DeepReducer: A linear transformer-based model for MEG denoising
Measuring event-related magnetic fields (ERFs) in magnetoencephalography (MEG) is crucial for investigating perceptual and cognitive information processing in both neuroscience research and clinical practice. However, the magnitude of the ERF in cortical sources is comparable to the noise in a singl...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-03-01
|
| Series: | NeuroImage |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1053811925000825 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823856884819427328 |
|---|---|
| author | Hui Xu Li Zheng Pan Liao Bingjiang Lyu Jia-Hong Gao |
| author_facet | Hui Xu Li Zheng Pan Liao Bingjiang Lyu Jia-Hong Gao |
| author_sort | Hui Xu |
| collection | DOAJ |
| description | Measuring event-related magnetic fields (ERFs) in magnetoencephalography (MEG) is crucial for investigating perceptual and cognitive information processing in both neuroscience research and clinical practice. However, the magnitude of the ERF in cortical sources is comparable to the noise in a single trial. Consequently, numerous repetitive recordings are needed to distinguish these sources from background noise, requiring lengthy time for data acquisition. Herein, we introduce DeepReducer, a linear transformer-based deep learning model designed to reliably and efficiently denoise ERFs, thereby reducing the number of required trials. DeepReducer was trained on a mix of limited-trial and multi-trial averaged ERFs, employing mean squared error as the loss function to effectively capture and model the complex signal fluctuations inherent in MEG recordings. Validation on both semi-synthetic and experimental task-related MEG data showed that DeepReducer outperforms conventional trial-averaging techniques, significantly improving the signal-to-noise ratio of ERFs and reducing source localization errors. The practical significance of DeepReducer encompasses optimizing MEG data acquisition by reducing participant stress (particularly for patients) and minimizing associated artifacts. |
| format | Article |
| id | doaj-art-1f51c8e52b5b4003a096b110cd327fa4 |
| institution | Kabale University |
| issn | 1095-9572 |
| language | English |
| publishDate | 2025-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | NeuroImage |
| spelling | doaj-art-1f51c8e52b5b4003a096b110cd327fa42025-02-12T05:30:42ZengElsevierNeuroImage1095-95722025-03-01308121080DeepReducer: A linear transformer-based model for MEG denoisingHui Xu0Li Zheng1Pan Liao2Bingjiang Lyu3Jia-Hong Gao4McGovern Institute for Brain Research, Peking University, Beijing 100871, PR China; Center for MRl Research, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing 100871, PR China; Beijing City Key Lab for Medical Physics and Engineering, Institution of Heavy Ion Physics, School of Physics, Peking University, Beijing 100871, PR ChinaState Key Laboratory of Brain and Cognitive Science, Institute of Biophysics, Chinese Academy of Sciences, Beijing 100101, ChinaCenter for MRl Research, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing 100871, PR China; Changping Laboratory, Beijing 102206, PR ChinaChangping Laboratory, Beijing 102206, PR ChinaMcGovern Institute for Brain Research, Peking University, Beijing 100871, PR China; Center for MRl Research, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing 100871, PR China; Beijing City Key Lab for Medical Physics and Engineering, Institution of Heavy Ion Physics, School of Physics, Peking University, Beijing 100871, PR China; State Key Laboratory of Brain and Cognitive Science, Institute of Biophysics, Chinese Academy of Sciences, Beijing 100101, China; Changping Laboratory, Beijing 102206, PR China; National Biomedical Imaging Center, Peking University, Beijing 100871, PR China; Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, PR China; Corresponding author: Integrated Science Research Center, Peking University, 5 Yiheyuan Road, Haidian District, Beijing 100871, PR China.Measuring event-related magnetic fields (ERFs) in magnetoencephalography (MEG) is crucial for investigating perceptual and cognitive information processing in both neuroscience research and clinical practice. However, the magnitude of the ERF in cortical sources is comparable to the noise in a single trial. Consequently, numerous repetitive recordings are needed to distinguish these sources from background noise, requiring lengthy time for data acquisition. Herein, we introduce DeepReducer, a linear transformer-based deep learning model designed to reliably and efficiently denoise ERFs, thereby reducing the number of required trials. DeepReducer was trained on a mix of limited-trial and multi-trial averaged ERFs, employing mean squared error as the loss function to effectively capture and model the complex signal fluctuations inherent in MEG recordings. Validation on both semi-synthetic and experimental task-related MEG data showed that DeepReducer outperforms conventional trial-averaging techniques, significantly improving the signal-to-noise ratio of ERFs and reducing source localization errors. The practical significance of DeepReducer encompasses optimizing MEG data acquisition by reducing participant stress (particularly for patients) and minimizing associated artifacts.http://www.sciencedirect.com/science/article/pii/S1053811925000825MEGERFDeep learningDenoiseTransformer |
| spellingShingle | Hui Xu Li Zheng Pan Liao Bingjiang Lyu Jia-Hong Gao DeepReducer: A linear transformer-based model for MEG denoising NeuroImage MEG ERF Deep learning Denoise Transformer |
| title | DeepReducer: A linear transformer-based model for MEG denoising |
| title_full | DeepReducer: A linear transformer-based model for MEG denoising |
| title_fullStr | DeepReducer: A linear transformer-based model for MEG denoising |
| title_full_unstemmed | DeepReducer: A linear transformer-based model for MEG denoising |
| title_short | DeepReducer: A linear transformer-based model for MEG denoising |
| title_sort | deepreducer a linear transformer based model for meg denoising |
| topic | MEG ERF Deep learning Denoise Transformer |
| url | http://www.sciencedirect.com/science/article/pii/S1053811925000825 |
| work_keys_str_mv | AT huixu deepreduceralineartransformerbasedmodelformegdenoising AT lizheng deepreduceralineartransformerbasedmodelformegdenoising AT panliao deepreduceralineartransformerbasedmodelformegdenoising AT bingjianglyu deepreduceralineartransformerbasedmodelformegdenoising AT jiahonggao deepreduceralineartransformerbasedmodelformegdenoising |