Comparison of Four Freely Available Frameworks for Image Processing and Visualization That Use ITK
IEEE Transactions on Visualization and Computer Graphics
| Hi-index | 0.00 |
Solutions to the source localization problem in magnetoencephalography (MEG) aim to noninvasively recover the location of neural activity within a living brain based on the magnetic field produced at detectors external to the head. Many properties and maladies of the human brain can be studied through the MEG source localization problem, including spontaneous activity, neuronal processing, and the sites of epileptic foci. Yet models currently used in MEG source localization often are inaccurate because they are overly simplified and are insufficient to represent the complexity of the physics and physiology involved in the human brain. In this dissertation, a novel modeling and simulation algorithm is developed and studied to determine the effects on the accuracy for magnetic field computations and source localization. The effects that three volume conductor model parameters have on accurate magnetic field calculations and in MEG source localization were examined in a realistic head model. The first study examines the importance of volume currents in the accurate calculation of the magnetic fields measured at the MEG detectors in a realistic head model, and in source localization using this simulated magnetic field data. The second study explores the importance of accurate geometry in brain and head models used for MEG forward calculations and source localization problems. The third study determines the significance of conductivity within the different tissue layers of the human brain, and how they influence both the magnetic field produced external to the brain and the source localization problem. The results show that volume currents, geometry, and conductivity all contribute significantly to accurate magnetic field computation and source localization in magnetoencephalography.