Una Europa Epilepsy Data Challenge

Epilepsy is one of the most common neurological disorders. Treatment is possible, but despite anti-seizure medication, around 30 percent of patients are not seizure-free. To improve therapeutic decisions, it is crucial to accurately detect and log seizures. However, seizure diaries have proven unreliable in clinical practice and longer-term full-scalp-EEG recordings are impossible in a home setting. The solution: Wearable devices that record EEG with behind-the-ear electrodes.

In this challenge, participants are asked to train Machine Learning (ML) models to accurately detect seizure events in data obtained from more than 200 epilepsy patients using a wearable device. The challenge is based on a unique dataset obtained from wearable devices used by patients of KU Leuven university hospitals.

The challenge will kick off on 3 February with an online workshop, followed by release of the dataset to participants. Throughout the challenge period, regular online Q&A sessions will be organised to provide guidance and foster exchange among participants. In April, participants and Una Europa experts from the fields of neuroscience, machine learning, signal processing and cognitive science will come together in a joint workshop, held in hybrid format. Here, participants will have the chance to present their findings and discuss their results with experts.

High-scoring participants within the first six weeks will be eligible to win a ‘wild card’ to be funded to travel to Leuven for the workshop. Results may be presented internationally at the 2025 EMBC Conference in Copenhagen in July.

The challenge is open to doctoral researchers and master’s students enrolled at an Una Europa partner university. We welcome applications from candidates working in the field of biomedical engineering, software engineering, electrical engineering, IT for health, or equivalent.

Participants will need to have basic programming skills; experience with biomedical signals is an advantage. You will be expected to submit a pre-trained model, either in Python or Matlab.

As participant, you will tackle a pressing medical challenge in an international, collaborative environment, working alongside leading experts in neuroscience, machine learning, signal processing, and cognitive science from across Europe. This challenge offers a rare opportunity to work on a high-quality, real-world dataset, enabling you to make an impact in reducing risks for patients.

As participant, you will be at the forefront of advancing neurological research and technology by contributing to the standardization of EEG data collection and analysis techniques as well as improved algorithms for brain activity detection and data interpretation.

Qualified doctoral researchers and master’s students can apply individually or as a team of up to three students by filling in this application form by 20 January.

If you are interested in teaming up with someone but unable to build a team yourself, you can indicate your wish to be matched in the registration form.

Photo credit: Laboratorium voor Epilepsie Onderzoek UZ KU Leuven