Joint hybrid recursive feature elimination based channel selection and ResGCN for cross session MI recognition
Duan Li1, Keyun Li1, Yongquan Xia2
1School of Computer Science and Technology, Zhengzhou University of Light Industry, Zhengzhou, Henan, China.
Scientific Reports
|October 9, 2024
Summary
This study introduces a novel Hybrid-Recursive Feature Elimination (H-RFE) method for selecting electroencephalography (EEG) channels in brain-computer interfaces (BCI). The approach significantly enhances motor imagery (MI) recognition accuracy by optimizing channel selection for BCI systems.
Area of Science:
- Neuroscience and Biomedical Engineering
- Signal Processing and Machine Learning
Background:
- Brain-computer interfaces (BCI) utilizing multi-channel electroencephalography (EEG) are crucial for motor imagery (MI) recognition in applications like sensory compensation.
- Individual variability and noisy EEG data present challenges to BCI performance, necessitating effective channel selection strategies.
Purpose of the Study:
- To develop and evaluate a novel channel selection method, Hybrid-Recursive Feature Elimination (H-RFE), for improving motor imagery recognition in BCI systems.
- To integrate H-RFE with residual graph neural networks to effectively utilize spatiotemporal EEG information.
Main Methods:
- A Hybrid-Recursive Feature Elimination (H-RFE) strategy was employed for adaptive channel selection, using random forest, gradient boosting, and logistic regression as evaluators.
- A graph neural network with residual blocks was utilized to process multi-channel EEG data and recognize motor imagery tasks.
- The proposed method was validated on the SHU and PhysioNet datasets.
Main Results:
- The H-RFE method achieved high cross-session MI recognition accuracy: 90.03% on the SHU dataset using 73.44% of channels and 93.99% on the PhysioNet dataset using 72.5% of channels.
- The proposed channel selection approach demonstrated substantial improvements in classification accuracy compared to traditional methods, including significant gains over correlation-based and mutual information-based selections.
- The method effectively reduced channel redundancy and noise, leading to enhanced BCI performance.
Conclusions:
- The proposed H-RFE combined with residual graph neural networks offers a superior approach for channel selection in motor imagery-based BCIs.
- This method enhances BCI accuracy and efficiency by adaptively selecting optimal EEG channels tailored to individual subjects.
- The findings suggest a promising direction for developing more robust and precise brain-computer interface systems.
Keywords:
Brain-computer interface (BCI)Channel selectionDeep learningGraph convolutional neural network (GCN)Motor imagery (MI)More Related Videos
Related Concept Videos
01:03IR Frequency Region: Fingerprint Region
IR spectra are divided into two main regions: the diagnostic region and the fingerprint region. The diagnostic region of the spectrum lies above 1500 cm−1. The absorptions resulting from single-bond vibrations of the N–H, C–H, and O–H stretch at higher wavenumbers and appear on the left side of the spectrum. The stretching absorptions of the C≡C and C≡N occur between 2100–2300 cm−1. In contrast, those arising from stretching absorptions of the C=O, C=N, and C=C occur between 1600–1850 cm−1.
The...
The...
01:19Masking and Demasking Agents
EDTA titrations may necessitate masking and demasking agents to temporarily protect a particular metal ion in a mixture from the EDTA reaction. These agents facilitate the sequential analysis of the metal ions by forming stable complexes with some—but not all—metal ions during certain steps.
There are many masking agents, such as cyanide, fluoride, triethanolamine, thiourea, and 2,3-bis(sulfanyl)propan-1-ol (formerly 2,3-dimercapto-1-propanol), with the masking agent chosen based on the metal...
There are many masking agents, such as cyanide, fluoride, triethanolamine, thiourea, and 2,3-bis(sulfanyl)propan-1-ol (formerly 2,3-dimercapto-1-propanol), with the masking agent chosen based on the metal...