Abstract Introduction Sleep contains rich information about health status, relevant to future health outcomes including dementia, cerebrocardiovascular diseases, psychiatric diseases, and mortality. We hypothesized that the risk of these outcomes is predictable from quantitative analysis of sleep microstructure. Methods We included participants who underwent a diagnostic study and age older than 18 years. We excluded participants with missing demographics or PSGs < 2.5 hours in duration. We considered 11 outcomes including dementia, mild cognitive impairment or dementia, ischemic stroke, intracranial hemorrhage, atrial fibrillation, myocardial infarction, type 2 diabetes, hypertension, bipolar disorder, depression, and mortality. The outcomes were determined using ICD codes, brain imaging reports, medications, and/or cognition scores. We extracted 86 spectral and time-domains features from overnight sleep EEG recordings, including 57 features from NREM sleep epochs, 21 from REM, and 8 covariates including age, sex, body mass index, and medication prescriptions including benzodiazepines, antidepressants, sedatives, antiseizure medications, and stimulants. We modeled risk using Cox survival analysis with death as a competing risk. Model calibration was assessed using the difference in 10-year cumulative incidence (10y-CI) between Cox estimate vs. Aalon-Johansen estimate (ground truth). Results There were 8673 participants with an average age of 51 years; 51% were female. Participants were partitioned into three groups: poor sleep (hazard > 3rd quartile (Q3)), average sleep (Q1 ≤ hazard ≤ Q3), and good sleep (hazard < Q1). The model was able to predict the 10y-CI not significantly different from the ground truth, except for the risk of intracranial hemorrhage in the poor sleep group. The outcome-wise mean prediction difference in 10y-CI was 2.3% for the poor sleep group, 0.5% for the average sleep group, and 1.3% for the good sleep group. The outcomes with the top three poor-to-average risk ratios (RR) were dementia (RR = 6.2 95% confidence interval [4.5 – 9.3]), mortality (RR = 5.7 [5.0 – 7.5]), and MCI or dementia (RR = 4.0 [3.2 – 4.9]). Conclusion Sleep EEGs contain decodable information about the risk for future incidence of mortality, dementia, cerebrocardiovascular and psychiatric diseases. The findings strengthen the concept of sleep as a window into brain and general health. Support (If Any) This work is supported by the AASM Foundation 2019 Strategic Research Award.