Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Protecting Multiple Types of Privacy Simultaneously in EEG-based Brain-Computer Interfaces

Lubin Meng, Xue Jiang, Tianwang Jia, Dongrui Wu

TL;DR

This work exposes a privacy vulnerability in EEG-based BCIs by showing that user identity, gender, and BCI-experience can be inferred from EEG signals. It proposes a perturbation-generation framework that builds small, class-specific perturbations for each privacy type and sums them to produce privacy-protected EEG data, preserving the primary BCI task. Experimental results demonstrate that private-information classifiers experience significant accuracy drops on perturbed data, while task-classifier performance remains largely unchanged, confirming practical privacy protection. The approach generalizes across model architectures and offers a viable path toward privacy-preserving data sharing and safer EEG-based BCI systems.

Abstract

A brain-computer interface (BCI) enables direct communication between the brain and an external device. Electroencephalogram (EEG) is the preferred input signal in non-invasive BCIs, due to its convenience and low cost. EEG-based BCIs have been successfully used in many applications, such as neurological rehabilitation, text input, games, and so on. However, EEG signals inherently carry rich personal information, necessitating privacy protection. This paper demonstrates that multiple types of private information (user identity, gender, and BCI-experience) can be easily inferred from EEG data, imposing a serious privacy threat to BCIs. To address this issue, we design perturbations to convert the original EEG data into privacy-protected EEG data, which conceal the private information while maintaining the primary BCI task performance. Experimental results demonstrated that the privacy-protected EEG data can significantly reduce the classification accuracy of user identity, gender and BCI-experience, but almost do not affect at all the classification accuracy of the primary BCI task, enabling user privacy protection in EEG-based BCIs.

Protecting Multiple Types of Privacy Simultaneously in EEG-based Brain-Computer Interfaces

TL;DR

This work exposes a privacy vulnerability in EEG-based BCIs by showing that user identity, gender, and BCI-experience can be inferred from EEG signals. It proposes a perturbation-generation framework that builds small, class-specific perturbations for each privacy type and sums them to produce privacy-protected EEG data, preserving the primary BCI task. Experimental results demonstrate that private-information classifiers experience significant accuracy drops on perturbed data, while task-classifier performance remains largely unchanged, confirming practical privacy protection. The approach generalizes across model architectures and offers a viable path toward privacy-preserving data sharing and safer EEG-based BCI systems.

Abstract

A brain-computer interface (BCI) enables direct communication between the brain and an external device. Electroencephalogram (EEG) is the preferred input signal in non-invasive BCIs, due to its convenience and low cost. EEG-based BCIs have been successfully used in many applications, such as neurological rehabilitation, text input, games, and so on. However, EEG signals inherently carry rich personal information, necessitating privacy protection. This paper demonstrates that multiple types of private information (user identity, gender, and BCI-experience) can be easily inferred from EEG data, imposing a serious privacy threat to BCIs. To address this issue, we design perturbations to convert the original EEG data into privacy-protected EEG data, which conceal the private information while maintaining the primary BCI task performance. Experimental results demonstrated that the privacy-protected EEG data can significantly reduce the classification accuracy of user identity, gender and BCI-experience, but almost do not affect at all the classification accuracy of the primary BCI task, enabling user privacy protection in EEG-based BCIs.

Paper Structure

This paper contains 15 sections, 4 equations, 4 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Method
  3. Problem Statement
  4. Perturbation Generation
  5. Experiments
  6. Datasets
  7. Models
  8. Performance Measure
  9. Experimental Settings
  10. Identity, Gender, and BCI-experience Privacy Mining
  11. Identity, Gender, and BCI-experience Privacy Protection
  12. BCI Task Performance
  13. Characteristics of the Privacy-Protected EEG Data
  14. Visualization of the Training Process
  15. Conclusions

Figures (4)

  • Figure 1: Original unperturbed EEG trials and their perturbed counterparts from the three BCI tasks, which are almost identical. (a) ERP; (b) MI; and, (c) SSVEP. The perturbations are magnified $10$ times for better visualization. The interval between each dashed line represents the amplitude of 8mV for EEG signals, and 0.5mV for perturbations.
  • Figure 2: Average Cz channel spectrograms of the original unperturbed EEG trials and the privacy-protected EEG trials on three BCI tasks. (a) ERP; (b) MI; and, (c) SSVEP.
  • Figure 3: Average topoplots of the original unperturbed EEG trials and the privacy-protected EEG trials on three BCI tasks. (a) ERP; (b) MI; and, (c) SSVEP.
  • Figure 4: The training and test BCAs of three Privacy-Classifiers and three Task-Classifiers on the original unperturbed EEG data and the privacy-protected EEG data. (a) Privacy-Classifiers; and, (b) Task-Classifiers.