Deep Learning for Non-invasive Cortical Potential Imaging

Razorenova, A.M.; Yavich, N.; Malovichko, Mikhail; Fedorov, Maxim V.; Koshev, Nikolay; Dylov, Dmitry V.

doi:10.1007/978-3-030-66843-3_5

Cited by 8 publications

(4 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a recent work, Razorenova et al (2020) showed that the inverse problem for cortical potential imaging could be solved using a deep U-Net architecture, once more providing a strong case for the feasibility of ANNs to solve the EEG inverse problem (Fedorov et al, 2020). Tankelevich (2019) showed that a deep feed-forward network can find the correct set of source clusters that produced a given scalp signal.…”

Section: Artificial Neural Network and Inverse Solutionsmentioning

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: Artificial Neural Network and Inverse Solutionsmentioning

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

“…A big number of comparatively effective methods were developed in the neuroimaging area. Recently, several promising approaches based on the solution of the ill-posed Cauchy problems for elliptic PDEs were proposed by various mathematicians [7, 10, 12, 13, 28, 33, 36], including some ML/AI-based methods [41]. One of the further works in the nearest future will be devoted to development of high-accuracy algorithms for reconstruction of the mass MNPs distribution and validation of it in the real MPI/MRX experiments.…”

Section: Discussionmentioning

confidence: 99%

Yttrium-iron garnet film magnetometer for magnetic microparticles in vivo registration studies

Koshev

Kapralov

Evstigneeva

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

In the current article, we present a new kind of magnetometer for quantitative determination of magnetic objects in biological fluids and tissues. The sensor is based on yttrium-iron garnet film with optical signal registration system. Inheriting the working principle of a fluxgate magnetometers, the sensor works at a room-temperature, its wide dynamic range allows the measurements in an unshielded environment. A small size of sensitive element combined with a short recovery time after the excitation coils are off provide us with a potentially high spatial and temporal resolution of measurements. We show the feasibility of the sensor by sensing the remanent magnetization of Magnetic Nanoparticles (MNPs) both in vitro (test tubes, dry MNPs) and in vivo (local injection of the MNPs into mice).

show abstract

“…With the large increase of computing power and data resources in the past two decades, ANNs have gained in popularity and are now used successfully in a variety of tasks, e.g., image classification (Krizhevsky, Sutskever, & Hinton, 2012) and classification of single trial EEG (Schirrmeister et al, 2017). With this leap in technology, ANNs are now being reconsidered to solve the inverse problem in M/EEG and various research groups are starting to develop and refine architectures (see Awan, Saleem, & Kiran, 2019; Fedorov, Koshev, & Dylov, 2020; Razorenova et al, 2020; Zorzos, Kakkos, Ventouras, & Matsopoulos, 2021 for reviews). Deep ANNs were considered for inverse problems in other domains, too (Jin, McCann, Froustey, & Unser, 2017).…”

Section: Introductionmentioning

confidence: 99%

Evaluation of Long-Short Term Memory Networks for M/EEG Source Imaging with Simulated and Real EEG Data

Hecker

Rupprecht

Elst

et al. 2022

Preprint

View full text Add to dashboard Cite

Magneto- and electroencephalography (M/EEG) are widespread techniques to measure neural activity in vivo at a high temporal resolution but relatively low spatial resolution. Locating the sources underlying the M/EEG poses an inverse problem, which is itself ill-posed. In recent years, a new class of source imaging methods was developed based on artificial neural networks. We present a long-short term memory (LSTM) network to solve the M/EEG inverse problem. It integrates several aspects essential for qualitative inverse solutions: Low computational cost, exploitation of both spatial and temporal information, input flexibility and robustness to noise. Using simulation data, the LSTM shows higher accuracy on multiple metrics and for varying numbers of neural sources, compared to classical inverse solutions but also compared to our alternative architectures without integration of temporal information. It successfully integrates the temporal context given by sequential data, which is particularly useful with noisy data. Real data of a median nerve stimulation paradigm was used to show that the LSTM predicts plausible sources that are in concordance with classical inverse solutions. The performance of the LSTM regarding its robustness to noise renders it a promising and easy-to-apply inverse solution to be considered in future source localization studies and for clinical applications.

show abstract

Deep Learning for Non-invasive Cortical Potential Imaging

Cited by 8 publications

References 27 publications

ConvDip: A Convolutional Neural Network for Better EEG Source Imaging

ConvDip: A Convolutional Neural Network for Better EEG Source Imaging

Yttrium-iron garnet film magnetometer for magnetic microparticles in vivo registration studies

Evaluation of Long-Short Term Memory Networks for M/EEG Source Imaging with Simulated and Real EEG Data

Contact Info

Product

Resources

About