The use of personal and personally identifiable data, which also and especially relates to biometric data, is strictly regulated in the context of the General Data Protection Regulation (GDPR). However, in many cases, this strong protection also results in a more or less severely restricted usability of the data, which in turn conflicts with its social value for medical knowledge gain and the further development of technologies. To address and resolve this dilemma, technical solutions are needed that enable the provision and use of data in accordance with data protection regulations in the spirit of “open data,” thereby addressing the requirements of data protection and data use at the same time.
The NEMO project therefore focuses on researching and validating new methods for re-identification analysis and adaptive anonymization of biosignals, in particular EEG data. Given the growing importance of this data for analysis and medical applications, as well as the proliferation of consumer devices for sleep monitoring, the project emphasizes the legal and ethical challenges regarding re-identification risks and the potential derivation of sensitive medical information. The goal is to develop methods for quantifying re-identification risks based on privacy metrics and for the adaptive anonymization of EEG data. The NEMO toolbox, which combines statistical analysis, differential privacy, and anonymization techniques, is being developed and evaluated. The trade-off between data protection and user benefits is explained, and tools for application-specific data exploration are being developed. The integrated data platform must take data protection and scaling aspects into account and ensure high maintainability. The NEMO toolbox is being iteratively evaluated and improved based on analysis workflows, particularly for sleep phase detection, and further developed through feedback from practical use. The project demonstrates how adequately anonymized medical data can contribute to the advancement of knowledge in data-based medicine in the context of open science.
The Institute for Medical Informatics and Statistics (IMIS), the Center for Integrative Psychiatry/Sleep Laboratory (ZIP), and the Department of Neurology at UKSH are participating in the project. The goal of the UKSH partners is to make electroencephalogram (EEG) data already available in the clinic and associated research groups accessible for future research projects via the IMIS Medical Data Integration Center (MeDIC). Particular attention is being paid to minimizing the risk of re-identification of patients and test subjects.
