Working Group "Personalized Digital Health and Telemedicine"
SecureNeuroAI
SecureNeuroAI: Advanced IT Security for Medical Seizure Detection
The SecureNeuroAI project focuses on developing innovative, IT-secure technologies for the real-time detection of epileptic seizures. The goal is to protect AI-based medical devices from cyberattacks, ensuring the safety of patients and the reliability of medical care.
By integrating multimodal data—including EEG, heart rate, and respiratory rate—the project develops AI models capable of not only accurately detecting seizures but also identifying and mitigating manipulation attempts. Alongside technical innovation, the project emphasizes legal and ethical considerations to safeguard patient data and ensure compliance with regulatory requirements.
The role of the Krüger group at UKB
The Personalized Digital Health and Telemedicine Group, plays a central role in the project. The group’s contributions include:
Data Collection and Analysis: Systematic acquisition of multimodal data in clinical settings, including EEG and vital signs, to create realistic foundations for developing and validating AI models.
Development and Integration: Adapting AI models to clinical requirements and integrating them into the IT infrastructure of everyday clinical practice.
Pilot Studies: Testing and evaluation of the developed technologies in both clinical and home environments.
Interdisciplinary Collaboration: Close collaboration with medical personnel and technical experts to ensure user-friendly and practical solutions.
Collaborations
The project is a collaborative effort with the Computer Science Department at the University of Bonn, involving two leading research groups:
The IT Security Group, led by Prof. Dr. Michael Meier, contributing expertise in cybersecurity, risk assessment, and attack detection throughout the AI lifecycle.
The Data Science and Intelligent Systems Group, led by Prof. Dr. Elena Demidova, specializing in the development of robust and explainable AI models for multimodal data analysis.
With SecureNeuroAI, we make a significant contribution to improving healthcare and enhancing the resilience of modern medical technologies against cyber threats.
Funded by:
Publications:
2025
Greß, Hannah; Demidova, Elena; Meier, Michael; Krüger, Björn
In: Ohm, Marc (Ed.): Proceedings of the 15th graduate workshop of the special interest group Security - Intrusion Detection and Response (SIDAR) of the German Informatics Society (GI) (SPRING 2025), pp. 22-24, GI SIG SIDAR, Nuremberg, April, 2025, ISSN: 2190-846X.
@inproceedings{Greß2025,
title = {SecureNeuroAI: Advanced Security Framework for AI-Powered Multimodal Real-Time Detection of Medical Seizure Events},
author = {Hannah Greß and Elena Demidova and Michael Meier and Björn Krüger},
editor = {Marc Ohm},
url = {https://fg-sidar.gi.de/publikationen/sidar-reports},
issn = {2190-846X},
year = {2025},
date = {2025-05-12},
urldate = {2025-05-12},
booktitle = { Proceedings of the 15th graduate workshop of the special interest group Security - Intrusion Detection and Response (SIDAR) of the German Informatics Society (GI) (SPRING 2025)},
pages = {22-24},
publisher = {GI SIG SIDAR},
address = {Nuremberg, April},
abstract = {In today's interconnected world, medical devices are increasingly equipped with novel digital technologies and AI-powered methods to improve the users' quality of life.
Despite the increased possibilities and features these devices offer due to the technical progress, cyberattacks on medical devices will increase as well with possibly severe outcomes for the patients.
At the same time, AI-based technologies could help to detect and mitigate these attacks on medical systems and their data in real-time.
Therefore, our project "SecureNeuroAI" aims to detect epileptic seizures using multimodal sensor data and AI models while also considering possible cyberattacks on this system resulting in an IT-secure system.
Our results will serve as an example for future AI-supported medical devices and systems to enhance their security and to strengthen their trustworthiness towards their (future) users.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In today's interconnected world, medical devices are increasingly equipped with novel digital technologies and AI-powered methods to improve the users' quality of life.
Despite the increased possibilities and features these devices offer due to the technical progress, cyberattacks on medical devices will increase as well with possibly severe outcomes for the patients.
At the same time, AI-based technologies could help to detect and mitigate these attacks on medical systems and their data in real-time.
Therefore, our project "SecureNeuroAI" aims to detect epileptic seizures using multimodal sensor data and AI models while also considering possible cyberattacks on this system resulting in an IT-secure system.
Our results will serve as an example for future AI-supported medical devices and systems to enhance their security and to strengthen their trustworthiness towards their (future) users.
