Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis
The motor imagery (MI) classification has been a prominent research topic in brain-computer interfaces (BCIs) based on electroencephalography (EEG). Over the past few decades, the performance of MI-EEG classifiers has seen gradual enhancement. In this study, we amplify the geometric deep-learning-ba...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/179051 http://arxiv.org/abs/2211.02641v4 |
| _version_ | 1811685876631076864 |
|---|---|
| author | Ju, Ce Guan, Cuntai |
| author2 | College of Computing and Data Science |
| author_facet | College of Computing and Data Science Ju, Ce Guan, Cuntai |
| author_sort | Ju, Ce |
| collection | NTU |
| description | The motor imagery (MI) classification has been a prominent research topic in brain-computer interfaces (BCIs) based on electroencephalography (EEG). Over the past few decades, the performance of MI-EEG classifiers has seen gradual enhancement. In this study, we amplify the geometric deep-learning-based MI-EEG classifiers from the perspective of time-frequency analysis, introducing a new architecture called Graph-CSPNet. We refer to this category of classifiers as Geometric Classifiers, highlighting their foundation in differential geometry stemming from EEG spatial covariance matrices. Graph-CSPNet utilizes novel manifold-valued graph convolutional techniques to capture the EEG features in the time-frequency domain, offering heightened flexibility in signal segmentation for capturing localized fluctuations. To evaluate the effectiveness of Graph-CSPNet, we employ five commonly used publicly available MI-EEG datasets, achieving near-optimal classification accuracies in nine out of 11 scenarios. The Python repository can be found at https://github.com/GeometricBCI/Tensor-CSPNet-and-Graph-CSPNet. |
| first_indexed | 2024-10-01T04:51:29Z |
| format | Journal Article |
| id | ntu-10356/179051 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T04:51:29Z |
| publishDate | 2024 |
| record_format | dspace |
| spelling | ntu-10356/1790512024-07-17T07:59:32Z Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis Ju, Ce Guan, Cuntai College of Computing and Data Science School of Computer Science and Engineering S-Lab Computer and Information Science Geometric deep learning Motor imagery classification Riemannian geometry The motor imagery (MI) classification has been a prominent research topic in brain-computer interfaces (BCIs) based on electroencephalography (EEG). Over the past few decades, the performance of MI-EEG classifiers has seen gradual enhancement. In this study, we amplify the geometric deep-learning-based MI-EEG classifiers from the perspective of time-frequency analysis, introducing a new architecture called Graph-CSPNet. We refer to this category of classifiers as Geometric Classifiers, highlighting their foundation in differential geometry stemming from EEG spatial covariance matrices. Graph-CSPNet utilizes novel manifold-valued graph convolutional techniques to capture the EEG features in the time-frequency domain, offering heightened flexibility in signal segmentation for capturing localized fluctuations. To evaluate the effectiveness of Graph-CSPNet, we employ five commonly used publicly available MI-EEG datasets, achieving near-optimal classification accuracies in nine out of 11 scenarios. The Python repository can be found at https://github.com/GeometricBCI/Tensor-CSPNet-and-Graph-CSPNet. Agency for Science, Technology and Research (A*STAR) Submitted/Accepted version This work was supported in part by the RIE2020 Industry Alignment Fund–Industry Collaboration Projects (IAF-ICP); and in part by the RIE2020 Advanced Manufacturing and Engineering (AME) Programmatic Grant, Singapore, under Grant A20G8b0102. 2024-07-17T07:59:32Z 2024-07-17T07:59:32Z 2023 Journal Article Ju, C. & Guan, C. (2023). Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis. IEEE Transactions On Neural Networks and Learning Systems. https://dx.doi.org/10.1109/TNNLS.2023.3307470 2162-237X https://hdl.handle.net/10356/179051 10.1109/TNNLS.2023.3307470 37725740 2-s2.0-85173008700 http://arxiv.org/abs/2211.02641v4 en A20G8b0102 IEEE Transactions on Neural Networks and Learning Systems © 2023 IEEE. All rights reserved. This article may be downloaded for personal use only. Any other use requires prior permission of the copyright holder. The Version of Record is available online at http://doi.org/10.1109/TNNLS.2023.3307470. application/pdf |
| spellingShingle | Computer and Information Science Geometric deep learning Motor imagery classification Riemannian geometry Ju, Ce Guan, Cuntai Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis |
| title | Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis |
| title_full | Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis |
| title_fullStr | Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis |
| title_full_unstemmed | Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis |
| title_short | Graph neural networks on SPD manifolds for motor imagery classification: a perspective from the time–frequency analysis |
| title_sort | graph neural networks on spd manifolds for motor imagery classification a perspective from the time frequency analysis |
| topic | Computer and Information Science Geometric deep learning Motor imagery classification Riemannian geometry |
| url | https://hdl.handle.net/10356/179051 http://arxiv.org/abs/2211.02641v4 |
| work_keys_str_mv | AT juce graphneuralnetworksonspdmanifoldsformotorimageryclassificationaperspectivefromthetimefrequencyanalysis AT guancuntai graphneuralnetworksonspdmanifoldsformotorimageryclassificationaperspectivefromthetimefrequencyanalysis |