DeepReducer: A linear transformer-based model for MEG denoising

被引：0
作者：
Xu, Hui [1 ,2 ,3 ]
Zheng, Li [4 ]
Liao, Pan [2 ,5 ]
Lyu, Bingjiang [5 ]
Gao, Jia-Hong [1 ,2 ,3 ,4 ,5 ,6 ,7 ]
机构：
[1] Peking Univ, McGovern Inst Brain Res, Beijing 100871, Peoples R China
[2] Peking Univ, Acad Adv Interdisciplinary Studies, Ctr MRl Res, Beijing 100871, Peoples R China
[3] Peking Univ, Inst Heavy Ion Phys, Sch Phys, Beijing City Key Lab Med Phys & Engn, Beijing 100871, Peoples R China
[4] Chinese Acad Sci, Inst Biophys, State Key Lab Brain & Cognit Sci, Beijing 100101, Peoples R China
[5] Changping Lab, Beijing 102206, Peoples R China
[6] Peking Univ, Natl Biomed Imaging Ctr, Beijing 100871, Peoples R China
[7] Hefei Comprehens Natl Sci Ctr, Inst Artificial Intelligence, Hefei, Peoples R China
来源：
NEUROIMAGE | 2025年 / 308卷
基金：
中国国家自然科学基金;
关键词：
MEG; ERF; Deep learning; Denoise; Transformer; TEMPORAL DYNAMICS; RESPONSES; MAGNETOENCEPHALOGRAPHY; HABITUATION; POTENTIALS;
D O I：
10.1016/j.neuroimage.2025.121080
中图分类号：
Q189 [神经科学];
学科分类号：
071006 ;
摘要：
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.
引用
收藏
页数：14
相关论文
共 47 条
  • [1] Ahonen L., Huotilainen M., Brattico E., Within-and between-session replicability of cognitive brain processes: an MEG study with an N-back task, Physiol. Behav., 158, pp. 43-53, (2016)
  • [2] Al-Tahan H., Mohsenzadeh Y., Reconstructing feedback representations in the ventral visual pathway with a generative adversarial autoencoder, PLoS. Comput. Biol., 17, (2021)
  • [3] Baillet S., Magnetoencephalography for brain electrophysiology and imaging, Nat. Neurosci., 20, pp. 327-339, (2017)
  • [4] Beppi C., Violante I.R., Scott G., Sandrone S., EEG, MEG and neuromodulatory approaches to explore cognition: current status and future directions, Brain Cogn., 148, (2021)
  • [5] Boudewyn M.A., Luck S.J., Farrens J.L., Kappenman E.S., How many trials does it take to get a significant ERP effect?, Psychophysiology., 55, (2018)
  • [6] Bourbon W.T., Will K.W., Gary H.E., Papanicolaou A.C., Habituation of auditory event-related potentials: a comparison of self-initiated and automated stimulus trains, Electroencephalogr. Clin. Neurophysiol., 66, pp. 160-166, (1987)
  • [7] Brown T.B., (2020)
  • [8] Chang P.C., Chang J.R., Chen P.Y., Cheng L.K., Hsieh J.C., Yu H.Y., Chen L.F., Chen Y.S., Decoding neural representations of rhythmic sounds from magnetoencephalography, IEEE international conference on acoustics, speech and signal processing, pp. 1280-1284, (2021)
  • [9] Dale A.M., Liu A.K., Fischl B.R., Buckner R.L., Belliveau J.W., Lewine J.D., Halgren E., Dynamic statistical parametric mapping: combining fMRI and MEG for high-resolution imaging of cortical activity, Neuron, 26, pp. 55-67, (2000)
  • [10] Dawson G.D., A summation technique for the detection of small evoked potentials, Electroencephalogr. Clin. Neurophys., 6, pp. 65-84, (1954)
12345