A groundbreaking smartphone tool could revolutionise the way paramedics identify strokes, offering a much quicker and more precise method than current technologies. Strokes, which impact millions globally, result from an interrupted or reduced blood supply to part of the brain, depriving brain tissue of oxygen and nutrients. Prompt identification is crucial, as even a few minutes’ delay can cause irreversible damage to brain cells.
The AI technology behind this innovative tool was developed by a team of biomedical engineers at RMIT University. Their findings have been published in the journal Computer Methods and Programs in Biomedicine.
Led by PhD scholar Guilherme Camargo de Oliveira from RMIT and São Paulo State University, and supervised by Professor Dinesh Kumar, the team has created a face-screening tool that can detect strokes with 82% accuracy. Kumar, from RMIT’s School of Engineering, emphasised the importance of early detection. “Timely treatment can vastly improve recovery outcomes, reduce the risk of long-term disability, and save lives,” he said.
The smartphone tool allows paramedics to quickly determine if a patient has suffered a stroke and to alert the hospital even before the ambulance departs. While it won’t replace comprehensive clinical diagnostics, it offers a significant advantage by identifying those in need of urgent care much sooner.
Symptoms of a stroke, such as confusion, movement control loss, speech impairments, and diminished facial expressions, can be subtle and easily missed. Studies have shown that nearly 13% of strokes go undetected in emergency departments, and 65% of patients without a documented neurological examination remain undiagnosed. This rate is even higher in regional centres where initial care is often provided by first responders under non-ideal conditions.
The AI-driven technology leverages facial expression recognition to detect strokes by analysing facial symmetry and specific muscle movements. Using the Facial Action Coding System (FACS), the tool examines facial movements by the contraction or relaxation of muscles, providing a detailed analysis of expressions. “A key indicator for stroke patients is that their facial muscles typically become unilateral, causing one side of the face to behave differently,” explained de Oliveira.
The study involved video recordings of facial expressions from 14 post-stroke individuals and 11 healthy controls. The AI tools and image processing algorithms identified changes in smile symmetry, which were crucial for detection.
Looking ahead, the team aims to develop this technology into a user-friendly app in collaboration with healthcare providers. This app could potentially detect other neurological conditions that affect facial expressions. “We strive for the highest sensitivity and specificity,” Kumar said. “Our next step is to enhance the AI tool with additional data and to consider other diseases. Collaboration with healthcare providers will be vital to integrate this app into emergency response protocols.”
This research was a collaborative effort between RMIT University and São Paulo State University in Brazil, showcasing the global importance and potential impact of this innovative stroke detection tool.
