AI-STROKE

A digital neurologist for detecting strokes

The problem

Stroke is the 2nd leading cause of death and disability worldwide, and 25% of us will have at least 1 stroke in our lifetime. However, globally, fewer than 5% of acute ischemic stroke patients receive effective treatment within the critical 4-6 hour window for clinical intervention.

The clinical signs of a stroke vary greatly from patient to patient as they depend on the specific area of the brain that is affected. Very often, the patient, relatives, and even emergency professionals do not understand that a stroke has occurred. What made things worse is that the current pathway for stroke patients is very complicated, and often requires patients to be transferred through different health providers, wasting critical time. While we try to make the right diagnosis, patients lose up to 2 million brain cells each minute and their ischemic brain ages 3.6 years for each hour of delayed stroke care.

The key problem is that stroke diagnosis is complicated, but EMS first responders have no available tool, causing significant loss.

Our solution

AI-Stroke has created the world’s largest dataset of stroke patient videos, with over 23,000 videos and 4.5 million images, to train its AI. Embedded in a tablet or smartphone, it is designed for EMS first responders (e.g., ambulances, 911 operators & emergency rooms) to identify strokes and large vessel occlusions in seconds. Patients perform exercises while being filmed, and the AI gives stroke diagnostics, specifies its kind, and selects the best hospital for treatment, significantly shortening the time to treatment and saving the lives of stroke patients.