An application developed by experts from the Faculty of Applied Sciences at UWB will help doctors detect early-stage dementia in patients. The introduction of the application into practice will provide patients with more accessible medical care while reducing the costs of treatment.
An application that helps users recognize the symptoms of early neurodegenerative diseases such as dementia. All from the comfort of home, without the need to visit a doctor. Experts at the Faculty of Applied Sciences of the University of West Bohemia in Pilsen (FAV ZČU) are currently developing just such a tool.
“What is new in this area is that a wide range of the Czech population will be able to examine themselves at home. Additionally, the examination results will be evaluated completely automatically. Early detection of cognitive disorders can then be followed by more detailed diagnosis and appropriate treatment,” said Luboš Šmídl from the Department of Cybernetics at FAV UWB.
The set of tests in the application gradually examines various areas of cognitive function, including
vision, hearing, comprehension, basic attention, short-term memory, visual recognition, and the ability to name objects.
“A key part is a detailed description of a color picture of a scene by a lake, which assesses narrative organization, language fluency, the ability to maintain a coherent storyline, visual-spatial perception, and attention,” explained Šmídl. The application thus comprehensively maps basic sensory functions such as memory, language skills, executive functions, and speed of information retrieval—key areas that are often impaired in dementia.
The Faculty Hospital Královské Vinohrady and the Institute of Physics of the Czech Academy of Sciences are also participating in the project entitled Telemedical Self-Examination of Speech and Memory for Rapid Detection of Cognitive Disorders Using Machine Learning Methods.
"The collaboration between experts from the medical and technical faculties has resulted in a unique application that can mimic an examination similar to that performed by a doctor. The patient does not communicate with a person, but with a computer that gives them verbal and written instructions. The answers are recorded and evaluated by artificial intelligence. This opens up a completely new way of examination that is in line with global trends and can be of significant help at a time when the population is aging and medical staff are dwindling. We believe that with this project, we are not only keeping up with the times, but even ahead of them," said Dr. Aleš Bartoš from the Memory Disorders Clinic at the Neurology Clinic of the Královské Vinohrady University Hospital and the Third Faculty of Medicine of Charles University in Prague, who is also a co-researcher on the project.
The fact that the use of artificial intelligence in healthcare is a hot topic is also evidenced by the awards received by experts from FAV at the 28th annual international conference Text, Speech and Dialogue (TSD 2025). Jan Tupý won the Best Poster Award, and Tomáš Lebeda received the Runner-up Poster Award. The conference is one of the 30 most important conferences in the world in the field of computational linguistics and natural language processing. This year, it was held in Erlangen, Germany, and representatives from the Department of Informatics and Computer Science at FAV were members of the international organizing team.
Doctoral students Jan Tupý and Tomáš Lebeda from the Department of Cybernetics and the NTIS Research Center at the Faculty of Applied Sciences, University of West Bohemia, are part of the team, and their approach combines automatic speech recognition, semantic analysis, and voice dialogue systems, paving the way for more effective and accessible screening methods in healthcare.
“I started working on this topic as part of my thesis. I found it interesting from a technical point of view and useful for real-world applications. And since then, I have actually stayed with this topic as part of my current doctoral studies,” said Tomáš Lebeda.
This project is co-financed from the state budget by the Technology agency of the Czech Republic under the
“Telemedical self-examination of speech and memory for rapid detection of cognitive disorders using machine learning methods”
Progamme.
