Loading...

EEG Source Localization Using Block Sparse Structure in Reduced Dimension Leadfield

Khanzamani Mohammadi, Ali | 2022

154 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 55393 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Babaiezadeh, Massoud; Ghazizadeh, Ali
  7. Abstract:
  8. Electroencephalogram (EEG) brain source localization carries many potential applications in systems and cognitive neuroscience, and for treatment of various neurological problems such as epilepsy. According to some recent studies, determining the spatial extent of sources and estimating their true time courses have proved challenging. This master's thesis proposes a method for localizing extended brain sources. Cortical surface parcellation has been used to reduce the dimension of the inverse problem without losing much information. The active regions are assumed to be sparse and the time course of the sources exhibits a correlation structure. The reduced dimension problem was then solved by a block sparse solution method. The proposed method was tested using some realistic simulations, then compared to some commonly used methods. Real EEG data was also used to evaluate the method, and the results were compared to fMRI images. The proposed method has demonstrated promising results in determining the active regions and the time course of the sources based on simulations and real data.
  9. Keywords:
  10. Brain Source Localization ; Electroencphalogram Signal ; Singular Value Decomposition (SVD) ; Block Sparse Signal ; Spatio-Temporal Source Localization ; Brain Sources with Extent ; Leadfield Dimension Reduction

 Digital Object List

 Bookmark

No TOC