US researchers develop AI tool for early stroke detection during sleep

American scientists have developed an AI-based model that assesses stroke and heart disease risk using data from sleep ECGs, even with just one chest electrode. The research involved training the software on data from approximately 16,000 individuals who underwent sleep ECG monitoring, subsequently testing its efficacy on two groups totaling around 23,000 patients.

The AI algorithm detected risk markers through alterations in heart rhythm during sleep. The findings indicate a heightened risk of stroke by 19 to 40 percent for each standard deviation increase in the risk score, a 69 to 133 percent rise in heart failure risk, and a 103 to 172 percent surge in atrial fibrillation risk. The study also highlights the importance of early sleep stages, as rapid nervous system changes during this phase may facilitate the detection of concealed heart rhythm disorders.

Researchers are optimistic that this approach could facilitate the early identification of individuals at increased risk of stroke and other cardiovascular diseases, potentially years before symptoms manifest.