In the nationally funded EPICONN project, AIT brings its expertise in AI-supported EEG analysis to develop novel biomarkers for focal epilepsy. The aim is to identify, early in the course of treatment, which patients respond to antiepileptic drugs and who requires alternative therapeutic options. The Clinical Research Group, funded by the Ludwig Boltzmann Society, brings together neurology, neuroimaging, neurosurgery, genetics, and machine learning to optimise individual treatment pathways.
Epilepsy is one of the most common neurological disorders worldwide. In around one third of patients with focal epilepsy, antiepileptic drugs are not sufficiently effective—but this drug resistance often only becomes apparent after a protracted process of medication adjustment, which is extremely burdensome for patients and imposes high costs on the healthcare system. A reliable early detection method that predicts this from the outset of treatment could enable a much earlier switch to alternative therapy options.
“Currently, our understanding of the complex neural networks underlying epilepsy is still too limited,” explains brain-signal expert Gerhard Gritsch. “Our goal is to gain new insights and, using AI, identify subtle biomarkers to individualise and optimise treatment pathways at an early stage.”
Network Biomarkers and AI-EEG Analysis
EPICONN develops multimodal biomarkers based on imaging, EEG, and molecular data. Central to the project is the creation of a multimodal method that maps structural, functional, and metabolic changes and links them to genetic variants. AIT is responsible for developing and validating machine-learning algorithms for the automatic detection of EEG biomarkers based on high-frequency and infra-slow EEG oscillations recorded at the scalp.
- Multidisciplinary Study: In addition to existing data, 300 patients with focal epilepsy will be followed and examined prospectively over the full study period in the first four years.
- AI-Supported EEG Analysis: AIT’s Medical Signal Analysis unit will train neural networks on EEG data from the early treatment stage to predict resistance to drug therapies.
- Validation and Implementation: In the second project phase, the developed method will be validated in parallel at several Austrian epilepsy centres, and its integration into clinical practice will be advanced.
Here, AIT collaborates with the Departments of Neurology, Neurosurgery, Radiology, and Nuclear Medicine at the Medical University of Vienna. The project team comprises 11 internationally recognised researchers and brings expertise in clinical and research imaging, genetics and cell biology, as well as digital methods and deep learning. Together, they cover all areas of epilepsy research, making this research group ideally positioned to achieve the goals of EPICONN.
For the first time, this project creates an integrated analysis and prognosis toolkit to support treating physicians in choosing efficient, individualised therapy pathways. In the long term, the method is intended to be embedded in routine workflows and made available as a digital decision-support system in epilepsy centres.
Through the combination of innovative neuroimaging technology, molecular characterisation, and AI-based EEG evaluation, EPICONN provides a significant impetus for data-driven, patient-centred epilepsy care.
Link to the Ludwig Boltzmann Society press release “Patient-Oriented Health Research: LBG and BMFWF Present Three New, Outstanding Clinical Research Groups” (German Version): https://www.ots.at/presseaussendung/OTS_20250707_OTS0062/patientenorientierte-gesundheitsforschung-lbg-und-bmfwf-praesentieren-drei-neue-herausragende-klinische-forschungsgruppen