Related Experiment Video
Updated: Jul 16, 2026
05:59
New Framework for Understanding Cross-Brain Coherence in Functional Near-Infrared Spectroscopy (fNIRS) Hyperscanning Studies
Published on: October 6, 2023
Fast feature selection to compare broadband with narrowband phase synchronization in brain-computer interfaces.
1Swiss Center for Electronics and Microtechnology, Rue Jaquet-Droz 1, 2007 Neuchâtel, Switzerland.
Methods of Information in Medicine
|March 10, 2007
Summary
This study compares brain activity features for brain-computer interfaces (BCIs). Power spectral density and phase locking value features effectively distinguish mental tasks, improving BCI communication for motor-disabled individuals.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Brain-computer interfaces (BCIs) offer communication solutions for individuals with motor disabilities.
- BCIs utilize brain activity, not muscle signals, for device control.
- Effective feature extraction is crucial for BCI performance.
Purpose of the Study:
- To compare phase synchronization and power spectral density (PSD) features for discriminating mental tasks.
- To evaluate features extracted from narrowband and broadband filtered electroencephalography (EEG) signals.
- To assess the efficacy of the Fast Correlation Based Filter (FCBF) for feature selection in BCIs.
Main Methods:
- EEG signals recorded from five subjects during left/right hand movement imagination and word generation.
- Application of a modified Fast Correlation Based Filter (FCBF) for feature selection.
- Computation of PSD and phase synchronization features from narrowband (8-12 Hz) and broadband (8-30 Hz) filtered EEG signals.
Main Results:
- Selected features originated from electrodes over the motor cortex, aligning with neurophysiological evidence.
- PSD features showed higher discrimination in narrowband filtered signals.
- Phase locking value (PLV) features were more discriminative in broadband filtered signals.
Conclusions:
- The FCBF algorithm demonstrated comparable generalization performance to SVM-rfe.
- FCBF is faster and selects fewer features than SVM-rfe.
- FCBF shows potential as a valuable tool for advancing BCI technology.
