Researchers from the Korea Advanced Institute of Science and Technology (KAIST) have developed an artificial intelligence system that may help detect the early warning signs of cerebrovascular disease before serious symptoms appear.
The study, published in the journal npj Digital Medicine, suggests that small changes in an older person’s daily routine, sleep and home environment could reveal that their risk is increasing long before they visit a hospital.
Cerebrovascular disease affects the blood vessels that supply the brain.
It includes conditions such as stroke, which is one of the leading causes of death and long-term disability worldwide.
When blood flow to the brain is blocked or reduced, brain cells can quickly become damaged. Fast treatment is extremely important, but many people do not realise that warning signs have already started to appear because these early changes are often very subtle.
Hospital tests usually take place only after symptoms become obvious, meaning valuable time may already have been lost.
To explore whether these hidden warning signs could be detected earlier, Professor Lisa Lim and colleagues worked with researchers from Sungkyunkwan University and Korea University Anam Hospital. Instead of relying only on medical scans or blood tests, they focused on how older adults live in their own homes.
The team analysed information from 1,224 older adults collected by LivOn Care. They examined 13,362 two-week lifelog records gathered in real homes.
These records included daily movement, sleeping habits, activity patterns across the day, indoor humidity and other environmental measurements. Personal information such as age and existing chronic diseases was also included to improve the AI model.
The researchers trained AI to recognise patterns linked with the early stage of cerebrovascular disease. They found that older adults entering this stage often stayed active between 10 p.m. and 2 a.m., when most people should normally be sleeping. This suggested disrupted body clocks and delayed sleep schedules.
As diagnosis became closer, evening activity between 6 p.m. and 10 p.m. became lower while inactive time increased. Dry indoor air, reflected by low humidity, also appeared to be an important clue.
One of the strongest findings involved timing. The researchers compared data collected within four weeks before diagnosis with information collected about 12 weeks earlier. The AI distinguished these two periods with an accuracy of 96.53 percent, suggesting that lifestyle changes may become increasingly noticeable as diagnosis approaches.
Unlike many AI systems, this one also explained why it reached its conclusions. This allowed researchers to identify which behaviours and environmental conditions contributed most to the estimated risk instead of simply producing a prediction.
The researchers believe this technology could become a useful digital health tool for older adults, especially those who may struggle to describe changes in their health. Caregivers and doctors could receive earlier warnings and encourage medical assessment before major symptoms develop.
The findings are encouraging because they show that ordinary daily behaviour may provide valuable health information that cannot always be captured during a short medical appointment.
However, the study does not predict exactly when cerebrovascular disease will occur and cannot replace a doctor’s diagnosis. Larger prospective studies are still needed before the system can be widely used in healthcare. Even so, the work represents an important step toward preventing disease instead of only treating it after serious damage has already happened.
If you care about stroke, please read studies about how to eat to prevent stroke, and diets high in flavonoids could help reduce stroke risk.
For more health information, please see recent studies about how Mediterranean diet could protect your brain health, and wild blueberries can benefit your heart and brain.
