Differential privacy protects patient data in clinical EEG features
Differentially Private EEG Feature Anonymization: A Privacy-Utility Case Study in Clinical Neurophysiology
Cryptography and SecurityMachine Learning
Summary
EEG brain data can reveal private information about patients, so keeping it safe when shared or analyzed is important. The authors studied ways to add controlled noise to EEG features to protect patient identity while still keeping the data useful for medical analysis. They tested different methods to see which balance keeping data private and still good for AI tools. Their results show protecting privacy is possible but tricky, especially with small or uneven patient data. This work helps understand how to better share sensitive brain data safely.
What this means in practice
- •For hospital data teams: Add privacy protection to patient EEG data shared across hospitals while preserving clinical analysis usefulness.
- •For medical ai developers: Improve development of AI clinical systems by integrating differentially private EEG features to protect patient identity.
Authors
Noman Sadiq, Mohsen Toorani
Abstract
Clinical electroencephalography (EEG) data are valuable for healthcare research and for developing artificial intelligence (AI)-based clinical decision-support systems, but EEG recordings and derived features may contain sensitive patient-specific information. This creates privacy risks when data are reused, analyzed, or shared across clinical and research environments. Conventional anonymization methods are often insufficient for high-dimensional biomedical signals, since removing direct identifiers does not necessarily prevent re-identification, linkage, or inference risks. At the same time, strong privacy protection may distort clinically relevant signal characteristics and reduce data utility. This paper studies subject-level differential privacy for protecting clinical EEG-derived feature representations using Gaussian and Laplace perturbations. The proposed framework considers three deployment scenarios: client-side anonymization, centralized server-side anonymization, and decentralized local training. Following EEG preprocessing and feature extraction, Gaussian and Laplace perturbations are applied to the resulting patient-level EEG feature representations. The Laplace experiments evaluate the implemented noise scales, while the scales required for formal full-vector calibration are derived separately. The effects of both perturbations are assessed using statistical utility measures and a downstream machine-learning-based utility check. The results show that differentially private perturbation can be integrated into EEG processing workflows, but the selected mechanism, privacy parameters, and sensitivity calibration strongly influence data utility. The study highlights the practical privacy-utility trade-off in DP-based EEG feature anonymization and the challenges of preserving downstream utility in small and imbalanced clinical EEG datasets.