Montpellier: AI Stroke start-up accelerates stroke detection

Le fondateur Cédric Javault lors de la remise de son prix pour le plan France 2030. DR – DR

Based and created in Montpellier, the AI ​​Stroke start-up is developing an application based on artificial intelligence, which analyzes the gestures and facial expressions of a patient in the process of starting a stroke. The software's goal is to speed up the management of these cases by emergency caregivers.

Based in Montpellier, the young company AI Stroke focuses on a major public health challenge: reducing the time taken to treat strokes. The second leading cause of death worldwide, strokes require rapid intervention to avoid serious or even fatal after-effects. The start-up has developed a “digital neurologist”, a mobile application designed to help first responders detect the signs of a stroke more quickly.

The start-up AI Stroke is developing an innovative digital solution for early detection of strokes. Its mobile application, using artificial intelligence (AI), aims to reduce diagnostic times and improve patient care by guiding first responders in quickly identifying symptoms.

200% Deposit Bonus up to €3,000 180% First Deposit Bonus up to $20,000

“Our application guides the patient in performing simple neurological tests, records videos and makes a prediction”explains Cédric Javault, co-founder and director of AI Stroke. The tool will film the patient performing simple gestures such as “making a big smile” or “holding your arms out in front”, standard tests for detecting the first signs of a stroke. Artificial intelligence will then analyze these videos to establish a preliminary diagnosis.

Example of analysis of a patient's face by the application of the start-up AI Stroke. DR – DR

A crucial public health issue

“Today in France, we have one in three patients who are not immediately understood as having a stroke by first responders” deplores the manager. The company aims to “divide this error rate by two” thanks to its application, hoping to thus “gain up to 1h30 on average per patient”. A crucial time frame, as emergency treatments must be provided within a 4 to 5 hour window after the first symptoms.

To develop this innovative technology, the start-up was able to count on public subsidies, notably from the France 2030 investment plan. “We are one of the companies that would not be here if it were not France 2030” acknowledges Javault. AI Stroke also collaborates with several cutting-edge institutions such as the Nîmes University Hospital, a CNRS laboratory and the University of Florida.

Development and perspectives

“To train an AI, we need to provide it with… a large number of videos of stroke patients” explains the manager. A long-term project carried out in partnership with these medical institutions. While the first tests are currently taking place with the Hérault firefighters, the company hopes to launch it on the market in 2026, after the regulatory steps for a medical device.

“For health, as a diagnostic aid, it's much less scary than other AI applications” believes the founder, seeing this tool as a valuable aid for healthcare workers, without the aim of replacing them. “We're going to save time and increase efficiency, in a context where there is a shortage of caregivers.”

I subscribe to read the rest