Strokes, affecting millions globally, occur when the brain's blood supply is compromised, depriving brain tissue of essential oxygen and nutrients. Common symptoms include confusion, loss of movement control, speech impairments, and reduced facial expressions. Early stroke detection is crucial; prompt treatment significantly improves recovery, reduces the risk of permanent disability, and saves lives. Yet, around 13% of strokes go undetected in emergency departments and community hospitals, and 65% of patients lacking a documented neurological exam are not diagnosed with stroke. The subtlety of symptoms and potential biases against different races or genders further complicate timely recognition. Now, an innovative smartphone-based facial screening tool promises to help paramedics identify stroke signs much sooner and more accurately than current technologies.

Developed by biomedical engineers at RMIT University (Melbourne, Australia), this technology leverages artificial intelligence (AI) to analyze facial symmetry and muscle movements, enhancing stroke detection through facial expression recognition. Stroke often causes unilateral facial muscle impairment, making one side of the face act differently than the other. As reported in a study published in Computer Methods and Programs in Biomedicine, the research utilized video analyses of facial expressions from 14 post-stroke individuals and 11 healthy controls. Using the Facial Action Coding System (FACS) from the 1970s, which categorizes facial movements by muscle contractions or relaxations, the AI assesses changes in facial expression symmetry, particularly in smiles. The

Paramedics can use this simple smartphone tool to quickly assess if a patient has had a stroke, potentially notifying hospitals before the ambulance leaves the patient’s house. While this tool, with an 82% accuracy rate in stroke detection, is unlikely to replace detailed clinical diagnostics, it could significantly speed up the identification of patients requiring urgent care. The researchers are planning to expand this technology into an app, in collaboration with healthcare providers, to detect other neurological conditions affecting facial expressions. They are also aiming to enhance the AI's capabilities by integrating more data and extending its diagnostic reach to other diseases.

“We have developed a simple smartphone tool that paramedics can use to instantly determine whether a patient is post-stroke and then inform the hospital before the ambulance leaves the patient’s house,” said team leader Professor Dinesh Kumar from RMIT’s School of Engineering. “Our face-screening tool has a success rate for detecting stroke that compares favorably to paramedics.” 

http://www.gzjiayumed.com/en/index.asp .