The George B. Moody PhysioNet Challenge 2026 invites teams to develop algorithmic approaches for using polysomnography (PSG), which records various physiological signals during sleep studies, to predict future diagnoses of cognitive impairment.
Sleep disturbances are associated with a higher risk of developing mild cognitive impairment (MCI) and other forms of cognitive impairment, and poor sleep can double or triple the odds of developing MCI. While MCI can cause or exacerbate sleep disruption, poor sleep also accelerates the accumulation of amyloid-beta plaques in the brain, creating a cycle that promotes cognitive decline.
The 2026 Challenge data are from the Human Sleep Project database and include clinical PSG data from five different U.S. institutions: Beth Israel Deaconess Medical Center, Emory University, Kaiser Permanente, Massachusetts General Brigham, and Stanford University. Each study includes synchronized physiological signals and human and automated annotations of sleep stages, arousals, and respiratory and limb movement events. The Complete AI Sleep Report (CAISR), a comprehensive automated AI framework designed for high-fidelity sleep analysis, provided sleep stages and clinical event markers for each PSG. All data were de-identified in accordance with the HIPAA Safe Harbor method.
Each team was asked to develop and implement an open-source algorithm that, given a patient's PSG, automated CAISR annotations, and basic demographic data, predicts whether or not the patient will receive a cognitive impairment diagnosis within 3 to 7 years after the PSG. An automated scoring code generated an area under the receiver operating characteristic curve (AUROC) and provided up to 10 scores on the validation data, which were posted on the public leaderboard. The teams will be asked to select their preferred algorithm for final scoring on the hidden test data, and winners will be announced at CinC 2026.