EEG source localization using spatio-temporal neural network

Cui, Song; Duan, Lijuan; Gong, Bei; Qiao, Yuanhua; Xu, Fan; Chen, Jun-Cheng; Wang, Changming

doi:10.23919/jcc.2019.07.011

Cited by 27 publications

(22 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Possible further advances could be made by preserving the 3D structure of the output space by, instead of a flattened output layer. This can be realized using a deconvolutional neural network for 2D-3D projections (Lin et al, 2018 ) or by using spatiotemporal information with a LSTM network (Cui et al, 2019 ). One obvious further way to improve the performance of ConvDip may be to increase the capacity of the neural network, e.g., by adding more layers.…”

Section: Discussionmentioning

confidence: 99%

“…Only very recently several studies about ANN-based solutions to the inverse problem have been published. Cui et al ( 2019 ) showed that a neural network can be trained to reconstruct the position and time course of a single source using a long-short term memory (LSTM) recurrent neural network architecture (Hochreiter and Schmidhuber, 1997 ). LSTMs allow to not only use single time instances of (e.g., EEG) data but instead learn from temporally lagged information.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

ConvDip: A Convolutional Neural Network for Better EEG Source Imaging

et al. 2021

View full text Add to dashboard Cite

The electroencephalography (EEG) is a well-established non-invasive method in neuroscientific research and clinical diagnostics. It provides a high temporal but low spatial resolution of brain activity. To gain insight about the spatial dynamics of the EEG, one has to solve the inverse problem, i.e., finding the neural sources that give rise to the recorded EEG activity. The inverse problem is ill-posed, which means that more than one configuration of neural sources can evoke one and the same distribution of EEG activity on the scalp. Artificial neural networks have been previously used successfully to find either one or two dipole sources. These approaches, however, have never solved the inverse problem in a distributed dipole model with more than two dipole sources. We present ConvDip, a novel convolutional neural network (CNN) architecture, that solves the EEG inverse problem in a distributed dipole model based on simulated EEG data. We show that (1) ConvDip learned to produce inverse solutions from a single time point of EEG data and (2) outperforms state-of-the-art methods on all focused performance measures. (3) It is more flexible when dealing with varying number of sources, produces less ghost sources and misses less real sources than the comparison methods. It produces plausible inverse solutions for real EEG recordings from human participants. (4) The trained network needs <40 ms for a single prediction. Our results qualify ConvDip as an efficient and easy-to-apply novel method for source localization in EEG data, with high relevance for clinical applications, e.g., in epileptology and real-time applications.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

ConvDip: A Convolutional Neural Network for Better EEG Source Imaging

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Only very recently several studies about ANN-based solutions to the inverse problem have been published. Cui et al (2019) showed that a neural network can be trained to reconstruct the position and time course of a single source using a long-short term memory recurrent neural network (LSTM) architecture (Hochreiter & Schmidhuber, 1997). LSTMs allow to not only use single time instances of (e.g.…”

Section: Artificial Neural Network (Ann) and Inverse Solutionsmentioning

confidence: 99%

“…Possible further advances could be made by preserving the 3D structure of the output space by, instead of a flattened output layer. This can be realized using a deconvolutional neural network for 2D-3D projections (Lin, Kong, & Lucey, 2018) or by using spatiotemporal information with a LSTM network (Cui et al, 2019). One obvious further way to improve the performance of ConvDip may be to increase the capacity of the neural network, e.g.…”

Section: Further Perspectives For Improvementmentioning

confidence: 99%

ConvDip: A convolutional neural network for better EEG Source Imaging

Hecker

Rupprecht

Elst

et al. 2020

Preprint

View full text Add to dashboard Cite

EEG and MEG are well-established non-invasive methods in neuroscientific research and clinical diagnostics. Both methods provide a high temporal but low spatial resolution of brain activity. In order to gain insight about the spatial dynamics of the M/EEG one has to solve the inverse problem, which means that more than one configuration of neural sources can evoke one and the same distribution of EEG activity on the scalp. Artificial neural networks have been previously used successfully to find either one or two dipoles sources. These approaches, however, have never solved the inverse problem in a distributed dipole model with more than two dipole sources. We present ConvDip, a novel convolutional neural network (CNN) architecture that solves the EEG inverse problem in a distributed dipole model based on simulated EEG data. We show that (1) ConvDip learned to produce inverse solutions from a single time point of EEG data and (2) outperforms state-of-the-art methods (eLORETA and LCMV beamforming) on all focused performance measures. (3) It is more flexible when dealing with varying number of sources, produces less ghost sources and misses less real sources than the comparison methods. (4) It produces plausible inverse solutions for real-world EEG recordings and needs less than 40 ms for a single forward pass. Our results qualify ConvDip as an efficient and easy-to-apply novel method for source localization in EEG and MEG data, with high relevance for clinical applications, e.g. in epileptology and real time applications.

show abstract

“…Finally, the most popular trend today is to use the temporal information as an additional constraint and to apply Markov models, or their approximations, in a recurrent network configuration (e.g. LSTM) [18,13,5].…”

Section: State-of-the-artmentioning

confidence: 99%

Deep Learning for Non-Invasive Cortical Potential Imaging

Razorenova

Yavich

Malovichko

et al. 2020

Preprint

View full text Add to dashboard Cite

0000−0002−3386−6914] , Nikolay Yavich 1[0000−0002−8913−7710] , Mikhail Malovichko 1[0000−0002−7618−0998] , Maxim Fedorov 1[0000−0003−3901−3565] , Nikolay Koshev 1[0000−0001−7241−3304] , and Dmitry V. Dylov 1[0000−0003−2251−3221]Abstract. Electroencephalography (EEG) is a well-established non-invasive technique to measure the brain activity, albeit with a limited spatial resolution. Variations in electric conductivity between different tissues distort the electric fields generated by cortical sources, resulting in a smeared potential measurements on the scalp. One needs to solve an ill-posed inverse problem to recover the original neural activity. In this article, we present a generic methodology of recovering the cortical potentials from the EEG measurement by introducing a new inverse-problem solver based on deep convolutional neural networks (CNN) in a U-Net configuration. The solver was trained on a paired dataset from a synthetic head conductivity model by solving the diffusion equations with the finite element method (FEM). This is the first method that provides robust translation of EEG data to the cortex surface using deep learning. Supplying a fast and accurate interpretation of the tracked EEG signal, the proposed approach is a candidate for future non-invasive brain-computer interface devices.

show abstract

EEG source localization using spatio-temporal neural network

Cited by 27 publications

References 0 publications

ConvDip: A Convolutional Neural Network for Better EEG Source Imaging

ConvDip: A Convolutional Neural Network for Better EEG Source Imaging

ConvDip: A convolutional neural network for better EEG Source Imaging

Deep Learning for Non-Invasive Cortical Potential Imaging

Contact Info

Product

Resources

About