Related Experiment Video
Updated: Oct 18, 2025
06:40
Microstate and Omega Complexity Analyses of the Resting-state Electroencephalography
Published on: June 15, 2018
10.4K
The Time-Robustness Analysis of Individual Identification Based on Resting-State EEG
Yang Di1, Xingwei An1, Wenxiao Zhong2
1Tianjin International Joint Research Center for Neural Engineering, Academy of Medical Engineering and Translational Medicine, Tianjin University, Tianjin, China.
Frontiers in Human Neuroscience
|September 30, 2021
Summary
Electroencephalogram (EEG) signals show time-robustness for individual identification. Features like Power Spectral Density and Channel Coherence achieve high classification accuracy, demonstrating EEG
Area of Science:
- Neuroscience
- Biometrics
- Signal Processing
Background:
- Growing interest in biosignal-based individual identification, including electroencephalogram (EEG) and magnetic resonance imaging (MRI).
- Previous research suggests brain activity during resting states (eyes open/closed) contains unique individual information.
- Noisy EEG signals can impact experimental accuracy, necessitating investigation into feature stability and time-robustness.
Purpose of the Study:
- To investigate the stability and time-robustness of inter-individual EEG features for reliable individual identification.
- To evaluate different measures for extracting individual features from EEG data.
- To determine optimal frequency ranges for accurate classification across different experimental sessions.
Main Methods:
- Conducted three experiments with at least a two-week interval between sessions.
- Extracted individual features using Power Spectral Density, Cross Spectrum, Channel Coherence, and Phase Lags.
- Utilized Pearson Correlation Coefficient (PCC) for intra-individual correlation and Support Vector Machine (SVM) for classification accuracy.
Main Results:
- Achieved high intra-experiment classification accuracies (85-100%) and fusion experiment accuracies (80-100%).
- Identified optimized frequency ranges for inter-experiment classification: 13-40 Hz for resting-state eyes open (REO) and 8-40 Hz for resting-state eyes closed (REC) for Power Spectral Density, Channel Coherence, and Cross Spectrum.
- Phase Lags demonstrated significantly lower classification results compared to the other three features.
Conclusions:
- EEG signals exhibit significant time-robustness, making them suitable for individual identification systems.
- Power Spectral Density, Cross Spectrum, and Channel Coherence are reliable features for time-robust individual identification.
- The findings support the development of EEG-based biometric systems.
