AI Models

Artificial Intelligence (AI) is transforming the way we understand the brain and treat neurological disorders. Our research innovation 'AI Models' brings together neuroscientists, clinicians, data scientists, and engineers to develop algorithms that can detect patterns in complex data, make accurate predictions, and support decision-making in healthcare. By turning vast and diverse datasets into actionable insights, AI enables earlier diagnoses, more personalised treatments, and better long-term outcomes.

Our researchers work on a close integration of AI developments with actual clinical and research settings. We combine large-scale data from medical imaging, physiological monitoring, behavioural assessments, and genetic profiles to create models that are both robust and clinically relevant. This approach turns data into clear, actionable insights, helping clinicians make faster, better-informed decisions and ultimately improving outcomes for patients.

AI is a powerful ally in advancing neurological health, capable of revealing insights that are impossible to detect with the human eye alone. By embedding these models in clinical workflows, we ensure that AI innovations lead to faster diagnoses, smarter treatments, and more personalised care.

Breakthroughs and impact

By combining information from a multitude of sources, we aim to develop models that are not only accurate in the lab but also practical in everyday clinical settings. Our research spans a wide range of applications, from identifying early signs of psychiatric disorders through speech analysis, to predicting epilepsy outcomes, to monitoring the wellbeing of newborns in intensive care. AI also powers innovative technologies such as brain-computer interfaces, enabling communication for people with severe impairments.

BCIs for effective communication
Brain-Computer Interfaces (BCIs) are promising solutions for people with motor impairments caused by for example amyotrophic lateral sclerosis, stroke, spinal cord injury and cerebral palsy. Brain signals can be recorded using implanted electrodes and translated into commands to control computers and other digital devices, enabling effective communication and access to the world. As such, this technology enhances independence, participation and quality of life of people with severe motor impairments.
AI to improve infant’s quality of life
The harnesses the power of AI to improve early prediction of long-term outcomes in newborns with brain injury. The model integrates clinical data, biomarkers, and advanced imaging and thereby supports personalized prognoses, early intervention, and better-informed care at a time when the developing brain is still highly responsive due to neuroplasticity. In close collaboration with parents, the study also explores how families wish to receive sensitive information about their child’s future. It empowers early action, communication and care, thereby maximizing each infant’s potential and improving quality of life.
Epilepsy prediction tools
Precision medicine in epilepsy focuses on individualized predictions of diagnosis and outcome. In this study, data from hundreds of patients with epilepsy, collected across multiple European centers, have been used to develop five prediction models. These models are freely available online and are applied in everyday clinical practice worldwide. With these, we can predict chances of epilepsy after a first seizure, risks of medication withdrawal, and cognitive outcomes after epilepsy surgery in individual patients.
Sleep Well Baby software
Sleep Well Baby is a bedside monitoring solution designed to assess sleep–wake states in preterm infants using physiological signals already captured in the neonatal intensive care unit (NICU). These are for example heart rate, respiratory rate, and oxygen saturation. Through extensive clinical and algorithmic validation, the system applies machine learning to provide real-time, non-invasive classification of active sleep, quiet sleep, and wakefulness. In parallel, the project examines links between sleep and brain development with the ultimate aim to prospectively evaluate the clinical impact of protecting sleep on neurodevelopmental outcomes.
“Talk to me and I’ll tell you how you’re doing”
Schizophrenia spectrum disorders, characterized by (auditory) hallucinations and cognitive decline, have been linked to changes in the brain’s language areas. Language, specifically (free) speech, data can be easily obtained in real-world settings, allowing clinical use for early detection in youth at risk or a relapse. Using this data, researchers were able to identify individuals with schizophrenia spectrum disorder by combining (large) language models and other machine learning techniques.
BASIC study
The BASIC project explores how university students develop the skills and awareness needed to work effectively and responsibly with generative AI. Using video, log, and survey data, the project identifies how students interact with AI in hybrid learning settings. These insights reveal how learners take ownership of AI tools, critically evaluate their use, and apply them meaningfully. Using data from two European higher education systems, its findings support responsible and sustainable AI use in education and professional life.