Vidi grants for UM research on AI models and our brains
Two Maastricht University (UM) scientists will receive a €850,000 Vidi grant for their research projects. This was announced by the Dutch Research Council (NWO) on Thursday. Researcher Laure Wynants (CAPHRI/ FHML) will work with her team on more reliable AI models for healthcare. And neuroscientist Michelle Moerel will study at FSE and FPN how our brain filters sound.
Michelle Moerel (FPN/FSE)
In noisy environments, we effortlessly focus on interesting sounds while ignoring the rest. This remarkable ability does not stem from our ears, but from sound processing in the brain. Using MRI and computational models, researchers will investigate how the brain accomplishes this. Findings are relevant for the 10% of the population struggling with tuning into sound.
“As an auditory neuroscientist, I am intrigued by how the brain rapidly adapts to optimally process the sounds around us. Small brain regions hidden deep in the brain likely play a crucial role in this auditory flexibility, but studying these brain regions in humans is challenging because they are so small. With the support of the VIDI grant, I aim to overcome these challenges and investigate how auditory flexibility works in individuals with healthy hearing. Ultimately, I hope to shed light on what goes wrong in those who struggle to hear in noisy environments.”
Laure Wynants (CAPHRI/ FHML)
Artificial intelligence (AI) can support diagnostics and prognostics, optimising patient care. Whether it succeeds depends on the reliability of an AI model. Often, AI testing is done on small numbers, and AI does not prove equally suitable for all sites and populations. Reliability can be made measurable with value-of-information measures. These reflect the risks of uncertainty, such as the expected number of misdiagnoses. In this project, researchers are developing these value-of-information measures. This will allow doctors and policymakers to decide for each AI model: is it ready for use, or is more research needed?
‘In my research so far, I have mainly been developing new prediction models for diagnostic and prognostic applications, and assessing existing models. In doing so, I have found that with the current assessment criteria, it remains very unclear when a model is now good enough to support clinical decisions in practice. With the VIDI grant, I can establish a team that will develop better measures to assess: is a model ready for use, or is more research needed? With this, I ultimately want to contribute to the optimal integration of AI in healthcare.’