Scientists from several institutions, including Stanford University in America, have together created an Artificial Intelligence model (AI) named ‘Sleep FM’. According to research published in ‘Nature Medicine’, this model can predict the risk of 130 diseases based on a person’s one night’s sleep. The accuracy of the AI model’s prediction on the C-index has been greater than 0.75. In simple terms, it has been more than 75% accurate. This also includes the prediction of the risk of death from any cause.
What is Sleep FM?
Sleep FM predicts the risk of disease based on PSG (Polysomnography) data. It has been trained using PSG records of over 5.85 lakh hours from 65 thousand people. Their health records were taken from Stanford’s Sleep Clinic. According to researchers, the records contained over a thousand disease categories. Out of these, 130 were accurately predicted based on sleep data. Also read: Received an unknown parcel delivery call or message? The ‘Dial 21’ courier scam forwards your financial details to scammers
Model combines various types of data Emmanuel Mignot, Professor of Sleep Medicine and Senior Writer of Research in the Department of Psychiatry and Behavioral Sciences at Stanford University, said – A large number of signals are recorded in sleep studies. An individual who is completely in control is monitored for 8 hours. He said that the AI model understands by combining various types of data. It reads the hidden physiology of sleep and patterns changing over time based on brain activity (EEG/EOG), heart rate (ECG/EKG), muscle activity (EMG), and respiratory signals.
The post appeared first on .
