While image analysis and machine learning research domains are developing rapidly, objective diagnosis and prognosis in neuropsychiatry remain challenging. MRI brain imaging is improving in terms of quality and speed, which is promising for handling this challenge. Apart from the novel acquisition techniques, innovations are needed in the analysis techniques for brian images. To this end, we perform research on so called neurodynamics: we explore the resting state of the brain in order to detect and describe dynamic patterns in functional network interactions. While structural or functional changes can be detected when the brain is obviously affected by a disease, the neurodynamics analysis offers an approach to describe less evident but essential changes in the brain which affect the behaviour and cognition of a patient.

Our technology, combined with machine learning, leads to creation of explainable AI systems which allow diagnosis of, for example, autism or accelerated cognitive ageing. Exploration of disease prognosis options, based on interpretable features, is the next step that we are taking in the ongoing research.

This research is performed in close collaboration with Philips, Kempenhaeghe, GGZ/e, and other partners.