Detection of Movement Intention in EEG-Based Brain-Computer Interfaces Using Fourier-Based Synchrosqueezing
Nedime Karakullukcu1,2, Bülent Yilmaz1,3,4
1Electrical and Computer Engineering Department, Graduate School of Engineering and Sciences, Abdullah Gul University, 38080 Kayseri, Turkey.
International Journal of Neural Systems
|November 22, 2021
Summary
This study introduces a novel method using Fourier-based synchrosqueezing transform (FSST) to detect movement intention from electroencephalography (EEG) signals, enabling brain-computer interfaces (BCI) initiation. High accuracy was achieved, paving the way for intuitive BCI control.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Brain-computer interfaces (BCI) often require external aids for initiation, posing challenges for individuals with motor impairments.
- Current BCI systems necessitate significant preprocessing and external accessories, limiting their practical application.
Purpose of the Study:
- To identify and characterize movement intention from multichannel electroencephalography (EEG) signals for BCI initiation.
- To develop a BCI system that can be initiated without additional accessories or methodologies.
- To evaluate the efficacy of the Fourier-based synchrosqueezing transform (FSST) for discriminating resting and motor imagery states.
Main Methods:
- Utilized Fourier-based synchrosqueezing transform (FSST) as a feature extractor for EEG signals.
- Employed singular value decomposition (SVD) for feature selection and support vector machines (SVM) as the classifier.
- Investigated the performance of FSST on both clean and noisy EEG data without extensive preprocessing.
Main Results:
- Achieved high accuracy (99.8%) and f-measure (0.99) in discriminating resting and motor imagery states using FSST, SVD, and SVM.
- Demonstrated the effectiveness of the FSST-SVD combination even with noisy EEG data and minimal preprocessing.
- Identified specific EEG channels (F4-Fz-C3-Cz-C4-Pz) and statistical features with significant discrimination capabilities (p < 0.05).
Conclusions:
- The FSST-SVD combination enables real-time detection of movement preparation from EEG signals with minimal preprocessing.
- This approach offers a promising, accessory-free method for initiating BCI systems, enhancing accessibility for individuals with motor impairments.
- The findings highlight the potential of advanced signal processing techniques for improving BCI functionality and user experience.
