A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source Imaging
Electrophysiological source imaging (ESI) refers to the process of reconstructing underlying activated sources on the cortex given the brain signal measured by Electroencephalography (EEG) or Magnetoencephalography (MEG). Due to the ill-posed nature of ESI, solving ESI requires the design of neuroph...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-04-01
|
| Series: | Frontiers in Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnins.2022.867466/full |
| _version_ | 1818018974283071488 |
|---|---|
| author | Meng Jiao Meng Jiao Guihong Wan Guihong Wan Yaxin Guo Dongqing Wang Hang Liu Jing Xiang Feng Liu |
| author_facet | Meng Jiao Meng Jiao Guihong Wan Guihong Wan Yaxin Guo Dongqing Wang Hang Liu Jing Xiang Feng Liu |
| author_sort | Meng Jiao |
| collection | DOAJ |
| description | Electrophysiological source imaging (ESI) refers to the process of reconstructing underlying activated sources on the cortex given the brain signal measured by Electroencephalography (EEG) or Magnetoencephalography (MEG). Due to the ill-posed nature of ESI, solving ESI requires the design of neurophysiologically plausible regularization or priors to guarantee a unique solution. Recovering focally extended sources is more challenging, and traditionally uses a total variation regularization to promote spatial continuity of the activated sources. In this paper, we propose to use graph Fourier transform (GFT) based bidirectional long-short term memory (BiLSTM) neural network to solve the ESI problem. The GFT delineates the 3D source space into spatially high, medium and low frequency subspaces spanned by corresponding eigenvectors. The low frequency components can naturally serve as a spatially low-band pass filter to reconstruct extended areas of source activation. The BiLSTM is adopted to learn the mapping relationship between the projection of low-frequency graph space and the recorded EEG. Numerical results show the proposed GFT-BiLSTM outperforms other benchmark algorithms in synthetic data under varied signal-to-noise ratios (SNRs). Real data experiments also demonstrate its capability of localizing the epileptogenic zone of epilepsy patients with good accuracy. |
| first_indexed | 2024-04-14T07:46:48Z |
| format | Article |
| id | doaj.art-1c1ee27be22740a2b17eb88f7c984f9b |
| institution | Directory Open Access Journal |
| issn | 1662-453X |
| language | English |
| last_indexed | 2024-04-14T07:46:48Z |
| publishDate | 2022-04-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroscience |
| spelling | doaj.art-1c1ee27be22740a2b17eb88f7c984f9b2022-12-22T02:05:19ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2022-04-011610.3389/fnins.2022.867466867466A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source ImagingMeng Jiao0Meng Jiao1Guihong Wan2Guihong Wan3Yaxin Guo4Dongqing Wang5Hang Liu6Jing Xiang7Feng Liu8School of Systems and Enterprises, Stevens Institute of Technology, Hoboken, NJ, United StatesCollege of Electrical Engineering, Qingdao University, Qingdao, ChinaDepartment of Dermatology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, United StatesDepartment of Biomedical Informatics, Harvard Medical School, Boston, MA, United StatesSchool of Systems and Enterprises, Stevens Institute of Technology, Hoboken, NJ, United StatesCollege of Electrical Engineering, Qingdao University, Qingdao, ChinaDepartment of Electrical and Computer Engineering, Stevens Institute of Technology, Hoboken, NJ, United StatesMEG Center, Division of Neurology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, United StatesSchool of Systems and Enterprises, Stevens Institute of Technology, Hoboken, NJ, United StatesElectrophysiological source imaging (ESI) refers to the process of reconstructing underlying activated sources on the cortex given the brain signal measured by Electroencephalography (EEG) or Magnetoencephalography (MEG). Due to the ill-posed nature of ESI, solving ESI requires the design of neurophysiologically plausible regularization or priors to guarantee a unique solution. Recovering focally extended sources is more challenging, and traditionally uses a total variation regularization to promote spatial continuity of the activated sources. In this paper, we propose to use graph Fourier transform (GFT) based bidirectional long-short term memory (BiLSTM) neural network to solve the ESI problem. The GFT delineates the 3D source space into spatially high, medium and low frequency subspaces spanned by corresponding eigenvectors. The low frequency components can naturally serve as a spatially low-band pass filter to reconstruct extended areas of source activation. The BiLSTM is adopted to learn the mapping relationship between the projection of low-frequency graph space and the recorded EEG. Numerical results show the proposed GFT-BiLSTM outperforms other benchmark algorithms in synthetic data under varied signal-to-noise ratios (SNRs). Real data experiments also demonstrate its capability of localizing the epileptogenic zone of epilepsy patients with good accuracy.https://www.frontiersin.org/articles/10.3389/fnins.2022.867466/fullelectroencephalographysource localizationinverse problemgraph Fourier transformBiLSTM |
| spellingShingle | Meng Jiao Meng Jiao Guihong Wan Guihong Wan Yaxin Guo Dongqing Wang Hang Liu Jing Xiang Feng Liu A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source Imaging Frontiers in Neuroscience electroencephalography source localization inverse problem graph Fourier transform BiLSTM |
| title | A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source Imaging |
| title_full | A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source Imaging |
| title_fullStr | A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source Imaging |
| title_full_unstemmed | A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source Imaging |
| title_short | A Graph Fourier Transform Based Bidirectional Long Short-Term Memory Neural Network for Electrophysiological Source Imaging |
| title_sort | graph fourier transform based bidirectional long short term memory neural network for electrophysiological source imaging |
| topic | electroencephalography source localization inverse problem graph Fourier transform BiLSTM |
| url | https://www.frontiersin.org/articles/10.3389/fnins.2022.867466/full |
| work_keys_str_mv | AT mengjiao agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT mengjiao agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT guihongwan agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT guihongwan agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT yaxinguo agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT dongqingwang agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT hangliu agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT jingxiang agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT fengliu agraphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT mengjiao graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT mengjiao graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT guihongwan graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT guihongwan graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT yaxinguo graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT dongqingwang graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT hangliu graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT jingxiang graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging AT fengliu graphfouriertransformbasedbidirectionallongshorttermmemoryneuralnetworkforelectrophysiologicalsourceimaging |