AI specialists aim to safeguard medical technology from digital assaults
The SecureNeuroAI project, a collaboration between the University of Bonn, University Hospital Bonn, and FIZ Karlsruhe, is working to protect AI-based medical devices and applications from cyberattacks. Particularly, the project aims to safeguard devices that detect epileptic seizures in real-time.
Led by Professor Elena Demidova from the Data Science and Intelligent Systems working group at the University of Bonn, the project focuses on creating data authentication methods that verify the origin of information without interfering with AI processing. This ensures the integrity, availability, and reliability of AI-enabled medical systems.
The project involves a comprehensive analysis and logging of multimodal sensor data, including vital signs, to detect medical emergencies such as epileptic seizures securely. AI-powered real-time detection techniques combined with cyber-secure data handling are used to prevent tampering or data breaches that could jeopardize patient safety or the functionality of medical equipment.
Professor Michael Meier from the University of Bonn brings expertise in current IT security topics, focusing on potential vulnerabilities in networked medical devices and infrastructures. Professor Björn Krüger from UKB emphasises the importance of secure system thinking in healthcare, particularly with sensitive patient data. The research department Immaterial Property Rights (IGR) at FIZ Karlsruhe, led by Professor Franziska Boehm, analyses data protection and IT regulations, as well as legal issues of artificial intelligence.
Elena Demidova's team plays a key role in developing AI methods for data authentication and designing explainable AI models for detecting manipulations and seizure events. The Clinic and Polyclinic for Epileptology at UKB systematically collects and prepares multimodal data for seizure detection under clinical conditions.
The project's AI models are designed to reliably detect seizures and identify potential data manipulation. Beyond the detection of epileptic seizures, the project's overarching goal is to create a technological basis that improves the integrity, availability, and reliability of AI-based medical devices. This could find application beyond seizure detection, with potential uses in early detection of medical emergencies by speech and AI tests for Parkinson's in seconds, as well as opportunities for AI in stroke treatment and job openings in software development.
Funded with nearly €2.5 million by the German Ministry for Education and Research (BMFTR) over three years, the SecureNeuroAI project aims to significantly enhance the cybersecurity of AI-driven medical devices to improve patient safety and emergency response effectiveness without compromising AI data processing performance.
[1] BMFTR Funding Announcement [4] BMFTR Project Information
- The SecureNeuroAI project, in collaboration with the University of Bonn, University Hospital Bonn, and FIZ Karlsruhe, aims to protect AI-based medical devices and applications from cyberattacks, especially focusing on real-time seizure detection devices.
- The project, led by Professor Elena Demidova, is developing AI methods for data authentication, creating explainable AI models, and analyzing data protection and IT regulations to ensure the integrity, availability, and reliability of AI-enabled health-and-wellness systems.
- Funded by the German Ministry for Education and Research (BMFTR), the SecureNeuroAI project hopes to significantly enhance the cybersecurity of AI-driven medical devices for various conditions, such as epilepsy, Parkinson's, and stroke, and create job opportunities in software development.
