referral management, and report customization with our comprehensive video library.
For Radiology & Sleep Medicine Professionals
Streamline your clinical workflow with secure patient data sharing and collaborative review
Upload Patient Studies
Upload polysomnography recordings individually or in batches for multiple patients
Automated Analysis
AI-powered sleep staging and disease risk assessment completed in minutes
Review & Annotate
Review detailed visualizations, compare historical studies, and generate insights
Secure Sharing
Generate time-limited secure links to share results with referring physicians or specialists
Batch Processing
Process up to 50 patient studies simultaneously with queue management and progress tracking
Longitudinal Analysis
Compare multiple studies over time to track disease progression and treatment efficacy
HIPAA-Ready Security
End-to-end encryption, secure storage, and time-limited sharing links for patient privacy
Professional Reports
Generate comprehensive PDF reports with sleep architecture analysis and risk assessments
Custom Templates
Create and share analysis templates for standardized protocols across your practice
Dr. Sarah Chen, Sleep Medicine Specialist, uses the platform to manage her sleep lab's workflow. She uploads overnight PSG studies for 15 patients each week using batch upload, reviews the AI-generated sleep staging and disease risk predictions, and generates secure sharing links for referring cardiologists and neurologists. The longitudinal comparison feature helps her track treatment outcomes for patients with sleep apnea, while the custom templates ensure consistent analysis protocols across her team.
Ready to Analyze Your Sleep Data?
Sign in to start using the SleepFM Inference Platform
Get Started Now