McMaster Researchers Use AI to Detect Long-Term Concussion Signs
Researchers from McMaster University, in collaboration with the Vector Institute for Artificial Intelligence, have developed a machine learning algorithm capable of detecting signs of concussions in retired athletes decades after the initial injury. Traditionally viewed as short-term issues, concussions are now recognized as chronic health problems with lasting effects on brain electrical signals. The new study, published in IEEE, utilizes data from retired Canadian Football League players to identify individual cases with an 81% accuracy rate, surpassing current clinical tools. Unlike typical black-box AI models, this explainable AI approach identifies specific brain responses responsible for the diagnosis, validating the method and uncovering previously undocumented signs of concussion. This breakthrough offers significant potential for helping individuals who were previously misdiagnosed or unaware of the severity of their long-term brain injuries, marking a shift from group-level statistics to precise single-subject prediction in neurological health assessment.
Wire timeline
McMaster Researchers Use AI to Detect Long-Term Concussion Signs
Researchers from McMaster University, in collaboration with the Vector Institute for Artificial Intelligence, have developed a machine learning algorithm capable of detecting signs of concussions in retired athletes decades after the initial injury. Traditionally viewed as short-term issues, concussions are now recognized as chronic health problems with lasting effects on brain electrical signals. The new study, published in IEEE, utilizes data from retired Canadian Football League players to identify individual cases with an 81% accuracy rate, surpassing current clinical tools. Unlike typical black-box AI models, this explainable AI approach identifies specific brain responses responsible for the diagnosis, validating the method and uncovering previously undocumented signs of concussion. This breakthrough offers significant potential for helping individuals who were previously misdiagnosed or unaware of the severity of their long-term brain injuries, marking a shift from group-level statistics to precise single-subject prediction in neurological health assessment.
Vector Institute for Artificial Intelligence