Neurological movement disorders such as Parkinson’s, Alzheimer’s and Huntington’s disease are not always easy to distinguish for doctors. Giving the right diagnosis is therefore complex. By introducing the AI RONDO lab, several institutions such as Radboud University, het Radboudumc and partners like Info Support, aim to diagnose movement disorders better and quicker.
Info Support is involved in the AI RONDO project as business partner artificial intelligence. Joop Snijder, head of the research center at Info Support: “We will focus on building AI models that recognize movement disorders based on video footage of the patient. Signs of certain movement disorders are for example shaking of the hands and decreased articulation. We can train an algorithm in such a manner that even subtle symptoms are detected on video. An algorithm is often better capable of recognizing relationships and patterns in those different symptoms than humans are. It could very well even be that these self-learning AI algorithms discover new relationships and patterns for neurological movement disorders.”
AI models can not only contribute to diagnosing movement disorders, but also offer help for patients that are already diagnosed. Marjan Meinders, one of three academic directors at the lab: “There is a lot of data on patients that suffer from Parkinson’s and Alzheimer’s available already, and by applying AI algorithms and models to this information, we can find new relationships, and for example designate groups that have a higher risk of developing complications. Using this enriched data enables us post diagnosis to prescribe the patient a treatment that is specifically tailored to his personal risk profile. We can also analyze signs – such as speaking softer, articulating less clearly, a change in gait or heartbeat – that indicate something is going wrong.”
Info Support starts a preliminary phase in June, starting with analyzing videos and researching possible classifications. Joop Snijder, Info Support: “We are very pleased to be part of this project. AI algorithms are more often used for diagnosing diseases, so that is not new. There are smart AI algorithms that can recognize cancer on lung x rays for example, but there are not too many AI models in health care yet that are able to diagnose based on video footage. This project brings health care another step forward. This is also a great challenge for us. It is a complex 4-to-5-year research trajectory, that will teach us incredibly much.”