Aristovich, Kirill (2011). Development of Methodologies for the Solution of the Forward Problem in Magnetic-Field Tomography (MFT) Based on Magnetoencephalography (MEG). (Unpublished Doctoral thesis, City University London)
Download (19MB) | Preview
The prime topic of research presented in this report is the development and validation of methodologies for the solution of the forward problem in Magnetic field Tomography based on Magnetoencephalography. Throughout the report full aspects of the accurate solution are discussed, including the development of algorithms and methods for realistic brain model, development of realistic neuronal source, computational approaches, and validation techniques.
Every delivered methodology is tested and analyzed in terms of mathematical and computational errors. Optimizations required for error minimization are performed and discussed. Presented techniques are successfully integrated together for different test problems. Results were compared to experimental data where possible for the most of calculated cases.
Designed human brain model reconstruction algorithms and techniques, which are based on MRI (Magnetic Resonance Imaging) modality, are proved to be the most accurate among existing in terms of geometrical and material properties. Error estimations and algorithm structure delivers the resolution of the model to be the same as practical imaging resolution of the MRI equipment (for presented case was less than 1mm).
Novel neuronal source modelling approach was also presented with partial experimental validation showing improved results in comparison to all existing methods. At the same time developed mathematical basis for practical realization of discussed approach allows computer simulations of any known neuronal formation. Also it is the most suitable method for Finite Element Method (FEM) which was proved to be the best computer solver for complex bio-electrical problems.
The mathematical structure for Inverse problem solution which is based on integrated human brain modelling technique and neuronal source modelling approach is delivered and briefly discussed.
In the concluding part of the report the practical application case of developed techniques is performed and discussed.
Actions (login required)
Downloads per month over past year