Unlocking Security for Comprehensive Electroencephalogram-Based User Authentication Systems
Adnan Elahi Khan Khalil1, Jesus Arturo Perez-Diaz1, Jose Antonio Cantoral-Ceballos1
1School of Engineering and Sciences, Tecnologico de Monterrey, Monterrey 64700, Nuevo Leon, Mexico.
Sensors (Basel, Switzerland)
|January 8, 2025
Summary
This study introduces a novel electroencephalogram (EEG)-based authentication system using a neural network. The system accurately identifies and authenticates users based on brain signals, achieving 97% accuracy for enhanced security.
Area of Science:
- Neuroscience
- Computer Science
- Biometrics
Background:
- Growing need for robust security systems due to AI advancements.
- Increasing interest in brain signal (EEG) analysis for user authentication.
- Limitations of previous EEG-based methods in achieving high accuracy.
Purpose of the Study:
- To develop and evaluate an EEG-based user authentication scheme.
- To utilize P300 potentials and a multi-layer perceptron feedforward neural network (MLP FFNN).
- To achieve high accuracy in both user identification and authentication.
Main Methods:
- Utilized electroencephalogram (EEG) signals focusing on P300 potentials.
- Employed a multi-layer perceptron feedforward neural network (MLP FFNN).
- Feature extraction using mutual information (MI) on power spectral density (PSD) across five frequency bands.
- Two-phase process: user identification (multi-class classification) and user authentication (probability assessment).
Main Results:
- Achieved 97% accuracy in EEG-based user identification.
- Achieved 97% accuracy in EEG-based user authentication.
- The scheme accommodates new users without retraining.
Conclusions:
- The proposed EEG-based authentication scheme offers a reliable and accurate method for safeguarding individual assets.
- The combination of P300 potentials, MLP FFNN, and MI feature extraction provides robust authentication.
- This approach represents a significant advancement in biometric security using brain-computer interfaces.
