Abstract
We propose a new method for EEG source localization. An efficient solution to this problem requires choosing an appropriate regularization term in order to constraint the original problem. In our work, we adopt the Bayesian framework to place constraints; hence, the regularization term is closely connected to the prior distribution. More specifically, we propose a new sparse prior for the localization of EEG sources. The proposed prior distribution has sparse properties favoring focal EEG sources. In order to obtain an efficient algorithm, we use the variational Bayesian (VB) framework which provides us with a tractable iterative algorithm of closed-form equations. Additionally, we provide extensions of our method in cases where we observe group structures and spatially extended EEG sources. We have performed experiments using synthetic EEG data and real EEG data from three publicly available datasets. The real EEG data are produced due to the presentation of auditory and visual stimulus. We compare the proposed method with well-known approaches of EEG source localization and the results have shown that our method presents state-of-the-art performance, especially in cases where we expect few activated brain regions. The proposed method can effectively detect EEG sources in various circumstances. Overall, the proposed sparse prior for EEG source localization results in more accurate localization of EEG sources than state-of-the-art approaches.
Highlights
Brain imaging techniques are important tools since they give us the ability to understand the neural mechanisms of complex human behavior in cognitive neuroscience
Where y ∈ RN is the EEG measurement vector acquired by the N electrodes, x ∈ R3M contains the amplitudes of M dipoles along the three spatial dimensions, and H ∈ RN×3M is the lead field matrix that describes the propagation of electromagnetic field from the sources to the sensors and it contains information related to the geometry and conductivity properties of the head. e vector e is an additive white Gaussian noise. e EEG inverse problem of the observation model of equation (1) consists of estimating the vector x given the data y and the lead field matrix H
We describe an approach for this process by using the variational Bayesian (VB) framework
Summary
Brain imaging techniques are important tools since they give us the ability to understand the neural mechanisms of complex human behavior in cognitive neuroscience. Dipolefitting models represent the brain activity using a small number of dipoles and try to estimate the amplitudes, the orientations, and the position of a few dipoles that explain the data [4, 5]. These methods are sensitive to the initial guess of the number of dipoles and their initial locations. On the other hand, distributed-source methods use a large number of dipoles with fixed positions and try to estimate their amplitudes by solving a linear inverse problem [4, 5]. On the other hand, distributed-source methods use a large number of dipoles with fixed positions and try to estimate their amplitudes by solving a linear inverse problem [4, 5]. e EEG linear inverse problem is ill-posed since the number of EEG sources is much larger than the number of Computational Intelligence and Neuroscience
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
