Scientists at the Max Planck Institute of Psychiatry in Munich, working with international colleagues, have developed an artificial intelligence algorithm capable of predicting a migraine attack the following day with up to 91 percent accuracy. The findings were published in the scientific journal Nature Medicine, according to infohub.kz.
The large-scale experiment involved 1,014 people suffering from chronic or frequent migraines. Over several months, volunteers wore wearable devices and regularly filled out special diaries in a mobile app. Digital sensors recorded sleep quality, physical activity levels, heart rate variability, heart rate and other physiological indicators.
The deep learning model was able to identify individual patterns in biometric data that precede the onset of headache. The main risk factors turned out to be sleep disturbances, sharp changes in stress levels and fluctuations in physical activity. The algorithm demonstrated high performance: sensitivity of 87 percent and accuracy of 91 percent.
The authors of the study, led by Dr. Florian Holsinger, note that early warning of migraine risk will allow patients to take preventive medication in time or adjust their daily routine to prevent or lessen an attack. In the near future, the researchers plan to begin creating a mobile app based on this algorithm for everyday use.