How AI-powered software supports early detection of dementia in clinical practice

Detecting dementia early – with AI and MRI analysis: An innovative demonstrator supports medical professionals in diagnosing Alzheimer’s disease.

Neurodegenerative diseases such as Alzheimer’s are among the most common conditions in old age – around 1.6 million people in Germany alone are affected. These diseases lead to characteristic changes in the brain, which can be visualized using magnetic resonance imaging (MRI). However, identifying such patterns requires a great deal of experience, time, and a strong 3D imagination.

This is where the research by Prof. Stefan Teipel and Dr. Martin Dyrba at the University Medicine Rostock, in collaboration with the German Center for Neurodegenerative Diseases (DZNE), comes in. Dr. Dyrba has developed AI-based software that automatically analyzes MRI data and highlights relevant brain areas in color. This makes it much easier for medical professionals to interpret the scans.

The software uses a neural network architecture that detects brain regions relevant to Alzheimer’s and visualizes them. The results are displayed in an intuitive interface with relevance maps that indicate which brain areas contribute most to the AI model’s decision.

Although currently intended for research purposes only, the system offers great potential for future clinical applications, especially in the early detection of dementia.

A publicly accessible demonstrator allows users to test the technology with sample data. With proper preparation, even personal MRI scans can be uploaded and analyzed.

The goal is to make complex MRI data more transparent and understandable through artificial intelligence. This not only supports more accurate diagnoses but also lays the groundwork for new therapeutic approaches.