I-Corps: Improving Detection of Stroke using Artificial Intelligence

University Of Virginia Main Campus, Charlottesville VA

Investigators

Abstract

The broader impact/commercial potential of this I-Corps project results from enabling first responders and emergency providers to have on-demand access to expert neurological knowledge when assessing stroke patients. Missed strokes in the emergency setting can result in patients suffering from increased long-term disability and cost the healthcare systems millions of dollars in medical expenses and lost productivity. By using artificial intelligence and mobile devices to bring the experience and knowledge of vascular neurologists closer to the patient at the time of their stroke, this technology can help reduce the number of strokes missed. In particular, rural and low population density areas can benefit from increased access to expert neurological knowledge that may not be readily or reliably accessible in their communities. Additionally, the increasingly aging population is at an overall higher risk for stroke further increasing the demand and potential impact of this technology in preventing the long-term consequences of stroke on patients and the healthcare system. The underlying technology can also be expanded to include assessment of other neurological disease such as multiple sclerosis, headaches and Parkinson's disease thereby opening up new market opportunities in the the diagnosis and monitoring of these conditions. This I-Corps project is based on technology that combines knowledge from the fields of neurology and computer science to create a set of algorithms that will accurately assess patients for signs of stroke. The algorithms use computer vision techniques to analyze videos of patients undergoing a stroke assessment. Machine learning classification techniques are then used to decide if the video is likely showing a positive or negative stroke assessment. The algorithms are trained and tested with data validated by vascular neurologists thereby learning directly from human experts. Proof-of-concept has been demonstrated for algorithms detecting individual signs of stroke but development of a comprehensive stroke assessment algorithm is underway. Further development of the technology for specific applications requires a thorough understanding of customer and key stakeholder needs and will be focus of this I-Corps Team project. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.