新型人工智能模型可通过睡眠数据预测疾病

Core Insights - An AI model named SleepFM has been developed by researchers at Stanford University to predict the risk of approximately 130 diseases based on a single night's sleep data, including heart disease, dementia, and certain cancers [1][2] - The model was trained using data from 65,000 participants, totaling nearly 600,000 hours of polysomnography data, marking the first use of AI to analyze such a large dataset of sleep information [1] Group 1 - The SleepFM model utilizes polysomnography, which records various physiological signals related to sleep, making it the "gold standard" for sleep assessment [1] - The training data includes records from 35,000 individuals at the Stanford Sleep Center, who have had their sleep and health monitored over 25 years [2] - The model has shown exceptional predictive capabilities for diseases such as Parkinson's, dementia, developmental delays, and cardiovascular diseases, with high accuracy in predicting prostate, breast, and skin cancers [2] Group 2 - Previous research on sleep and disease often focused on single indicators and specific diseases, neglecting the complexity of sleep physiology [2] - The new findings indicate that AI models can understand the "language" of sleep data, enabling flexible and efficient disease prediction [2]